Loading ...
Sorry, an error occurred while loading the content.

Reducing shimmering

Expand Messages
  • blueslander
    Shimmering is often times a big problem in panos, since the images are moving and zoom levels change etc. This can be reduced significantly by applying the
    Message 1 of 30 , Jan 16, 2011
    • 0 Attachment
      Shimmering is often times a big problem in panos, since the images are moving and zoom levels change etc.

      This can be reduced significantly by applying the "power of two" to the images used.

      Cubefaces should only be any of the following sizes:

      512x512
      1024x1024
      2048x2048
      4096x4096
      8192x8192
      etc...

      Equirectangular images should be:

      512x256
      1024x512
      2048x1024
      4096x2048
      8192x4096
      etc...

      Other graphics (hotspots) should be using sizes based on the same "power of two" concept:

      2
      4
      8
      16
      32
      64
      128
      256
      512
      1024
      etc.

      So a hotspot could be 64x128 or 32x256 or any other combination of the above numbers.

      This will not REMOVE the shimmering, but it does absolutely reduce it, and make things look better.

      Here you can see a test, where the upper image is sized according to the above "power of two" principle, and the lower image is not:

      http://www.kaourantin.net/swf/mipmap.html

      Make your own tests to confirm this.
    • Ken Warner
      That makes no sense to me. Could you provide your rational for this assertion?
      Message 2 of 30 , Jan 16, 2011
      • 0 Attachment
        That makes no sense to me. Could you provide your rational for
        this assertion?

        blueslander wrote:
        > Shimmering is often times a big problem in panos, since the images are moving and zoom levels change etc.
        >
        > This can be reduced significantly by applying the "power of two" to the images used.
        >
        > Cubefaces should only be any of the following sizes:
        >
        > 512x512
        > 1024x1024
        > 2048x2048
        > 4096x4096
        > 8192x8192
        > etc...
        >
        > Equirectangular images should be:
        >
        > 512x256
        > 1024x512
        > 2048x1024
        > 4096x2048
        > 8192x4096
        > etc...
        >
        > Other graphics (hotspots) should be using sizes based on the same "power of two" concept:
        >
        > 2
        > 4
        > 8
        > 16
        > 32
        > 64
        > 128
        > 256
        > 512
        > 1024
        > etc.
        >
        > So a hotspot could be 64x128 or 32x256 or any other combination of the above numbers.
        >
        > This will not REMOVE the shimmering, but it does absolutely reduce it, and make things look better.
        >
        > Here you can see a test, where the upper image is sized according to the above "power of two" principle, and the lower image is not:
        >
        > http://www.kaourantin.net/swf/mipmap.html
        >
        > Make your own tests to confirm this.
        >
        >
      • Fulvio Senore
        +1. It does not make sense to me too.
        Message 3 of 30 , Jan 16, 2011
        • 0 Attachment
          +1.

          It does not make sense to me too.

          Il 16/01/2011 18.45, Ken Warner ha scritto:
          > That makes no sense to me. Could you provide your rational for
          > this assertion?
          >
          > blueslander wrote:
          >> Shimmering is often times a big problem in panos, since the images are moving and zoom levels change etc.
          >>
          >> This can be reduced significantly by applying the "power of two" to the images used.
          >>
          >> Cubefaces should only be any of the following sizes:
          >>
          >> 512x512
          >> 1024x1024
          >> 2048x2048
          >> 4096x4096
          >> 8192x8192
          >> etc...
          >>
          >> Equirectangular images should be:
          >>
          >> 512x256
          >> 1024x512
          >> 2048x1024
          >> 4096x2048
          >> 8192x4096
          >> etc...
          >>
          >> Other graphics (hotspots) should be using sizes based on the same "power of two" concept:
          >>
          >> 2
          >> 4
          >> 8
          >> 16
          >> 32
          >> 64
          >> 128
          >> 256
          >> 512
          >> 1024
          >> etc.
          >>
          >> So a hotspot could be 64x128 or 32x256 or any other combination of the above numbers.
          >>
          >> This will not REMOVE the shimmering, but it does absolutely reduce it, and make things look better.
          >>
          >> Here you can see a test, where the upper image is sized according to the above "power of two" principle, and the lower image is not:
          >>
          >> http://www.kaourantin.net/swf/mipmap.html
          >>
          >> Make your own tests to confirm this.
          >>
          >>
          >
          >
          > ------------------------------------
          >
        • Rick Drew
          This sounds like moire as much as anything else. When lines overlap at opposing angles, the shimmering effect occurs. It s also evident in many Blu-ray movies
          Message 4 of 30 , Jan 16, 2011
          • 0 Attachment
            This sounds like moire as much as anything else. When lines overlap at
            opposing angles, the shimmering effect occurs. It's also evident in many
            Blu-ray movies on hidef TV's - ever see the opening of the last batman
            flick? When the Chicago skyline moves across the screen, the buildings with
            vertical lines look crazy. The more detail in the pano, the more I see this
            effect. Rare a couple of years ago, a lot more common today.



            Take a look at http://en.wikipedia.org/wiki/Moir%C3%A9_pattern



            Several good examples, plus the math for those of you so inclined.



            Rick Drew



            [Non-text portions of this message have been removed]
          • Trausti Hraunfjord
            The technical explanations: http://help.adobe.com/en_US/as3/mobile/WS4bebcd66a74275c3-576ba64d124318d7189-7ff7.html
            Message 5 of 30 , Jan 16, 2011
            • 0 Attachment
              The technical explanations:

              http://help.adobe.com/en_US/as3/mobile/WS4bebcd66a74275c3-576ba64d124318d7189-7ff7.html

              http://number-none.com/product/Mipmapping,%20Part%201/index.html

              http://number-none.com/product/Mipmapping,%20Part%202/index.html

              http://www.kaourantin.net/2007/06/mip-map-what.html

              http://www.kaourantin.net/swf/mipmap.html

              http://blog.nobien.net/examples/mipmapping/

              And as initially stated, you can perform your own tests to see how much
              reduction of the shimmering effect you can get. Try using cubefaces of
              2048x2048 vs.2000x2000. I did, and there was a very noticable difference.

