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

RE: Fwd: [regsaudioforum] TMs measurement

Expand Messages
  • jeffstakehifi
    Hi Tom. To find the frequency that is 1/6 octave higher than any frequency, multiply by (2^(1/6)), or about 1.1224620483. best, jeff ---In
    Message 1 of 4 , Oct 2, 2013
    • 0 Attachment

      Hi Tom.

      To find the frequency that is 1/6 octave higher than any frequency, multiply by (2^(1/6)), or about 1.1224620483.

      best,

      jeff



      ---In regsaudioforum@yahoogroups.com, <regsaudioforum@yahoogroups.com> wrote:

      I posted another graph after working with the DSPeaker Dual Core's parametric equalization controls last night.  The new picture of the SynRTA-graphed response is dated 9-30-13 and can be found at: 


      This graph was made using the Dual Core's Typical calibration, with bass dip compensation adjusted to plus 6 dB, and with various parametric equalization tweaks after that.  I think I used about 10 different bands of parametric EQ to reach this result.  

      It's a bit tedious to work with the parametric controls since they are not automatic like the Typical or Advanced calibration of the Dual Core or the target curve equalization of the TacT.  But the parametric controls will give you more control with probably less tedium than a 1/3-octave graphic equalizer.  

      The Dual Core screen is not up to fine adjustments like this since it is not big enough, but even if the Dual Core's screen were bigger I don't think using the Dual Core's screen would be as easy as adjusting the response via SynRTA anyway.  The Dual Core's frequency response graphing is not done in real time whereas SynRTA and True RTA provide real time display of the results as you adjust the filters.  You can work much more quickly with a real time display of your adjustment results.

      The process of fine tuning the response is complicated by the fact that I obviously cannot sit in the listening chair to listen to the results of tuning until the process is done since the microphone occupies the spot where my head would be.  Then there is the fact that I have to make a judgment about whether the graphed response will be more like or less like what I would hear from the listening seat if I stay in the room or leave the room when the test tones are playing.  I usually leave the room.  Plus, in the case of my whacky basement bunker room, the measured response below 200 Hz or so is quite a bit different depending on whether the single door to the room is open or shut.  (The open door vents bass pressure and smooths out the bass bumps and dips a bit.)  Since I usually listen with that door shut, that is how I do the measurements.

      One unexpected problem I ran into was translating the 1/6-octave bar graph of the SynRTA display to Hz figures for purposes of adjusting the midpoint of the parametric bands of the Dual Core.  It was unexpectedly hard for me to locate a table specifying the standard 1/6-octave frequency bands.  The best reference I found was the numerical markers on an old graph REG published using his Liberty Audio suite.  See the frequencies shown on the lower edge of the 1/6-octave measurement of REG's Harbeth M40s on the following page:


      As a reminder, when you do measurements, you normally should only have one channel active.  The more symmetrical your set up (and mine is quite symmetrical) and the more truly centered the measuring microphone is, the further down in frequency you will get at least a bit of higher frequency cancellation between the two channels if you attempt to measure with both channels running.  Even at the 700 Hz frequency, which was one of the main targets of my fine tuning last night, there was a bit of change in measured response between one and both channels operating.  The null at high frequencies was well over 20 dB.

      With the Dual Core, with the press of the bypass button it is easy A/B unequalized versus equalized signals.  Here is a picture of the unequalized response with the speakers in the same position:


      Yes, the equalized version sounds better.  But, honestly, with Alex's recommended positioning, once I get the bass level relatively correct with respect to the upper frequencies, the differences you can see in the graphs I've posted are huge compared to the significance of most of the audible differences--to my ears at least.  

      The largest audible difference is the elimination of the 80 Hz peak.  If I were sitting a few inches higher, this would be reduced without the aid of electronic EQ.  

      The second most audible effect of the EQ is the  taming of what REG calls the shouting of the midrange by pulling down the 700 - 800 Hz peak.  However, I think that at least some listeners might prefer the "shouty" midrange because it thrusts centered vocalists a bit forward in the stage, putting them more in a soloist position and increasing the apparent front-to-back depth of the stage.  Without A/Bing, I didn't notice the "shout" as a problem, whereas I was quite aware of the egregiousness of the 80 Hz peak.  

