Le Fri, 19 Dec 2008 13:22:48 +0200, Sergei Steshenko <steshenko_ser...@list.ru> a écrit :
> On Fri, 19 Dec 2008 11:06:06 +0100 > Matthijs ten Berge <tenberg...@yahoo.com> wrote: > > > Paulo Moura Guedes wrote: > > > BTW, Benchmark DAC1 resamples internally to 110kHz: > > > > > > "On the question of: Why does the DAC1 re-sample to 110 kHz? Here is > > > why: it is the highest frequency to maintain the full oversampling of > > > the D-A chip. EVERY D-to-A chip on the market cuts the oversampling rate > > > in half to accommodate 192 kHz. > > > > > > So these guys are running the DAC at 96 kHz + 14.6%? Sounds to me like > > having an 'overclocked' system, very close to (or maybe even past) the > > limits of the specs. How much headroom (clock-wise) will there be left > > for e.g. temperature fluctuations or component tolerances? > > > > The argument about 192 kHz is true, but why not run the DAC at the > > 'normal' 96 kHz then? I would prefer a 48 -> 96 conversion over a 48 -> > > 110 conversion!. > > And how about interference: won't a 96 -> 110 conversion give any > > interference at 14 kHz? Right in the audible band! > > > > Matthijs > > > > I think these guys running DAC at 110kHz simply do not understand what they > are doing. > > Upsampling makes sense for better analog reproduction, but when playing at > CD sample rate (44.1kHz) they should have upsampled 2x to 88.2kHz - such > upsampling does not introduce IMD. > > Regards, > Sergei. > Upsampling does nothing for analog reproduction because you cannot get more informations that what you get from the DAC output. Upsampling is only a matter of cost: higher the ADC frequency, cheaper is the output filter for approximately the same analog result at the output. Most (all?) cheap DAC convert the PCM linear into Pulse Width Modulation or better into delta-sigma modulation because the output filter become the cheapest of all: a simple RC low-pass filter. See http://en.wikipedia.org/wiki/Digital-to-analog_converter Now, another aspect of a good sounding digital system is the number of bits, 24 bits will sound definitely better than 16 bits, but it must be 24 bits all the way. The S/N ratio as well that the dynamic inclusive the headromm will depend of that number. It is why all musical CD are recorded at the maximum level from the beginning to the end: 16 bits linear PCM is not enough in order to get a good sound when playing pianissimo. It use integer numbers and low level sounds will only use 8 bits or less. Who will a 8 bits sound on a 16 bits system? If you really care about sound quality and want to do serious work, it is best to use jack. The jack daemon use 32 bits float numbers as internal format. You will get a much better dynamic when mixing or applying filters. The conversion to your sound card format (16 or 24 bits integer numbers) will be done at the output of jack. Cheers, Dominique > ------------------------------------------------------------------------------ > _______________________________________________ > Alsa-user mailing list > Alsa-user@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/alsa-user ------------------------------------------------------------------------------ _______________________________________________ Alsa-user mailing list Alsa-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/alsa-user
