Is a USB audio DAC interface needed just for DSD streams?

[hdo]

34 minutes ago, Kal Rubinson said:

What does that mean?

The sound difference is miniscule but is there. To hear the difference, you need very detailed headphones.

[Kal Rubinson]

That's OK, then. It is probably an anomaly.

[Allan F]

7 hours ago, hdo said:

I can clearly hear this. As you know, DSD is represented in a single bit. It cannot change volume level quickly. It has to change bit by bit, where as PCM can go from zero to maximum. Although you hear from loud speakers, you may not notice the difference.

 

This is absolute nonsense. You simply don't know what you are talking about.

[hdo]

1 hour ago, Allan F said:

This is absolute nonsense. You simply don't know what you are talking about.

I know DSDers won't believe whatever truth is.

[hdo]

5 hours ago, Kal Rubinson said:

That's OK, then. It is probably an anomaly

I tested using "Quiet Winter Night" from 2L website, because it provides fair comparison from the same original. You may try yourself with detailed headphones.

[HardrockInMiniMac]

9 hours ago, hdo said:

You cannot compare SACD/DSD to CD/PCM. DSDs are high definition. So it naturally sounds better than CD PCM tracks. If you try the same recordings from DXD converted to DSD vs PCM, you will hear the difference.

You are 100% right with every single one of your posts!

 

Underlying issue is, a lot of people believe, by upsampling PCM to DSD they will get DSD quality "for free".  Funny people. I love it.

 

You are also 100% right that 44,1 etc. should be fed 44,1 to a DAC without any treatment but that - to a large extend - is not the mainstream focus on this forum ;-) because then we would only be hearing about source, cable, DAC, a simple player software on the source. But then there would never be 1Mio posts on HQ Player, Audirvana, upsampling, USB gadgets and all those sponsored discussions. And I love that too.

 

My equipment makes almost all of these tricks technically boring as I do not need to tweak a signal so that something useful comes out of my speakers. Still I own them all for entertainment. I find it funny what PCM upsampling does with Alan Parsons or other "already perfect music".

 

The HQ Player programmer is a huge ABBA fan and that tells me, he has got a working pair of ears for perfect harmonies. That is why HQ Player does the best job when you want to upsample.

 

And then I grab the original vinyl of the music I like and play it through the analogue chain with the same room correction applied and that makes it easy to prove to anyone with ears that analog is still superior when it comes to pure dynamic (given it is not a remastered, loudness levelled version...) - but that is off topic and not compliant with the fundamental assumption of this forum. 

 

Still, this is the most fun HIFI forum and I like the people around here..they are more forgiving and liberal than on most other forums I know.

 

Chris

[Kal Rubinson]

7 hours ago, hdo said:

You may try yourself with detailed headphones.

I never listen to headphones (except on a plane) for many reasons but, primarily since they distort the soundstage.  

[Allan F]

18 hours ago, hdo said:

I know DSDers won't believe whatever truth is.

 

You are confusing your ignorance with truth and substituting your uninformed opinion for facts.

[rayl1234]

On 1/1/2018 at 4:50 PM, hdo said:

I can clearly hear this. As you know, DSD is represented in a single bit. It cannot change volume level quickly. It has to change bit by bit, where as PCM can go from zero to maximum. Although you hear from loud speakers, you may not notice the difference.

 

I am not a DSD person (bec most of my music genres are mixed and mixing is always in PCM so I focus on PCM DACs)... but I have to say I dsagree with this statement. A “valid” PCM signal cannot “change volume quickly” as that would violate the Nyquist sampling theorem (need signal to be bandwidth limited to half sampling rate to get a correct reconstruction)

 

Thus 44.1 kHz sampling rate cannot contain content greater than 22.05kHz. A square wave has infinite frequency content and will be reproduced at some actual frequency and shape determined by the filters in the DAC along with physical limitations.

 

 

[mordante]

IMHO the "best" digital connection is dual AES/ebu like dCS uses.

 

The second best I've heard was if I remember correctly a Aurender streamer feeding a EMM labs DAC using AES/ebu.

 

That doesn't mean I don't like USB. I use a Aries LE and Schiit Multibit gen5 dac with a Ansuz A series USB cable. It sounds good enough for me. 100% of my digital music library is 16/44.1 PCM

[WuNgUn]

8 hours ago, mordante said:

 100% of my digital music library is 16/44.1 PCM

So what is your DAC outputting them to your preamp at? 16/44.1?

Or are you software oversampling to the DAC? 

[clipper]

On 1/2/2018 at 2:52 AM, HardrockInMiniMac said:

You are 100% right with every single one of your posts!

 

Underlying issue is, a lot of people believe, by upsampling PCM to DSD they will get DSD quality "for free".  Funny people. I love it.