      Even with A/Bing, yes, the effect of the EQ on the midrange peak is clearly audible and I prefer the flatter response, but it is not at all obvious to me that everyone would. REG is much more sensitive to frequency response problems in this area than I am.  I think many audiophile speakers have emphasized mids for just the reason I mentioned:  enhanced "reach out and touch" the soloist effects, even if the tonality is a bit skewed.  It's all a matter of making the midrange peak large enough the move the soloist forward a bit and enhance the solidity of the image of that soloist, but not adding enough of a peak to make the tonality obviously wrong to most people.




      ---------- Forwarded message ----------
      From: Tom Mallin <tmallin@...>
      Date: Mon, Sep 30, 2013 at 4:22 PM
      Subject: Fwd: [regsaudioforum] TMs measurement
      To: Tom Mallin <tmallin4@...>




      ---------- Forwarded message ----------
      From: Tom Mallin <tmallin4@...>
      Date: Mon, Sep 30, 2013 at 4:20 PM
      Subject: Re: [regsaudioforum] TMs measurement
      To: "regsaudioforum@yahoogroups.com" <regsaudioforum@yahoogroups.com>


      Fair enough.  I'll experiment with the parametric controls of the Dual Core to see if I can improve those areas.

      I did experiment with using the Dual Core's "Advanced" calibration option to extend the range of correction up through 500 Hz, the highest that the automatic correction can go.  That resulted in SynRTA measurements which were even lower in level by about 2 dB in the 100 to 400 Hz region than what this photo shows.  And of course the Advanced calibration did nothing to the 800 Hz region.  I thus reverted to the Typical calibration of the Dual Core.  I believe Tim Ryan of Simplifi Audio recommends the Typical calibration because he likes the results of that algorithm better than what the Dual Core does above the range handled by the Typical calibration.




      On Mon, Sep 30, 2013 at 1:35 PM, <regtas43@...> wrote:
       

      Too midrange-y!

      You need to pull down the 800 Hz region

      and pull up what is below it. (100-400 Hz )

      This is going to shout at you!

      Bass is nice---but this is really not right.

      and it will not sound like what is on the recordings

      either, comes to that. One definitely does not

      want a lump of midrange projection. .

       

      REG




    • regtas43
      Correct. Pitch(as in octaves) runs on a log scale. That is , the distance in octaves between two frequencies f1 and f2 is log (f1/f/2) where log is to the base
      Message 2 of 4 , Oct 3, 2013
      • 0 Attachment

         Correct. Pitch(as in octaves) runs on a log scale.

        That is , the distance in octaves between two

        frequencies f1 and f2 is    log (f1/f/2) where log is

        to the base 2. So if f1 is twice f2 they are one octave part

        if f1 is square root two times f2 they are 1/2 octave apart and so on.

        All this is available on various calculator sites(or on a hand calculator)

        Here is one that turned up easily

        http://logbase2.blogspot.com/2008/08/log-calculator.html

        but there are lots.

        REG



        ---In regsaudioforum@yahoogroups.com, <regsaudioforum@yahoogroups.com> wrote:

        Hi Tom.

        To find the frequency that is 1/6 octave higher than any frequency, multiply by (2^(1/6)), or about 1.1224620483.

        best,

        jeff



        ---In regsaudioforum@yahoogroups.com, <regsaudioforum@yahoogroups.com> wrote:

        I posted another graph after working with the DSPeaker Dual Core's parametric equalization controls last night.  The new picture of the SynRTA-graphed response is dated 9-30-13 and can be found at: 


        This graph was made using the Dual Core's Typical calibration, with bass dip compensation adjusted to plus 6 dB, and with various parametric equalization tweaks after that.  I think I used about 10 different bands of parametric EQ to reach this result.  

        It's a bit tedious to work with the parametric controls since they are not automatic like the Typical or Advanced calibration of the Dual Core or the target curve equalization of the TacT.  But the parametric controls will give you more control with probably less tedium than a 1/3-octave graphic equalizer.  

        The Dual Core screen is not up to fine adjustments like this since it is not big enough, but even if the Dual Core's screen were bigger I don't think using the Dual Core's screen would be as easy as adjusting the response via SynRTA anyway.  The Dual Core's frequency response graphing is not done in real time whereas SynRTA and True RTA provide real time display of the results as you adjust the filters.  You can work much more quickly with a real time display of your adjustment results.

        The process of fine tuning the response is complicated by the fact that I obviously cannot sit in the listening chair to listen to the results of tuning until the process is done since the microphone occupies the spot where my head would be.  Then there is the fact that I have to make a judgment about whether the graphed response will be more like or less like what I would hear from the listening seat if I stay in the room or leave the room when the test tones are playing.  I usually leave the room.  Plus, in the case of my whacky basement bunker room, the measured response below 200 Hz or so is quite a bit different depending on whether the single door to the room is open or shut.  (The open door vents bass pressure and smooths out the bass bumps and dips a bit.)  Since I usually listen with that door shut, that is how I do the measurements.