              Trausti





              On Sun, Jan 16, 2011 at 1:49 PM, Rick Drew <rick@...> wrote:

              >
              >
              > This sounds like moire as much as anything else. When lines overlap at
              > opposing angles, the shimmering effect occurs. It's also evident in many
              > Blu-ray movies on hidef TV's - ever see the opening of the last batman
              > flick? When the Chicago skyline moves across the screen, the buildings with
              > vertical lines look crazy. The more detail in the pano, the more I see this
              > effect. Rare a couple of years ago, a lot more common today.
              >
              > Take a look at http://en.wikipedia.org/wiki/Moir%C3%A9_pattern
              >
              > Several good examples, plus the math for those of you so inclined.
              >
              > Rick Drew
              >
              > [Non-text portions of this message have been removed]
              >
              >
              >


              [Non-text portions of this message have been removed]
            • ptgroup
              This was also mentioned somewhere with the Immervision-Tools. Faster loading and better display- was what they said I guess. Ok- it was concerning JAVA and I
              Message 6 of 30 , Jan 16, 2011
              • 0 Attachment
                This was also mentioned somewhere with the Immervision-Tools.
                Faster loading and better display- was what they said I guess.
                Ok- it was concerning JAVA and I don´t know if this also affected their Flash.

                But it made sense to me, since we are in a binary world.

                Ciao
                Mike
                ----------------------------
                360° VR Fotografie:
                http://www.360de.de

                NEU: Abstrakte Fotografie unter:
                http://www.abstraktfoto.de
                -----------------------------
                Aktuelles vom Virtugrafen:
                http://virtugraf.wordpress.com

                -----Ursprüngliche Nachricht-----
                Von: PanoToolsNG@yahoogroups.com [mailto:PanoToolsNG@yahoogroups.com]Im
                Auftrag von Trausti Hraunfjord
                Gesendet: Sonntag, 16. Januar 2011 19:57
                An: PanoToolsNG@yahoogroups.com
                Betreff: Re: [PanoToolsNG] Reducing shimmering


                The technical explanations:

                http://help.adobe.com/en_US/as3/mobile/WS4bebcd66a74275c3-576ba64d124318d7189-7ff7.html

                http://number-none.com/product/Mipmapping,%20Part%201/index.html

                http://number-none.com/product/Mipmapping,%20Part%202/index.html

                http://www.kaourantin.net/2007/06/mip-map-what.html

                http://www.kaourantin.net/swf/mipmap.html

                http://blog.nobien.net/examples/mipmapping/

                And as initially stated, you can perform your own tests to see how much
                reduction of the shimmering effect you can get. Try using cubefaces of
                2048x2048 vs.2000x2000. I did, and there was a very noticable difference.

                Trausti





                On Sun, Jan 16, 2011 at 1:49 PM, Rick Drew <rick@...> wrote:

                >
                >
                > This sounds like moire as much as anything else. When lines overlap at
                > opposing angles, the shimmering effect occurs. It's also evident in many
                > Blu-ray movies on hidef TV's - ever see the opening of the last batman
                > flick? When the Chicago skyline moves across the screen, the buildings with
                > vertical lines look crazy. The more detail in the pano, the more I see this
                > effect. Rare a couple of years ago, a lot more common today.
                >
                > Take a look at http://en.wikipedia.org/wiki/Moir%C3%A9_pattern
                >
                > Several good examples, plus the math for those of you so inclined.
                >
                > Rick Drew
                >
                > [Non-text portions of this message have been removed]
                >
                >
                >


                [Non-text portions of this message have been removed]



                ------------------------------------

                --
              • Ken Warner
                I think that the question is: Is shimmering caused by what interpolator is used in down sampling or is shimmering caused by moire with the pixels on your
                Message 7 of 30 , Jan 16, 2011
                • 0 Attachment
                  I think that the question is:

                  Is shimmering caused by what interpolator is used in
                  down sampling or is shimmering caused by moire with
                  the pixels on your screen?

                  Jaggies are caused by down sampling/interpolator interactions
                  that can be reduced with careful programming that will increase
                  CPU effort in some instances. But I think most shimmering is caused by moire.

                  Just my point of view....

                  Trausti Hraunfjord wrote:
                  > The technical explanations:
                  >
                  > http://help.adobe.com/en_US/as3/mobile/WS4bebcd66a74275c3-576ba64d124318d7189-7ff7.html
                  >
                  > http://number-none.com/product/Mipmapping,%20Part%201/index.html
                  >
                  > http://number-none.com/product/Mipmapping,%20Part%202/index.html
                  >
                  > http://www.kaourantin.net/2007/06/mip-map-what.html
                  >
                  > http://www.kaourantin.net/swf/mipmap.html
                  >
                  > http://blog.nobien.net/examples/mipmapping/
                  >
                  > And as initially stated, you can perform your own tests to see how much
                  > reduction of the shimmering effect you can get. Try using cubefaces of
                  > 2048x2048 vs.2000x2000. I did, and there was a very noticable difference.
                  >
                  > Trausti
                  >
                  >
                  >
                  >
                  >
                  > On Sun, Jan 16, 2011 at 1:49 PM, Rick Drew <rick@...> wrote:
                  >
                  >>
                  >> This sounds like moire as much as anything else. When lines overlap at
                  >> opposing angles, the shimmering effect occurs. It's also evident in many
                  >> Blu-ray movies on hidef TV's - ever see the opening of the last batman
                  >> flick? When the Chicago skyline moves across the screen, the buildings with
                  >> vertical lines look crazy. The more detail in the pano, the more I see this
                  >> effect. Rare a couple of years ago, a lot more common today.
                  >>
                  >> Take a look at http://en.wikipedia.org/wiki/Moir%C3%A9_pattern
                  >>
                  >> Several good examples, plus the math for those of you so inclined.
                  >>
                  >> Rick Drew
                  >>
                  >> [Non-text portions of this message have been removed]
                  >>
                  >>
                  >>
                  >
                  >
                  > [Non-text portions of this message have been removed]
                  >
                  >
                  >
                  > ------------------------------------
                  >
                • Keith Martin
                  ... When using QTVR (remember that? ;-) if you left the quality level to a medium or low setting then the image would drop to a harsher, more pixellated render
                  Message 8 of 30 , Jan 16, 2011
                  • 0 Attachment
                    Sometime around 16/1/11 (at 11:44 -0800) Ken Warner said:

                    >I think most shimmering is caused by moire.

