Why does the soundstage sound different (often better IMHO) in high rate DSD like DSD256 Vs native Redbook to a DAC with a Chip that upsamples to ultimately do SDM conversion.

[Shadorne]

15 hours ago, Miska said:

 

My point is that the difference doesn't solely attribute to any such single thing. Just one of the many factors contributing to the end result.

 

I agree. However, you might agree that of these many differences, some aspects might be more audible than others. That is my interest and query here - what is the major audible problem with nearly all chip-based DACs such that upsampling to high rate DSD on a computer is generally & anecdotally regarded to have better soundstage. This soundstage difference suggests TIMING errors. I honestly believe (perhaps wrongly) that some aspects of processing are audibly more critical in regard to soundstage than others.

 

Both @Miska @bogi and several others, if l paraphrase correctly, believe there is something wrong enough to be audibly detrimental with the quality of the processing in chip-based SDM DACs, and pc-based processing eliminates this problem. I FULLY AGREE - this is unproven but it seems 99% likely to be the issue.

 

However, can we dig deeper?  

 

My experience with a multitude of speakers, finds a great similarity between the improvement of PC-based high quality upsampling and eliminating baffle edge diffraction on a speaker. Two things sound similar - I therefore wonder if there is a common cause (an edge diffraction on a speaker or an echo of the signal share many similarities). This is my particular reasoning for suspecting a pre-echo problem with chip-based DAC converters.

 

I welcome other suggestions.

[Miska]

Type of the digital filter, digital filter ratio and modulator design...

 

[fas42]

On 9/3/2023 at 9:01 AM, Miska said:

 

No, there's no lack of knowledge.

 

 

IOW, if you were only able to extract a set of measurements, and never actually hear, listen to the playback, you could confidently predict whether you could live with the SQ, or not - yes?

[fas42]

10 hours ago, Shadorne said:

 

Both @Miska @bogi and several others, if l paraphrase correctly, believe there is something wrong enough to be audibly detrimental with the quality of the processing in chip-based SDM DACs, and pc-based processing eliminates this problem. I FULLY AGREE - this is unproven but it seems 99% likely to be the issue.

 

...

 

I welcome other suggestions.

 

My experience is that human hearing can accommodate all sorts of technical deficiencies of various sorts - but a combination of too many issues becomes too much ... the ability of the ear/brain to process the sound it's hearing, and handle the necessary unraveling is overloaded, and it "sounds bad!". There is no single anomaly in itself which means a thumbs up, or a thumbs down - two rigs which have totally different sets of 'problems' could both be entirely acceptable; or alternatively, could both be disturbing to listen to - it's the combinations which matter, not the individual 'shortcomings', in isolation.

[Miska]

7 hours ago, fas42 said:

IOW, if you were only able to extract a set of measurements, and never actually hear, listen to the playback, you could confidently predict whether you could live with the SQ, or not - yes?

 

You were talking about noise and interference.

 

[Shadorne]

18 hours ago, Miska said:

Type of the digital filter, digital filter ratio and modulator design...

 

This is what I would call scratching the surface - it is no deeper than what @bogi stated. Of course, I do understand that if you do actually know what are the more important factors in improving the soundstage then you might be reticent to share that. It would be like a proprietary trade secret to HQplayer and the longer it remains so, the longer your competitive advantage exists over regular chip-based processing commonly found in DACs. In fact, to protect trade secrets one might even obfuscate or redirect purposefully. Although I sense that has not been the case here, and I really appreciate your input given that your depth of knowledge is far greater than myself and many others.

[bogi]

Of course, Miska will not explain how he does source content analysis and the details of filter/modulator implementation.

I really don't expect any soundstage specific processing in HQPlayer. I also don't expect a simple filter or modularor parameter affecting soundstage in some specially significant extent. IMO better soundstage simply comes out of more accurate reconstruction of analog signal from source digital content.