        One unexpected problem I ran into was translating the 1/6-octave bar graph of the SynRTA display to Hz figures for purposes of adjusting the midpoint of the parametric bands of the Dual Core.  It was unexpectedly hard for me to locate a table specifying the standard 1/6-octave frequency bands.  The best reference I found was the numerical markers on an old graph REG published using his Liberty Audio suite.  See the frequencies shown on the lower edge of the 1/6-octave measurement of REG's Harbeth M40s on the following page:


        As a reminder, when you do measurements, you normally should only have one channel active.  The more symmetrical your set up (and mine is quite symmetrical) and the more truly centered the measuring microphone is, the further down in frequency you will get at least a bit of higher frequency cancellation between the two channels if you attempt to measure with both channels running.  Even at the 700 Hz frequency, which was one of the main targets of my fine tuning last night, there was a bit of change in measured response between one and both channels operating.  The null at high frequencies was well over 20 dB.

        With the Dual Core, with the press of the bypass button it is easy A/B unequalized versus equalized signals.  Here is a picture of the unequalized response with the speakers in the same position:


        Yes, the equalized version sounds better.  But, honestly, with Alex's recommended positioning, once I get the bass level relatively correct with respect to the upper frequencies, the differences you can see in the graphs I've posted are huge compared to the significance of most of the audible differences--to my ears at least.  

        The largest audible difference is the elimination of the 80 Hz peak.  If I were sitting a few inches higher, this would be reduced without the aid of electronic EQ.  

        The second most audible effect of the EQ is the  taming of what REG calls the shouting of the midrange by pulling down the 700 - 800 Hz peak.  However, I think that at least some listeners might prefer the "shouty" midrange because it thrusts centered vocalists a bit forward in the stage, putting them more in a soloist position and increasing the apparent front-to-back depth of the stage.  Without A/Bing, I didn't notice the "shout" as a problem, whereas I was quite aware of the egregiousness of the 80 Hz peak.  

        Even with A/Bing, yes, the effect of the EQ on the midrange peak is clearly audible and I prefer the flatter response, but it is not at all obvious to me that everyone would. REG is much more sensitive to frequency response problems in this area than I am.  I think many audiophile speakers have emphasized mids for just the reason I mentioned:  enhanced "reach out and touch" the soloist effects, even if the tonality is a bit skewed.  It's all a matter of making the midrange peak large enough the move the soloist forward a bit and enhance the solidity of the image of that soloist, but not adding enough of a peak to make the tonality obviously wrong to most people.




        ---------- Forwarded message ----------
        From: Tom Mallin <tmallin@...>
        Date: Mon, Sep 30, 2013 at 4:22 PM
        Subject: Fwd: [regsaudioforum] TMs measurement
        To: Tom Mallin <tmallin4@...>




        ---------- Forwarded message ----------
        From: Tom Mallin <tmallin4@...>
        Date: Mon, Sep 30, 2013 at 4:20 PM
        Subject: Re: [regsaudioforum] TMs measurement
        To: "regsaudioforum@yahoogroups.com" <regsaudioforum@yahoogroups.com>


        Fair enough.  I'll experiment with the parametric controls of the Dual Core to see if I can improve those areas.

        I did experiment with using the Dual Core's "Advanced" calibration option to extend the range of correction up through 500 Hz, the highest that the automatic correction can go.  That resulted in SynRTA measurements which were even lower in level by about 2 dB in the 100 to 400 Hz region than what this photo shows.  And of course the Advanced calibration did nothing to the 800 Hz region.  I thus reverted to the Typical calibration of the Dual Core.  I believe Tim Ryan of Simplifi Audio recommends the Typical calibration because he likes the results of that algorithm better than what the Dual Core does above the range handled by the Typical calibration.




        On Mon, Sep 30, 2013 at 1:35 PM, <regtas43@...> wrote:
         

        Too midrange-y!

        You need to pull down the 800 Hz region

        and pull up what is below it. (100-400 Hz )

        This is going to shout at you!

        Bass is nice---but this is really not right.

        and it will not sound like what is on the recordings

        either, comes to that. One definitely does not

        want a lump of midrange projection. .

         

        REG




      Your message has been successfully submitted and would be delivered to recipients shortly.