                    When using QTVR (remember that? ;-) if you left the quality level to
                    a medium or low setting then the image would drop to a harsher, more
                    pixellated render while it was moving. That would cause unpleasant
                    shimmering. Modern computers have no problem rotating and warping
                    images while also dealing with smoothing and antialiasing, so *that*
                    problem is something that can be avoided without performance worries.

                    @blueslander, is the problem in question something that can be
                    affected by render quality settings in some way? What VR format are
                    we talking about here? What player engine?

                    k
                  • Ken Warner
                    Java is particularly susceptible to shimmer and jaggies because there s no interface to the GPU which can do anti-aliasing and other nice smoothings very fast.
                    Message 9 of 30 , Jan 16, 2011
                    • 0 Attachment
                      Java is particularly susceptible to shimmer and jaggies
                      because there's no interface to the GPU which can do
                      anti-aliasing and other nice smoothings very fast.

                      So in my viewer (for example) I do all up and down sampling
                      with ordinary math done by the CPU and to do anything more
                      than the simplest interpolating slows the viewer down
                      so that it's unusable.

                      If there was a nice interface to the GPU in Java, one could
                      take advantage of all the slick graphics computations that
                      a typical GPU offers.

                      Oh well....

                      Keith Martin wrote:
                      > Sometime around 16/1/11 (at 11:44 -0800) Ken Warner said:
                      >
                      >> I think most shimmering is caused by moire.
                      >
                      > When using QTVR (remember that? ;-) if you left the quality level to
                      > a medium or low setting then the image would drop to a harsher, more
                      > pixellated render while it was moving. That would cause unpleasant
                      > shimmering. Modern computers have no problem rotating and warping
                      > images while also dealing with smoothing and antialiasing, so *that*
                      > problem is something that can be avoided without performance worries.
                      >
                      > @blueslander, is the problem in question something that can be
                      > affected by render quality settings in some way? What VR format are
                      > we talking about here? What player engine?
                      >
                      > k
                      >
                    • Wim Koornneef
                      Some time ago I tested Pano2VR 3 for making multi-res panos with normal sized equirectangulars. The test was not optimized to prevent shimmering, I used steps
                      Message 10 of 30 , Jan 17, 2011
                      • 0 Attachment
                        Some time ago I tested Pano2VR 3 for making multi-res panos with normal sized
                        equirectangulars.
                        The test was not optimized to prevent shimmering, I used steps of approx.
                        1.5 instead of 2 (the default of Pano2VR 3) to make the transition from one
                        tile level to the other invisible, but the effect of using multiple tile
                        sizes for the reduction of shimmering is very clear to see.

                        Here is a link to the test with examples that I posted on the Garden Gnome
                        forum:
                        http://gardengnomesoftware.com/forum/viewtopic.php?f=6&t=5172

                        Wim
                        --
                        View this message in context: http://panotoolsng.586017.n4.nabble.com/Reducing-shimmering-tp3220213p3220864.html
                        Sent from the PanoToolsNG mailing list archive at Nabble.com.
                      • Erik Krause
                        ... Interestingly the Java based PTViewer (since version 2.4) beats all Flash players in terms of shimmering... -- Erik Krause http://www.erik-krause.de
                        Message 11 of 30 , Jan 17, 2011
                        • 0 Attachment
                          Am 17.01.2011 00:58, schrieb Ken Warner:
                          > Java is particularly susceptible to shimmer and jaggies
                          > because there's no interface to the GPU which can do
                          > anti-aliasing and other nice smoothings very fast.

                          Interestingly the Java based PTViewer (since version 2.4) beats all
                          Flash players in terms of shimmering...

                          --
                          Erik Krause
                          http://www.erik-krause.de
                        • Fulvio Senore
                          ... Because (at least my version) by default it uses bilinear interpolation even when panning (slowly). This reduces shimmering a lot at the expense of some
                          Message 12 of 30 , Jan 17, 2011
                          • 0 Attachment
                            Il 17/01/2011 20.58, Erik Krause ha scritto:
                            > Am 17.01.2011 00:58, schrieb Ken Warner:
                            >> Java is particularly susceptible to shimmer and jaggies
                            >> because there's no interface to the GPU which can do
                            >> anti-aliasing and other nice smoothings very fast.
                            >
                            > Interestingly the Java based PTViewer (since version 2.4) beats all
                            > Flash players in terms of shimmering...
                            >
                            Because (at least my version) by default it uses bilinear interpolation
                            even when panning (slowly). This reduces shimmering a lot at the expense
                            of some speed.

                            You can also see the difference in FSPViewer: there is an option for
                            changing the interpolator.

                            Fulvio Senore
                          • Ken Warner
                            It s also going to depend on the size of the image and the size of the window in which the pano is displayed and what sacrifices are made for speed. With many
                            Message 13 of 30 , Jan 17, 2011
                            • 0 Attachment
                              It's also going to depend on the size of the image
                              and the size of the window in which the pano is
                              displayed and what sacrifices are made for
                              speed.

                              With many panos, going full window reduces shimmer.