 

Soundstage is created in our brains. I agree with the content @fas42 posted above. If the level of signal reconstruction accuracy is enough for our brain to allow to create better image of the scene, our subconsciousness will construct it automatically. I think quite generally, not only in audio area, we are underestimating the function of subconsciousness, all the automation it does for us to live easier. I support the idea fas42 brought into this topic about 'filling the gaps' by our subconsciousness.

[davide256]

3 hours ago, bogi said:

Of course, Miska will not explain how he does source content analysis and the details of filter/modulator implementation.

I really don't expect any soundstage specific processing in HQPlayer. I also don't expect a simple filter or modularor parameter affecting soundstage in some specially significant extent. IMO better soundstage simply comes out of more accurate reconstruction of analog signal from source digital content.

 

Soundstage is created in our brains. I agree with the content @fas42 posted above. If the level of signal reconstruction accuracy is enough for our brain to allow to create better image of the scene, our subconsciousness will construct it automatically. I think quite generally, not only in audio area, we are underestimating the function of subconsciousness, all the automation it does for us to live easier. I support the idea fas42 brought into this topic about 'filling the gaps' by our subconsciousness.

Sound stage is mostly a function of treble and treble overtones of lower frequencies. Speakers with fast treble drivers will sound stage better than slower speakers. Ergo if the solution you employ is fast, undistorted by resonances and accurate for dynamics it can sound stage well. As to the OP's original question, it remains

a no brainer that DSD up-sampling can sound stage better by pushing D-A  filter artifacts outside the range of human hearing, something which PCM up-sampling never seems to accomplish. Nit picking about much smaller order phenomena doesn't get us anywhere unless one can empirically demonstrate that the error

is humanly detectable

[fas42]

15 hours ago, Miska said:

 

You were talking about noise and interference.

 

 

The nub is this, that which I said earlier,

 

Quote

I don't think I have read anywhere where someone has done research on the resistance of a complete audio chain to noise/interference, as regards the impact on subjective SQ ...

 

Can you measure this, or not?

[fas42]

3 hours ago, davide256 said:

Sound stage is mostly a function of treble and treble overtones of lower frequencies. Speakers with fast treble drivers will sound stage better than slower speakers. Ergo if the solution you employ is fast, undistorted by resonances and accurate for dynamics it can sound stage well. As to the OP's original question, it remains

a no brainer that DSD up-sampling can sound stage better by pushing D-A  filter artifacts outside the range of human hearing, something which PCM up-sampling never seems to accomplish. Nit picking about much smaller order phenomena doesn't get us anywhere unless one can empirically demonstrate that the error

is humanly detectable

 

The nature of the speaker may help with soundstage perception, depending upon anomalies elsewhere - but it's only a very small player in the game. A sorted reproduction chain with a very low cost, normal speaker will present a soundstage vastly superior to that of an exotic speaker costing 100's of times more, if the latter is driven by a substandard set of electronics ... it's the overall integrity that determines the quality of the listening experience; not the 'quality' of individual components.

 

Once one has a handle on what is going on, it's quite easy to make the soundstage collapse, by simply 'damaging', degrading some link in the chain, anywhere in the path.

[Miska]

7 hours ago, fas42 said:

Can you measure this, or not?

 

Yes I can, it is not a problem. Noise/interference is pretty straightforward to measure.

 

[Archimago]

Interesting discussions guys at 3 pages already...

 

I'm wondering though, has there ever been a documented listening test with controls in place (eg. same output level) showing that there has been a change in soundstage for the "better" as assessed by multiple listeners, between straight PCM playback and the same data upsampled to DSD256? With which DAC? With what upsampling software and settings?

 

Personally, I feel that once I control the volume, upsampling to DSD256 (with flat frequency response) sounds more like just straight high-quality PCM playback. DSD64 with its elevated noise seems to add a subjective sense of "space" for me more than DSD128 or DSD256.

 

[STC]

9 hours ago, Archimago said:

I'm wondering though, has there ever been a documented listening test with controls in place (eg. same output level) showing that there has been a change in soundstage for the "better" as assessed by multiple listeners,

There was paper published by AES. [At certain frequencies hires seemed to give different soundstage but the opposite for others. I didn’t care much to dwell deeper into the paper as soundstage changes due to frequency is already known as you can sense with different speakers. ] items in the bracket figment of imagination. 
 