 

You are also 100% right that 44,1 etc. should be fed 44,1 to a DAC without any treatment but that - to a large extend - is not the mainstream focus on this forum ;-) because then we would only be hearing about source, cable, DAC, a simple player software on the source. But then there would never be 1Mio posts on HQ Player, Audirvana, upsampling, USB gadgets and all those sponsored discussions. And I love that too.

 

My equipment makes almost all of these tricks technically boring as I do not need to tweak a signal so that something useful comes out of my speakers. Still I own them all for entertainment. I find it funny what PCM upsampling does with Alan Parsons or other "already perfect music".

HardrockInMiniMac and hdo:

 

It's DAC dependent.

 

1) Some DACs sound best with no upsampling.  

2) Some DACs sound best with upsampling.  

3) Some DACs sound best when PCM is converted to DSD (usually the highest DSD rate the DAC can do).

4) Some DACs even sound best when DSD is converted to PCM.

[mordante]

6 hours ago, WuNgUn said:

So what is your DAC outputting them to your preamp at? 16/44.1?

Or are you software oversampling to the DAC? 

 

If there is any resampling it is done by my DAC

[HardrockInMiniMac]

7 hours ago, clipper said:

HardrockInMiniMac and hdo:

 

It's DAC dependent.

 

1) Some DACs sound best with no upsampling.  

2) Some DACs sound best with upsampling.  

3) Some DACs sound best when PCM is converted to DSD (usually the highest DSD rate the DAC can do).

4) Some DACs even sound best when DSD is converted to PCM.

 

Nope.

 

There is a huge difference between "sound best" and "sound right". According to numerous tests most people do not prefer "sound right" and like to add a little tweak here and there which is totally ok. People can do whatever they want with their gear. Young people prefer mp3 over traditional versions of the same song as we all know.

 

Technically, you cannot improve a signal by re-sampling / filtering. At best it stays as "right" as it was before. But still we do. While most DACs will again resample internally (with or without filters) the whole idea of upsampling/filtering and feeding a DAC eventually becomes apples+oranges=banana split or "throwing a coin" if you will. It is basically out of control as your list 1-4 illustrates.

 

Again, no issue as people seem to like the idea of getting some extra quality by chance. Technically a "sound right" by upsampling the master material does not exist but it may well be a "sound best" to someones ears. Fine.

 

 

[rayl1234]

5 hours ago, HardrockInMiniMac said:

 

Nope.

 

There is a huge difference between "sound best" and "sound right". According to numerous tests most people do not prefer "sound right" and like to add a little tweak here and there which is totally ok. People can do whatever they want with their gear. Young people prefer mp3 over traditional versions of the same song as we all know.

 

Technically, you cannot improve a signal by re-sampling / filtering. At best it stays as "right" as it was before. But still we do. While most DACs will again resample internally (with or without filters) the whole idea of upsampling/filtering and feeding a DAC eventually becomes apples+oranges=banana split or "throwing a coin" if you will. It is basically out of control as your list 1-4 illustrates.

 

Again, no issue as people seem to like the idea of getting some extra quality by chance. Technically a "sound right" by upsampling the master material does not exist but it may well be a "sound best" to someones ears. Fine.

 

 

 

While I agree with the general idea, I do not agree with some of the iron-clad statements.

 

I agree that a well designed standalone DAC should work best with native bit perfect input -- after all, that's what I would consider the definition of "well-designed"... else, the DAC will be lacking for certain types of digital inputs, hence, no longer "well-designed."

 

But, it is entirely possible to break the DSP out into a separate component. This might be done to improve noise isolation, etc.

 

One example of this is the Chord Blu Mk 2 and DAVE combo.  The DAVE is already a well designed DAC, but the Blu Mk 2 CD transport/upsampler can be inserted before it (even for USB or S/PDIF audio). Bec both pieces are from the same company, they know what DSP/noise shapers are in the DAVE and can adjust the upsampler in the Blu Mk 2 (and they have their own dual-S/PDIF interface between the two components).

 

But another example might be a very basic DAC box that has an inadequate anti-aliasing filter.  If you upsample (needed to apply a steeper anti-aliasing filter w/o undue latency), you can insert a better anti-aliasing filter in front of it. A "perfect" anti-aliasing filter (i.e. one that "sounds [as] right" as can be) is infinite... so there is no limit except due to computational limits and what the ear can hear. So in a sense, you can always make any existing DAC sound "more" right except we don't have infinite sampling rates, word sizes, computation, etc.

 

As another example -- say you want a perfect accuracy digital volume control and your DAC box doesn't have one.

 

Or -- given that most sigma delta DACs are based on a known small set of chips, it is entirely possible to design DSP in software on a PC in much the same manner that the Blu2/DAVE combo does it.

 

One might argue that  "DAC" should refer to the combo of all DSP/noise-shapers/analog filters instead of what's in the box named DAC....