                              And also, the cleverness of the programmer :-)

                              Fulvio Senore wrote:
                              > Il 17/01/2011 20.58, Erik Krause ha scritto:
                              >> Am 17.01.2011 00:58, schrieb Ken Warner:
                              >>> Java is particularly susceptible to shimmer and jaggies
                              >>> because there's no interface to the GPU which can do
                              >>> anti-aliasing and other nice smoothings very fast.
                              >> Interestingly the Java based PTViewer (since version 2.4) beats all
                              >> Flash players in terms of shimmering...
                              >>
                              > Because (at least my version) by default it uses bilinear interpolation
                              > even when panning (slowly). This reduces shimmering a lot at the expense
                              > of some speed.
                              >
                              > You can also see the difference in FSPViewer: there is an option for
                              > changing the interpolator.
                              >
                              > Fulvio Senore
                              >
                              >
                              > ------------------------------------
                              >
                            • dalileis
                              Oh, my, Trausti, if this is correct than you are onto something huge. Your example really says it all. Thank you and all kudos to you. This is a very important
                              Message 14 of 30 , Apr 10, 2011
                              • 0 Attachment
                                Oh, my, Trausti, if this is correct than you are onto something huge.

                                Your example really says it all.

                                Thank you and all kudos to you.

                                This is a very important post you made back in January 2011.


                                --- In PanoToolsNG@yahoogroups.com, "blueslander" <trausti.hraunfjord@...> wrote:
                                >
                                > Shimmering is often times a big problem in panos, since the images are moving and zoom levels change etc.
                                >
                                > This can be reduced significantly by applying the "power of two" to the images used.
                                >
                                > Cubefaces should only be any of the following sizes:
                                >
                                > 512x512
                                > 1024x1024
                                > 2048x2048
                                > 4096x4096
                                > 8192x8192
                                > etc...
                                >
                                > Equirectangular images should be:
                                >
                                > 512x256
                                > 1024x512
                                > 2048x1024
                                > 4096x2048
                                > 8192x4096
                                > etc...
                                >
                                > Other graphics (hotspots) should be using sizes based on the same "power of two" concept:
                                >
                                > 2
                                > 4
                                > 8
                                > 16
                                > 32
                                > 64
                                > 128
                                > 256
                                > 512
                                > 1024
                                > etc.
                                >
                                > So a hotspot could be 64x128 or 32x256 or any other combination of the above numbers.
                                >
                                > This will not REMOVE the shimmering, but it does absolutely reduce it, and make things look better.
                                >
                                > Here you can see a test, where the upper image is sized according to the above "power of two" principle, and the lower image is not:
                                >
                                > http://www.kaourantin.net/swf/mipmap.html
                                >
                                > Make your own tests to confirm this.
                                >
                              • Ken Warner
                                This is just wrong. Screen size, window size, pixel pitch, graphics hardware and certainly many other factors play into moire generation. Nice simple idea --
                                Message 15 of 30 , Apr 11, 2011
                                • 0 Attachment
                                  This is just wrong. Screen size, window size, pixel pitch,
                                  graphics hardware and certainly many other factors play into
                                  moire generation. Nice simple idea -- wish it were so.

                                  dalileis wrote:
                                  > Oh, my, Trausti, if this is correct than you are onto something huge.
                                  >
                                  > Your example really says it all.
                                  >
                                  > Thank you and all kudos to you.
                                  >
                                  > This is a very important post you made back in January 2011.
                                  >
                                  >
                                  > --- In PanoToolsNG@yahoogroups.com, "blueslander" <trausti.hraunfjord@...> wrote:
                                  >> Shimmering is often times a big problem in panos, since the images are moving and zoom levels change etc.
                                  >>
                                  >> This can be reduced significantly by applying the "power of two" to the images used.
                                  >>
                                  >> Cubefaces should only be any of the following sizes:
                                  >>
                                  >> 512x512
                                  >> 1024x1024
                                  >> 2048x2048
                                  >> 4096x4096
                                  >> 8192x8192
                                  >> etc...
                                  >>
                                  >> Equirectangular images should be:
                                  >>
                                  >> 512x256
                                  >> 1024x512
                                  >> 2048x1024
                                  >> 4096x2048
                                  >> 8192x4096
                                  >> etc...
                                  >>
                                  >> Other graphics (hotspots) should be using sizes based on the same "power of two" concept:
                                  >>
                                  >> 2
                                  >> 4
                                  >> 8
                                  >> 16
                                  >> 32
                                  >> 64
                                  >> 128
                                  >> 256
                                  >> 512
                                  >> 1024
                                  >> etc.
                                  >>
                                  >> So a hotspot could be 64x128 or 32x256 or any other combination of the above numbers.
                                  >>
                                  >> This will not REMOVE the shimmering, but it does absolutely reduce it, and make things look better.
                                  >>
                                  >> Here you can see a test, where the upper image is sized according to the above "power of two" principle, and the lower image is not:
                                  >>
                                  >> http://www.kaourantin.net/swf/mipmap.html
                                  >>
                                  >> Make your own tests to confirm this.
                                  >>
                                  >
                                  >
                                  >
                                • Yazan Sboul
                                  well it works in the example!! http://www.kaourantin.net/swf/mipmap.htmlI would like to know how this translates across browsers and what other factors are
                                  Message 16 of 30 , Apr 11, 2011
                                  • 0 Attachment
                                    well it works in the example!! http://www.kaourantin.net/swf/mipmap.htmlI would like to know how this translates across browsers and what other factors are involved (which we might also be able to address), but the pixel size certainly seems to be a significant factor - how else do you explain the example (works fine with firefox on 30" screen).
                                    Y.S

                                    To: PanoToolsNG@yahoogroups.com
                                    From: kwarner000@...
                                    Date: Mon, 11 Apr 2011 08:19:22 -0700
                                    Subject: Re: [PanoToolsNG] Re: Reducing shimmering




























                                    This is just wrong. Screen size, window size, pixel pitch,

                                    graphics hardware and certainly many other factors play into

                                    moire generation. Nice simple idea -- wish it were so.



                                    dalileis wrote:

                                    > Oh, my, Trausti, if this is correct than you are onto something huge.

                                    >

                                    > Your example really says it all.

                                    >

                                    > Thank you and all kudos to you.

                                    >

                                    > This is a very important post you made back in January 2011.

                                    >

                                    >

                                    > --- In PanoToolsNG@yahoogroups.com, "blueslander" <trausti.hraunfjord@...> wrote:

                                    >> Shimmering is often times a big problem in panos, since the images are moving and zoom levels change etc.

