adamdea;699192 Wrote: 
> The conclusion so far is that john Swenson considers that digital
> filters don't sound good unless they are on separate FPGAs -he doesn't
> like the ones on the dac ships themselves (also IIRC some time ago he
> said that the ones on the Sabre Dac are OK.)
> 
> AS I understand it conventional engineering thinking is that a well
> designed DAC is not source dependent. Well designed does not mean
> expensive.
> 
> Curiously the ones which might be source dependent are those designed
> for/by audiophiles. At that point, I suggest that you draw your own
> conclusions. 
> 
> [by way of a side turning i have been meaning for a while to see
> whether I could test JS's hypothesis by implementing a sinc like
> (linear phase, at least 80 db attenuation at nyquist) filter in Sox
> which would mean that the half band filter in my MF DAC did very little
> (there being no frequencies for it to attenuate- although I don;t
> suppose i could really cut it off entirely without having -80db at
> 20kHz; but presumably a transition band 20-22kHz would mainly take it
> out of the equation). I assume that filtering in Sox  should work at
> least as well as the separate FPGA.
> 
> I have never got round to working out how to do this  though especially
> given the upsampling to 96Khz which sox does at present courtesy of
> phil's setting for inguz.]

I would not say that the only DACs that are source sensitive are ones
designed by audiophiles. Any DAC that uses one of the off the shelf
S/PDIF receivers and directly feeds the data and clock into a DAC chip
is going to be quite sensitive to the outside world. There are a LOT of
those out there and many are still being built that way. 

A rough estimate is that maybe half of the DACS being made today are
using ASRC chips, this helps significantly but does not completely
eliminate outside influence. There are other factors involved such as
ground plane noise and the local clock itself. (the groundplane noise
frequently modulates the local clock)

The sabre chips are interesting, they seem to be the only ones that
have fairly decent digital filters (not perfect but quite good), BUT
they have a builtin ASRC which causes the same sort of issue I have
been discussing with ASRCs. There is one DAC maker I know of who has
figured out how to bypass the internal ASRC and uses his own very low
jitter circuit so the ASRC is not needed, this DAC sounds really good
but is very expensive. 

Using external software (SOX etc) to perform the filtering is a very
viable alternative, as long as the software uses enough precision in
its internal computations. A 32 bit float is NOT enough precision, a
double (64 bit float) IS sufficient. But after you do this you don't
want it going into through the same digtial filter built into the DAC
chip. Fortunately many chips use different filtering at 176/192 than
they do at lower sample rates, some don't do any filtering at the
highest sample rates. So if you use software to upsample to 176/192 you
might actually have a good chance of bypassing many of the problems with
the 44.1 filters in most DAC chips. 

This by the way is why I think high sample rate files are popular with
some people, its not that the file has more high frequency information,
but that the DAC is using a better sounding filter at that sample rate.


John S.


-- 
JohnSwenson
------------------------------------------------------------------------
JohnSwenson's Profile: http://forums.slimdevices.com/member.php?userid=5974
View this thread: http://forums.slimdevices.com/showthread.php?t=94352

_______________________________________________
audiophiles mailing list
[email protected]
http://lists.slimdevices.com/mailman/listinfo/audiophiles

Reply via email to