Corrected.  AES paper.  This on localization narrow vs wide. 
 

https://www.aes.org/e-lib/browse.cfm?elib=17737

[Miska]

On 10/14/2023 at 9:39 PM, Archimago said:

Personally, I feel that once I control the volume, upsampling to DSD256 (with flat frequency response) sounds more like just straight high-quality PCM playback. DSD64 with its elevated noise seems to add a subjective sense of "space" for me more than DSD128 or DSD256.

 

Do you use a DAC that converts both PCM and DSD to analog bit-perfect?

 

Have you compared with various different DAC architectures and compared how they behave at different rates?

 

I personally wouldn't bother to feed DAC at DSD64 rate.

 

With ESS PRO chip models you have massive DAC-to-DAC variations depending on the I/V stage design and how the 4+4 channel summing is done. Plus the analog filter design, plus how the DAC chip is clocked/configured.

 

[fas42]

All this sort of thing occurs because digital reproduction is very, very prone to altering the subjective clarity of low level detail in a recording - the slightest variation in "how it's done" impacts that clarity, which means there is this endless debate about the 'right way'.

 

It would be nice if human hearing wasn't so sensitive to this stuff ... but, it is. So, "pick your poison" - really accurate reproduction is possible, but requires more than the usual fiddling around with the technical settings of the design ...

[bogi]

On 10/14/2023 at 8:39 PM, Archimago said:

Personally, I feel that once I control the volume, upsampling to DSD256 (with flat frequency response) sounds more like just straight high-quality PCM playback. DSD64 with its elevated noise seems to add a subjective sense of "space" for me more than DSD128 or DSD256.

I personally prefer PCM to DSD upsampling already about 9 years. I started to use it initially with foobar2000, since it made sound more natural for my ears. With HQPlayer I perceive improved sound quality and slightly different sound with every filter or modulator. Both sound quality improvement and the ability to tune sound according to my personal preferences are the main points for me why I am using HQPlayer.
 

On 10/14/2023 at 8:39 PM, Archimago said:

I'm wondering though, has there ever been a documented listening test with controls in place (eg. same output level) showing that there has been a change in soundstage for the "better" as assessed by multiple listeners, between straight PCM playback and the same data upsampled to DSD256? With which DAC? With what upsampling software and settings?

Like I mentioned above in some post, soundstage is created in our brains. We are clearly coming into subjective waters here. Then we need to assume a listener to be part of a listening chain and attribute listening test results to a specific listener. IMO no generalization of individual listening test results is appropriate. I also cannot associate such a listening test results to word 'objective'.

It is nice if a group of listeners agrees on some commonly perceived listening impressions. But they are no more true than impressions of other persons whose listening experience is different. Listener group selection can also affect results.

 

Over the years I found so much demands on double blind listening tests in endless objective vs subjective discussions at many forums. But those discussions are mostly abstracting from what is happening in listeners head. It may be interesting to read about somebody else's experience, but then I take it only as an interesting reading and not as something what is or should be relevant to me as a person and to my listening chain. We are simply different persons with different hearing abilities, taste and preferences about what is good sound. And we are using different devices, software and their settings. Too much variables and too many individual preferences.

 

There are many aspects of sound quality, one of them is soundstage. People can easily disagree on anything sound quality related, including which aspects of sound quality are more important etc. Hunting for objectiveness does not bring sense to me.

[Archimago]

20 hours ago, Miska said:

 

Do you use a DAC that converts both PCM and DSD to analog bit-perfect?

 

Have you compared with various different DAC architectures and compared how they behave at different rates?

 

I personally wouldn't bother to feed DAC at DSD64 rate.

 

With ESS PRO chip models you have massive DAC-to-DAC variations depending on the I/V stage design and how the 4+4 channel summing is done. Plus the analog filter design, plus how the DAC chip is clocked/configured.

 

I've had listening tests with many AKM DACs using DSD-Direct over the years as well beyond the ESS chips.