                                    >>

                                    >> This can be reduced significantly by applying the "power of two" to the images used.

                                    >>

                                    >> Cubefaces should only be any of the following sizes:

                                    >>

                                    >> 512x512

                                    >> 1024x1024

                                    >> 2048x2048

                                    >> 4096x4096

                                    >> 8192x8192

                                    >> etc...

                                    >>

                                    >> Equirectangular images should be:

                                    >>

                                    >> 512x256

                                    >> 1024x512

                                    >> 2048x1024

                                    >> 4096x2048

                                    >> 8192x4096

                                    >> etc...

                                    >>

                                    >> Other graphics (hotspots) should be using sizes based on the same "power of two" concept:

                                    >>

                                    >> 2

                                    >> 4

                                    >> 8

                                    >> 16

                                    >> 32

                                    >> 64

                                    >> 128

                                    >> 256

                                    >> 512

                                    >> 1024

                                    >> etc.

                                    >>

                                    >> So a hotspot could be 64x128 or 32x256 or any other combination of the above numbers.

                                    >>

                                    >> This will not REMOVE the shimmering, but it does absolutely reduce it, and make things look better.

                                    >>

                                    >> Here you can see a test, where the upper image is sized according to the above "power of two" principle, and the lower image is not:

                                    >>

                                    >> http://www.kaourantin.net/swf/mipmap.html

                                    >>

                                    >> Make your own tests to confirm this.

                                    >>

                                    >

                                    >

                                    >
















                                    [Non-text portions of this message have been removed]
                                  • Ken Warner
                                    If you think it works, please explain why it works.
                                    Message 17 of 30 , Apr 11, 2011
                                    • 0 Attachment
                                      If you think it works, please explain why it works.

                                      Yazan Sboul wrote:
                                      > well it works in the example!! http://www.kaourantin.net/swf/mipmap.html
                                    • Sacha Griffin
                                      What s your problem Ken? You explain why the shimmer isn t visible when the graphic size is different. ;) s From: PanoToolsNG@yahoogroups.com
                                      Message 18 of 30 , Apr 11, 2011
                                      • 0 Attachment
                                        What's your problem Ken?

                                        You explain why the shimmer isn't visible when the graphic size is
                                        different.

                                        ;)



                                        s



                                        From: PanoToolsNG@yahoogroups.com [mailto:PanoToolsNG@yahoogroups.com] On
                                        Behalf Of Ken Warner
                                        Sent: Monday, April 11, 2011 2:03 PM
                                        To: PanoToolsNG@yahoogroups.com
                                        Subject: Re: [PanoToolsNG] Re: Reducing shimmering





                                        If you think it works, please explain why it works.

                                        Yazan Sboul wrote:
                                        > well it works in the example!! http://www.kaourantin.net/swf/mipmap.html





                                        [Non-text portions of this message have been removed]
                                      • Ben Newhouse
                                        I think both of you are right - OpenGL has been using MipMaps for decades and automatically generates mipmaps whenever you push down a texture (in addition to
                                        Message 19 of 30 , Apr 11, 2011
                                        • 0 Attachment
                                          I think both of you are right - OpenGL has been using MipMaps for decades
                                          and automatically generates mipmaps whenever you push down a texture (in
                                          addition to taking into account all of the things Ken mentioned). I don't
                                          however know by default what Flash's default rendering engine does as I
                                          would suppose it's not yet fully accelerated and thus is doing everything on
                                          the CPU (mipmap generation is a lot more expensive there).

                                          -Ben

                                          On Mon, Apr 11, 2011 at 11:03 AM, Ken Warner <kwarner000@...> wrote:

                                          >
                                          >
                                          > If you think it works, please explain why it works.
                                          >
                                          >
                                          > Yazan Sboul wrote:
                                          > > well it works in the example!! http://www.kaourantin.net/swf/mipmap.html
                                          >
                                          >
                                          >


                                          [Non-text portions of this message have been removed]
                                        • ptgroup
                                          This was also told in the Immervision forum (I guess I read it there) JAVA concerning, to keep the loading-time as short as possible and the quality as high as
                                          Message 20 of 30 , Apr 11, 2011
                                          • 0 Attachment
                                            This was also told in the Immervision forum (I guess I read it there) JAVA concerning, to keep the
                                            loading-time as short as possible and the quality as high as possible, JAVA concerning though.
                                            But I can imagine that it might be an issue in other viewers too.

                                            As far as I remember it was due to the memory management in their program, but I´m not sure - it´s
                                            some times
                                            ago.

                                            Ciao
                                            Mike

                                            http://www.virtugraf.de <http://www.virtugraf.de/>
                                            ----------------------------
                                            360° VR Fotografie:
                                            http://www.360de.de <http://www.360de.de/>

                                            NEU: Abstrakte Fotografie unter:
                                            http://www.abstraktfoto.de <http://www.abstraktfoto.de/>
                                            -----------------------------
                                            Aktuelles vom Virtugrafen:
                                            http://virtugraf.wordpress.com <http://virtugraf.wordpress.com/>


                                            _____

                                            Von: PanoToolsNG@yahoogroups.com [mailto:PanoToolsNG@yahoogroups.com] Im Auftrag von Ken Warner
                                            Gesendet: Montag, 11. April 2011 20:03
                                            An: PanoToolsNG@yahoogroups.com
                                            Betreff: Re: [PanoToolsNG] Re: Reducing shimmering




                                            If you think it works, please explain why it works.

                                            Yazan Sboul wrote:
                                            > well it works in the example!! http://www.kaourantin.net/swf/mipmap.html




                                            _____

                                            I am using the Free version of SPAMfighter <http://www.spamfighter.com/len> .
                                            SPAMfighter has removed 218 of my spam emails to date.

                                            Do you have a slow PC? <http://www.spamfighter.com/SLOW-PCfighter?cid=sigen> Try free scan!


