[HeadLover]

I see that some sound cards (and DACs) have 192KHz, and some have 96KHZ and some even have "only" 48KHZ
So
Is it really matter to the final output quality?
I mean, all CD audio is 44KHZ in the end, so how much up sampling or the fact that a DAC on the soundcard have more or less make it better ?

 

[Hardcoreckn]

I say it generally does not matter.

Nyquist theorem states that to reproduce the original signal a sampling of 2x the highest frequency is required. In this case we are dealing with audible frequencies which has an upper peak at 22kHz, which is why the 44kHz is the industry standard.

HOWEVER, there is an advantage to sampling at a higher frequency for audio, it's more forgiving of having poor filters used when recording.

 

[HeadLover]

Quote:

Originally Posted by Hardcoreckn /img/forum/go_quote.gif I say it generally does not matter. Nyquist theorem states that to reproduce the original signal a sampling of 2x the highest frequency is required. In this case we are dealing with audible frequencies which has an upper peak at 22kHz, which is why the 44kHz is the industry standard. HOWEVER, there is an advantage to sampling at a higher frequency for audio, it's more forgiving of having poor filters used when recording.

so
why do so much of the DACS today upsampling it ?
Why not putting just a better filters ?

 

[monolith]

Probably because people see high numbers and think it must mean it sounds better. It's easier to market.

 

[Toe]

You're right in that if all you're playing is 16/44 audio as is, there's not much need for a card/chip/whatever that supports higher resolution. A DAC that supports higher resolutions may or may not be better at 16/44 audio than one that only supports 16/44.

As for resampling to a higher rate, it's debatable. Resampling can certainly degrade the sound quality if done poorly. But the question of whether it can improve sound quality if done well, there's people on both sides of the argument.

 

[ericj]

DACs generally work better and produce less distortion when they're run at a higher rate, which is why some DACs like the ESS Sabre parts that are showing up lately convert the sample rate to a very very high rate.

However, upsampling the signal before it gets to the DAC processing stage can easily degrade the final SQ. It's best of you just get a good DAC and deliver the stream to it as unmolested as you can.

 

[dspargo]

Quote:

Originally Posted by ericj /img/forum/go_quote.gif DACs generally work better and produce less distortion when they're run at a higher rate, which is why some DACs like the ESS Sabre parts that are showing up lately convert the sample rate to a very very high rate. However, upsampling the signal before it gets to the DAC processing stage can easily degrade the final SQ. It's best of you just get a good DAC and deliver the stream to it as unmolested as you can.

I use J. River's Media Center and under the DSP Studio and the Output Format tab I have 24 bit and 96K selected. Do you recommend I not use this Output Format tab and let the data flow clean to my Benchmark DAC?

 

[ericj]

Quote:

Originally Posted by dspargo /img/forum/go_quote.gif I use J. River's Media Center and under the DSP Studio and the Output Format tab I have 24 bit and 96K selected. Do you recommend I not use this Output Format tab and let the data flow clean to my Benchmark DAC?

In general, yeah. But i don't know much about your DAC.

the 24-bit thing, well, if your DAC is really 24-bit, padding your 16-bit audio with 8 extra bits of 0's might allow the DAC to perform better - if it wasn't already padding it's input anyway.

If the DAC does any of it's own oversampling or other sample rate conversion, it may perform better if you don't mess with the sample rate yourself.

In order to do it's job best, the part of a dac that does oversampling and noise shaping has to know what has already happened to the signal. If it's expecting a virgin 96khz stream and it gets a stream that was converted from 44.1khz to 96khz, it will perform worse than it would have with the virgin stream. Whether or not that is worse than how it would have performed with a virgin 44.1khz stream is where we find the word "potentially".

Converting from 44.1khz to 96khz does far less damage than converting from 44.1khz to 48khz, anyhow.

 

[Nocturnal310]

well...can anyone of u make out 44.1khz from 48khz?

 

[Redo]

Quote:

Originally Posted by Nocturnal310 /img/forum/go_quote.gif well...can anyone of u make out 44.1khz from 48khz?

In what way? A poorly resampled 44.1khz to 48khz audio stream can have a noticeable impact. This was a big problem with most Creative sound cards except for the X-Fi.

In terms of a well mastered Reel to Reel tape recorded to 44.1khz, and then again recorded to 48khz. Trying to tell the difference between those two will be all but impossible (again, so long as no other filtering or other variables are entered into the equation).

 

[OutdoorXplorer]

I understand the RSA Predator's DAC is 16/48, correct me if I am wrond and it sounds excellent.