 

Yeah, I prefer not to listen to DSD64 either... Just an observation that subjectively it made more difference than DSD256 and sometimes I might even prefer that variation in sound quality even though I know technically it's limited.

[Archimago]

On 10/14/2023 at 8:30 PM, STC said:

There was paper published by AES. [At certain frequencies hires seemed to give different soundstage but the opposite for others. I didn’t care much to dwell deeper into the paper as soundstage changes due to frequency is already known as you can sense with different speakers. ] items in the bracket figment of imagination. 
 

Corrected.  AES paper.  This on localization narrow vs wide. 
 

https://www.aes.org/e-lib/browse.cfm?elib=17737

Interesting paper STC.

 

I dunno if this would apply to (for all intensive purposes), lossless, possibly hi-res transcoding between PCM --> DSD though. A quick look at the paper is suggesting changes in soundstage perception using MP3-encoded (LAME) content at 320/128/80/56/48kbps bitrates! As expected, accuracy went down with higher compression:

We're talking about much greater accuracy obviously.
 

[Miska]

57 minutes ago, Archimago said:

I've had listening tests with many AKM DACs using DSD-Direct over the years as well beyond the ESS chips.

 

These are naturally not native PCM, so you more like compare modulators (if you used same digital filter for both).

 

57 minutes ago, Archimago said:

Yeah, I prefer not to listen to DSD64 either... Just an observation that subjectively it made more difference than DSD256

 

For example the new AK4191 + AK4499EX chip combo always runs either at 128fs or 256fs conversion rate. With careful selection of settings this combo can also do great DSD Direct conversion at DSD128 or DSD256. Then also the modulator rates can match.

 

[fas42]

19 hours ago, bogi said:

Like I mentioned above in some post, soundstage is created in our brains. We are clearly coming into subjective waters here. Then we need to assume a listener to be part of a listening chain and attribute listening test results to a specific listener. IMO no generalization of individual listening test results is appropriate. I also cannot associate such a listening test results to word 'objective'.

 

 

My experience with 'soundstage' has always correlated with what others around me have experienced. Unless they were fully paid up 'audiophiles', . The latter have at times reacted positively to very obviously defective reproduction; and showed lack of interest, boredom with 'natural', realistic replay ... IME .

 

Good sound stage means that it's easy to relate what you're hearing to sounds made by musicians, instruments - it throws up a convincing illusion which in the best instances is impossible to "hear past". If you're always aware that you are listening to some sort of imitation, having a certain degree of "paleness", then you experiencing below what's possible soundstaging.

 

19 hours ago, bogi said:

There are many aspects of sound quality, one of them is soundstage. People can easily disagree on anything sound quality related, including which aspects of sound quality are more important etc. Hunting for objectiveness does not bring sense to me.

 

Good sound quality delivers everything. Including soundstage. The biggest problem is the general poor quality of measurement methods, which are in the league of the Blind Men and an Elephant parable - small parts of the story can be determined with great accuracy, but the gestalt, which is so obvious at a subjective level, is MIA ...

[bogi]

5 hours ago, fas42 said:

Good sound quality delivers everything. Including soundstage. The biggest problem is the general poor quality of measurement methods, which are in the league of the Blind Men and an Elephant parable - small parts of the story can be determined with great accuracy, but the gestalt, which is so obvious at a subjective level, is MIA ...

 

If there would be some software available which analyses measurement results and gives something like "soundstage quality index", then such a measurement results could be called "objective" - given we know and understand how the index was computed. The result would more or less reflect what most of people is perceiving, depending on the analysis method, but at least it would provide the same result in repeated measurements.

 

AFAIK we are very far from having a trustworthy software available giving us some higher level sound quality information.

 

As soon as we substitute measurement gear with group of people, then we are "measuring" opinions of those people and not the listening chain playing some audio content  - since the thing which changes are people listening to the same audio content using the same listening chain. It's like repeating measurements with different measurement gear of totally random quality including buggy ones, calibrated and not calibrated ones etc. and then attempting to "evaluate" soundstage relevant outcome of such repeated measurements in some way.