                                            [Non-text portions of this message have been removed]
                                          • Yazan Sboul
                                            Ken, I m with you on that - it would be nice to know exactly whats going on and I m sure that will help us fine tune the results. However It does have a
                                            Message 21 of 30 , Apr 11, 2011
                                            • 0 Attachment
                                              Ken, I'm with you on that - it would be nice to know exactly whats going on and I'm sure that will help us fine tune the results. However It does have a positive effect as seen on the example (so long as the images sizes in this example are not too different). I can't explain this positive improvement, I'm just reporting it working on my system. If you think what I'm seeing is an illusion or due to some other factor, the obligation to explain it is on you. :)
                                              Please explain.
                                              CheersY.S

                                              To: PanoToolsNG@yahoogroups.com
                                              From: kwarner000@...
                                              Date: Mon, 11 Apr 2011 11:03:05 -0700
                                              Subject: Re: [PanoToolsNG] Re: Reducing shimmering




























                                              If you think it works, please explain why it works.



                                              Yazan Sboul wrote:

                                              > well it works in the example!! http://www.kaourantin.net/swf/mipmap.html
















                                              [Non-text portions of this message have been removed]
                                            • Ken Warner
                                              I see it on my machine also. But that doesn t mean that there is a universal principle working. It may be just an accident of that particular image and the
                                              Message 22 of 30 , Apr 11, 2011
                                              • 0 Attachment
                                                I see it on my machine also. But that doesn't mean that there is a
                                                universal principle working. It may be just an accident of that
                                                particular image and the way it's being resized.

                                                One good example does not make a universal rule.

                                                Yazan Sboul wrote:
                                                > Ken, I'm with you on that - it would be nice to know exactly whats going on and I'm sure that will help us fine tune the results. However It does have a positive effect as seen on the example (so long as the images sizes in this example are not too different). I can't explain this positive improvement, I'm just reporting it working on my system. If you think what I'm seeing is an illusion or due to some other factor, the obligation to explain it is on you. :)
                                                > Please explain.
                                                > CheersY.S
                                                >
                                                > To: PanoToolsNG@yahoogroups.com
                                                > From: kwarner000@...
                                                > Date: Mon, 11 Apr 2011 11:03:05 -0700
                                                > Subject: Re: [PanoToolsNG] Re: Reducing shimmering
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                > If you think it works, please explain why it works.
                                                >
                                                >
                                                >
                                                > Yazan Sboul wrote:
                                                >
                                                >> well it works in the example!! http://www.kaourantin.net/swf/mipmap.html
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                >
                                                > [Non-text portions of this message have been removed]
                                                >
                                                >
                                                >
                                                > ------------------------------------
                                                >
                                              • Ken Warner
                                                It s not up to me to explain it. I didn t do it.
                                                Message 23 of 30 , Apr 11, 2011
                                                • 0 Attachment
                                                  It's not up to me to explain it. I didn't do it.


                                                  Sacha Griffin wrote:
                                                  > What's your problem Ken?
                                                  >
                                                  > You explain why the shimmer isn't visible when the graphic size is
                                                  > different.
                                                  >
                                                  > ;)
                                                • Yazan Sboul
                                                  Ken, might have a point. Need to test this on a pano. I ll do it today, will get the results up for everyone to see. But Ken, You can t say the results are
                                                  Message 24 of 30 , Apr 11, 2011
                                                  • 0 Attachment
                                                    Ken, might have a point. Need to test this on a pano. I'll do it today, will get the results up for everyone to see. But Ken, You can't say the results are pointless just because there are other variables. If you works, you explain it. I'll just use it. :)

                                                    To: PanoToolsNG@yahoogroups.com
                                                    From: kwarner000@...
                                                    Date: Mon, 11 Apr 2011 12:07:49 -0700
                                                    Subject: Re: [PanoToolsNG] Re: Reducing shimmering




























                                                    It's not up to me to explain it. I didn't do it.



                                                    Sacha Griffin wrote:

                                                    > What's your problem Ken?

                                                    >

                                                    > You explain why the shimmer isn't visible when the graphic size is

                                                    > different.

                                                    >

                                                    > ;)
















                                                    [Non-text portions of this message have been removed]
                                                  • Yazan Sboul
                                                    Dam! The results are in, and Ken is completely right. http://p2vhost.pan2view.com/i_hate_ken/smug_ken.html Basically, apologies to ken, the resizing image
                                                    Message 25 of 30 , Apr 11, 2011
                                                    • 0 Attachment
                                                      Dam! The results are in, and Ken is completely right.
                                                      http://p2vhost.pan2view.com/i_hate_ken/smug_ken.html
                                                      Basically, apologies to ken, the resizing image makes no sense. I don't know why it works.
                                                      If you look at the at the example above the fist pano that loads on the top of the page is in the recommended 2048x2048 "power of 2 rule" (which is eerily reminiscent of the language used by homeopathy). Look at the back of the scene to the the glitter shimmering wall by panning to the left.
                                                      The second scene at the bottom is also 2048x2048. But this scene will resize are you resize your browser window. Notice that as it is already bigger there is no shimmering. Meanwhile, if you make the window smaller shimmering will appear.
                                                      Now, if you use the picmenu to click to the second scene this is in 1300px (admittedly smaller, but not following the rule). Shimmering actually get better, but this may just be because its a smaller image. The interesting point is that as you make the screen smaller (switch to the second scene in the resizable pano) the shimmering gets worse again.
                                                      Finally, zoom! it seems to be all due to zoom. If you zoom in shimmering goes - even on larger images. If you zoom out (like making the frame smaller) shimmering intensifies.
                                                      If your cube faces are smaller = less shimmer, if larger = more shimmer. The "power of 2" Rule, as much as it pains me to say it.... is Rubbish. Ken was absolutely right.
                                                      Dam!... And thanks Ken.


