- I want to get an external sound adapter so I can run an Ensemble RXTX with almost any PC . Any recommendations? I see one by Philips for about $40 runningMessage 1 of 42 , Oct 30, 2011View SourceI want to get an external "sound adapter" so I can run an Ensemble RXTX with "almost any PC".
Any recommendations? I see one by Philips for about $40 running 96kHz.
Al Gerheim - K1QN
- The fact that consumer grade sound cards are crap is not the reason not to use them at 192kHz sampling. If one can look at 192kHz waterfall with a $50Message 42 of 42 , Jan 2View Source
The fact that consumer grade sound cards are crap is not the reason not to use them at 192kHz sampling. If one can look at 192kHz waterfall with a $50 soundcard and see useful signals, this is enough reason to use it.
Look at the explosion of applications and capabilities that Realtek DTV dongles caused. And those are absolute crap 8-bit joke of an ADC.The best soundcard for the Softrock is the one that you have and use. Even stereo is not a hard requirement, there are applications that can live with the image being present.
---In email@example.com, <see_how_deep_da_rabbithole_goes@...> wrote:
Therein lies the difference between Consumer gear and Professional gear.
In Professional recording gear the upper cutoff is usually around 100Khz for the reason that a filter with a cutoff lower, say 40K-50Khz will introduce phase shifts into the 20khz passband and that is especially true for 'brickwall' filters they use for AD-DA converters that inherently have a lot of ripple below the cutoff frequency. Even though you'll technically will have a full 20khz bandwidth the upper octave or two (10khz-20khz or 5khz-20khz) will sound 'smeared' because of phase shifts and ripple in the upper octave (And part of the reason early CD's sounded so harsh) ... For instance look at the specs of the upper grade of consumer power amps and you'll see most have a 100Khz bandwidth for this very reason. It's a type of distortion that doesn't show up in the specs (But can be measured with the right equipment) but a trained ear can hear it and recognize it as a problem where most people will just think it just doesn't sound quite right or experience ear fatigue (Poorly phase aligned speakers can do that also)
Generally as a rule real (And I don't mean Creative products or their E-mu in name only line) professional soundcards for recording work well for SDR while the consumer grade stuff doesn't because it's hard to get right especially in a noisy computer environment so they take the easy way out and filter just above 20khz. Personally I've been using a Delta 44 since before SDR's even hit the scene (I got mine in early 2001 with win98 drivers and software bundle) and I haven't seen much that can beat it nor another manufacturer that provides that kind of driver support for something that came out clear back in 1999 (It took a long time for everyone else to catch up)
Sure sometimes I long for a full 192khz bandwidth display but with that 100Khz limit I know is built into pro gear the performance won't be as good as the Delta at 96khz and likely won't be as good running at 96khz either.
Beware of any cards that just quote the theoretical dynamic range of the AD and DA converters in lieu of actually professional measurements of the finished product (Creative is infamous for this scam but so are most others) Also check on how they upgrade drivers when Windows changes, many don't support anything over a couple of years old (Again Creative is infamous for this) A lot of computer soundgear specs are as shaky as autosound specs for instance you can't have a Dolby 5.1 output and only provide outputs for 2 speakers ... I don't care if you run some Dolby licensed software it's still just a stereo output but the market is flooded with scams like this
It should also be noted that the reason for higher sample rates isn't to increase the bandwidth but to increase the number of samples in the upper couple of octaves which is why none of them go out past 100khz in bandwidth necessitating a specialized converter card like a couple of the ones talked about here that are specifically for SDR if you want to make 192khz actually be beneficial for more than just an enhanced display
Another thing I don';t hear anyone mention when talking about 192khz bandwidth is changing the QSD caps which can also cause 'sag' in the upper and lower part of the display. Read Dan Tayloe's work for how the values of the 'sampling caps' effects the frequency response as well as the dynamic range of a QSD
---In firstname.lastname@example.org, <mark@...> wrote:In the days of the early crystal controlled Softrock radios, the
bandwidth of your sound card was directly related to the amount of
band coverage your receiver got. If you are using one of the frequency
agile Ensemble radios with an Si750, that is no longer true because
the LO can move; you get full band coverage at any sound card sample
What higher sample rates get you is wider simultaneous band coverage.
You can see more of the band in the waterfall display at 96KHz (at
least 80KHz, 96 is the theoretical limit) than at 48KHz (40-48), and
even more at 192KHz (160-192). How close to the theoretical limit you
can get depends on the antialiasing filters in the sound card. In
theory, software can decode more than one signal within the passband
simultaneously. (With existing software you would either have to run
multiple copies at once, which would require a really powerful system,
or route the output of the Softrock to more than one computer. It has
been done.) You will also need 192KHz sampling if you want to
accurately decode wideband FM signals (wide like broadcast FM or
analog TV audio).
To get full advantage of 192KHz sampling, you need a sound card that
works WELL at that sample rate; not every device that offers it
qualifies. Two failings are common: analog filtering that cuts off a
little above 20KHz, or a noise floor that rises rapidly above 20KHz.
(The M-Audio Audiophile 192, which would seem on the surface to be a
more desirable device than the popular Delta 44, suffers from that
problem.) The rapidly rising noise is often caused by noise shaping, a
technique that lowers noise in the audio passband (0-20KHz) at the
cost of higher noise beyond that frequency; it works well for audio
but is counterproductive in SDR applications.
The SDR Widget (mentioned earlier in this thread) is an example of a
sound interface that DOES offer excellent 192KHz performance. It uses
high quality converters that do not employ noise shaping, so the noise
floor stays low all the way up. But you have to build it (when
available at all it is only available as a kit) and it's a fairly
challenging dense surface mount build.