 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
146.115.126.203
In Reply to: RE: Digital Volume Control - a different perspective posted by soundchekk on January 26, 2008 at 06:28:34
Looks like I've been guilty of scanning this thread too quickly. :-)
I went back and read more carefully what you have been saying.
Your solution of carefully matching the gain of your system so that the required attenuation
is mild is very compelling.
Follow Ups:
Digital VC when calculated correctly using double precision is perfect (resulting rounding error converting to integer is insignificant). I've experienced poor quality chip based digital VCs as well as bad software and can only attribute its poor performance to bad math. Foobar's VC is 64 bit as well as Soundchekk's brutefir using logarithmic formula.
If you strictly use 16 bits then digital VC will not work as DACs inability to handle least significant bits will be detrimental. Its better to use analogue VC as it attenuates full signal as processed by the DAC.
Increasing bit depth to 24 bits goes a long way to reduce dependency on lower order bits. Still, high resolution DACs is a must with digital VC's.
If DAC has 120db SNR giving 20 bits true resolution - lower 4 bits can't be processed. Digital VC at -24db implies first 4 high order bits are 0 (basic binary math). This leaves you with 16 bits resolution. Resulting error from DAC's inability to handle lowest 4 bits translates to a worst case 0.000252db error - I can't hear that! At lower signal amplitudes, this error increases: a signal at -55db followed by -24db attenuation (giving -79db) has worst case 0.142db error!
Burr Brown's 1792 DAC offers 127db SNR at 2V reducing lower order bit errors even further (above errors are more than halved).
I have an E-MU 0404 USB DAC rated at 117 db SNR. However online measurements using RMAA found a SNR of only 111 db while recording the analog output with a "mastering grade" professional audio interface (Lynx Aurora 8).
I'm sure the factory rating is correct, but probably in perfect conditions. In real life it seems you lose 6 db because of op amp/cables or something else (including the possibility of a bad test setup).
So I think it's a good idea to subtract 6 db from rated SNR when considering the "real-world" SNR.
E-MU 0404 USB ADC (BB 1820) SNR is 113db and DAC (AKM 4395) is 117db SNR. Your result of 111db reflects E-MU's ADC performance. (Its same as for PCI 0404 card). Yes there will be other losses which would apply across the board but I wouldn't use 6db as a guide - way too much. You should get close to its rated 113db performance with better power supply etc.. On EMU website they further qualify the 113 & 117 SNR performance.
Also, using digital VC gives no phase errors as pointed out by Soundchekk.
Hi cics.
The phase errors I mentioned are introduced by connectors, inductors
capacitors, soldering joints and whatsoever is sitting on the path if
you use external devices such as pre-amps and so forth.
Even my Stevens & Billington TX102 passive transformer pot I used to use slowed down transients a lot.
Now I got a 2 inch silverwire hooked on my DAC tower right at the chip which is running right into my AMP.
Cheers
| FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
||||||||||||||
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