                                                      > To: panotoolsng@yahoogroups.com
                                                      > From: yazansboul@...
                                                      > Date: Mon, 11 Apr 2011 19:24:07 +0000
                                                      > Subject: RE: [PanoToolsNG] Re: Reducing shimmering
                                                      >
                                                      >
                                                      > Ken, might have a point. Need to test this on a pano. I'll do it today, will get the results up for everyone to see. But Ken, You can't say the results are pointless just because there are other variables. If you works, you explain it. I'll just use it. :)
                                                      >
                                                      > To: PanoToolsNG@yahoogroups.com
                                                      > From: kwarner000@...
                                                      > Date: Mon, 11 Apr 2011 12:07:49 -0700
                                                      > Subject: Re: [PanoToolsNG] Re: Reducing shimmering
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      > It's not up to me to explain it. I didn't do it.
                                                      >
                                                      >
                                                      >
                                                      > Sacha Griffin wrote:
                                                      >
                                                      > > What's your problem Ken?
                                                      >
                                                      > >
                                                      >
                                                      > > You explain why the shimmer isn't visible when the graphic size is
                                                      >
                                                      > > different.
                                                      >
                                                      > >
                                                      >
                                                      > > ;)
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      >
                                                      > [Non-text portions of this message have been removed]
                                                      >
                                                      >
                                                      >
                                                      > ------------------------------------
                                                      >
                                                      > --
                                                      >
                                                      >
                                                      >


                                                      [Non-text portions of this message have been removed]
                                                    • Erik Krause
                                                      ... The correct name for what we call shimmering is aliasing. A short description can be found here: http://wiki.panotools.org/Aliasing There are better and
                                                      Message 26 of 30 , Apr 11, 2011
                                                      • 0 Attachment
                                                        Am 11.04.2011 20:39, schrieb Yazan Sboul:
                                                        > it would be nice to know exactly whats going on and I'm sure that
                                                        > will help us fine tune the results

                                                        The correct name for what we call "shimmering" is aliasing. A short
                                                        description can be found here: http://wiki.panotools.org/Aliasing

                                                        There are better and worse interpolators. Unfortunately the better ones
                                                        can't be used in real time viewing. That's why mip-mapping is used there
                                                        if possible. As far as I know that's not possible in standard Flash
                                                        (opposed to Java, Shockwave or proprietary viewers like QT or DevalVR).
                                                        Helmut Dersch introduced mip-mapping in his PTViewer exactly 10 years
                                                        ago. The webGL (HTML5) examples I watched aren't anti-aliased either...

                                                        --
                                                        Erik Krause
                                                        http://www.erik-krause.de
                                                      • dalileis
                                                        Very sorry that it is so, but Ken & Yazan are right. I have resized cubefaces to 2048 x 2048 (from 2096 x 2096), and there is NO improvement. Only loss in
                                                        Message 27 of 30 , Apr 11, 2011
                                                        • 0 Attachment
                                                          Very sorry that it is so, but Ken & Yazan are right.

                                                          I have resized cubefaces to 2048 x 2048 (from 2096 x 2096), and there is NO improvement. Only loss in quality because of more interpolation.

                                                          My FPP quality settings are at:

                                                          qualityStatic=high
                                                          qualityMotion=low
                                                          qualityStatic2=high
                                                          qualityMotion2=medium

                                                          And no, there is still no improvement in anti-aliasing even if I change motion parameters to high.

                                                          In fact, shimmering (aliasing) is occurring even with lossless tifs, in DevalVR.

                                                          So there. This ugly problem remains unsolved.

                                                          As to why Traustis example works so well? No idea.

                                                          I hoped to see the same dramatic effect on my own test cubefaces, but no such effect was evident, again, sorry to say...



                                                          --- In PanoToolsNG@yahoogroups.com, Yazan Sboul <yazansboul@...> wrote:
                                                          >
                                                          >
                                                          > Dam! The results are in, and Ken is completely right.
                                                          > http://p2vhost.pan2view.com/i_hate_ken/smug_ken.html
                                                          > Basically, apologies to ken, the resizing image makes no sense. I don't know why it works.
                                                          > If you look at the at the example above the fist pano that loads on the top of the page is in the recommended 2048x2048 "power of 2 rule" (which is eerily reminiscent of the language used by homeopathy). Look at the back of the scene to the the glitter shimmering wall by panning to the left.
                                                          > The second scene at the bottom is also 2048x2048. But this scene will resize are you resize your browser window. Notice that as it is already bigger there is no shimmering. Meanwhile, if you make the window smaller shimmering will appear.
                                                          > Now, if you use the picmenu to click to the second scene this is in 1300px (admittedly smaller, but not following the rule). Shimmering actually get better, but this may just be because its a smaller image. The interesting point is that as you make the screen smaller (switch to the second scene in the resizable pano) the shimmering gets worse again.
                                                          > Finally, zoom! it seems to be all due to zoom. If you zoom in shimmering goes - even on larger images. If you zoom out (like making the frame smaller) shimmering intensifies.
                                                          > If your cube faces are smaller = less shimmer, if larger = more shimmer. The "power of 2" Rule, as much as it pains me to say it.... is Rubbish. Ken was absolutely right.
                                                          > Dam!... And thanks Ken.
                                                          >
                                                          >
                                                          > > To: panotoolsng@yahoogroups.com
                                                          > > From: yazansboul@...
                                                          > > Date: Mon, 11 Apr 2011 19:24:07 +0000
                                                          > > Subject: RE: [PanoToolsNG] Re: Reducing shimmering
                                                          > >
                                                          > >
                                                          > > Ken, might have a point. Need to test this on a pano. I'll do it today, will get the results up for everyone to see. But Ken, You can't say the results are pointless just because there are other variables. If you works, you explain it. I'll just use it. :)
                                                          > >
                                                          > > To: PanoToolsNG@yahoogroups.com
                                                          > > From: kwarner000@...
                                                          > > Date: Mon, 11 Apr 2011 12:07:49 -0700
                                                          > > Subject: Re: [PanoToolsNG] Re: Reducing shimmering
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > > It's not up to me to explain it. I didn't do it.
                                                          > >
                                                          > >
                                                          > >
                                                          > > Sacha Griffin wrote:
                                                          > >
                                                          > > > What's your problem Ken?
                                                          > >
                                                          > > >
                                                          > >
                                                          > > > You explain why the shimmer isn't visible when the graphic size is
                                                          > >
                                                          > > > different.
                                                          > >
                                                          > > >
                                                          > >
                                                          > > > ;)
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > >
                                                          > > [Non-text portions of this message have been removed]
                                                          > >
                                                          > >
                                                          > >
                                                          > > ------------------------------------
                                                          > >
                                                          > > --
                                                          > >
                                                          > >
                                                          > >
                                                          >
                                                          >
                                                          > [Non-text portions of this message have been removed]
                                                          >
                                                        • no way
                                                          ... This just reminded me that back in the QTVRAS days, cylindrical panos had to have one of the dimensions(don t remember which one tho---)divisible by 4 or
                                                          Message 28 of 30 , Apr 11, 2011
                                                          • 0 Attachment
                                                            --- In PanoToolsNG@yahoogroups.com, "dalileis" <dalileis@...> wrote:
                                                            >
                                                            > Oh, my, Trausti, if this is correct than you are onto something huge.
                                                            >
                                                            > Your example really says it all.
                                                            >
                                                            > Thank you and all kudos to you.
                                                            >
                                                            > This is a very important post you made back in January 2011.
                                                            >
                                                            >
                                                            > --- In PanoToolsNG@yahoogroups.com, "blueslander" <trausti.hraunfjord@> wrote:
                                                            > >
                                                            > > Shimmering is often times a big problem in panos, since the images are moving and zoom levels change etc.
                                                            > >
                                                            > > This can be reduced significantly by applying the "power of two" to the images used.
                                                            > >
                                                            > > Cubefaces should only be any of the following sizes:
                                                            > >
                                                            > > 512x512
                                                            > > 1024x1024
                                                            > > 2048x2048
                                                            > > 4096x4096
                                                            > > 8192x8192
                                                            > > etc...
                                                            > >
                                                            > > Equirectangular images should be:
                                                            > >
                                                            > > 512x256
                                                            > > 1024x512
                                                            > > 2048x1024
                                                            > > 4096x2048
                                                            > > 8192x4096
                                                            > > etc...
                                                            > >
                                                            > > Other graphics (hotspots) should be using sizes based on the same "power of two" concept:
                                                            > >
                                                            > > 2
                                                            > > 4
                                                            > > 8
                                                            > > 16
                                                            > > 32
                                                            > > 64
                                                            > > 128
                                                            > > 256
                                                            > > 512
                                                            > > 1024
                                                            > > etc.
                                                            > >
                                                            > > So a hotspot could be 64x128 or 32x256 or any other combination of the above numbers.
                                                            > >
                                                            > > This will not REMOVE the shimmering, but it does absolutely reduce it, and make things look better.
                                                            > >
                                                            > > Here you can see a test, where the upper image is sized according to the above "power of two" principle, and the lower image is not:
                                                            > >
                                                            > > http://www.kaourantin.net/swf/mipmap.html
                                                            > >
                                                            > > Make your own tests to confirm this.

                                                            This just reminded me that back in the QTVRAS days, cylindrical panos had to have one of the dimensions(don't remember which one tho---)divisible by 4 or it wouldn't output. I'm not sure what the reason was. perhaps its related?
                                                            K
                                                            > >
                                                            >
                                                          • David
                                                            I m not exactly sure what is meant by shimmering , but I have noticed in many image applications that changing image view or size by power of 2 multiples of
                                                            Message 29 of 30 , Apr 12, 2011
                                                            • 0 Attachment
                                                              I'm not exactly sure what is meant by 'shimmering', but I have noticed in many image applications that changing image view or size by power of 2 multiples of the original size results in a better image.

                                                              I assume it is strongly dependent on the re-sampling algorithm used, and the fact that power of 2 resizing is easier and needs less pixel computation and hence is more accurate. Without actually looking at the algorithms I am just assuming that is the case, but it seems to often hold true in real life (digital real life that is).

                                                              A good algorithm may give results good enough that this effect is not noticed, but an inaccurate or optimized for speed on slow computers algorithm certainly degrades the image.

                                                              When viewing images we have the effect of the original processing as well as whatever our current viewing tool is doing as we resize or zoom in/out.

                                                              I'm not offended if anyone tells me I'm full of beans.

                                                              David B
                                                            • texas360dave
                                                              Several years ago I read that to determine the optimum cube face size is actually a mathematical equation. For example, Image width for example 5000 pixels
                                                              Message 30 of 30 , Apr 13, 2011
                                                              • 0 Attachment
                                                                Several years ago I read that to determine the optimum cube face size is actually a mathematical equation. For example, Image width for example 5000 pixels wide divided by Pi 3.1435. Or Image width 5000 divided by 4 (an even number like 2).

                                                                5000
                                                                Pi 3.1415 = 1591.59 or 1592 cube face dimension
                                                                4 = 1250 pixel cube face dimension

                                                                Some cube face conversion software (like Pano2VR) uses this approach.

                                                                Dave still at 360Texas.com

                                                                --- In PanoToolsNG@yahoogroups.com, "David" <dburton97128@...> wrote:
                                                                >
                                                                >
                                                                > I'm not exactly sure what is meant by 'shimmering', but I have noticed in many image applications that changing image view or size by power of 2 multiples of the original size results in a better image.
                                                                >
                                                                > I assume it is strongly dependent on the re-sampling algorithm used, and the fact that power of 2 resizing is easier and needs less pixel computation and hence is more accurate. Without actually looking at the algorithms I am just assuming that is the case, but it seems to often hold true in real life (digital real life that is).
                                                                >
                                                                > A good algorithm may give results good enough that this effect is not noticed, but an inaccurate or optimized for speed on slow computers algorithm certainly degrades the image.
                                                                >
                                                                > When viewing images we have the effect of the original processing as well as whatever our current viewing tool is doing as we resize or zoom in/out.
                                                                >
                                                                > I'm not offended if anyone tells me I'm full of beans.
                                                                >
                                                                > David B
                                                                >
                                                              Your message has been successfully submitted and would be delivered to recipients shortly.