Is there any budget DAC for DSD1024?

[sunjam]

It's 2024 now. DSD512 DAC has become mainstream consumer products. Budget DACs (less than $100) can support it easily.

How about DSD1024? It seems that only few makers are having products that support it? Is it because no ESS/AKM/CS chip that can support it?

In the past, DSD512 was not that important as there were only few content that are in that format. However, with the readily available of cheap CPU/GPU power together with professional real-time upsampling software (e.g. HQPlayer) for the general public, tons of quality DSD512 (or even DSD1024) contents are availble for us to enjoy (as we can just upsampling the CD quality source to DSD ourselves).

I am looking for a DAC that can support DSD1024 but seems that it would cost you a lot as of today (or did I miss any budget option?).

Btw, I highly suggest to try out HQPlayer for upsampling. It is really amazing. (Note: DSDNative uses it for upsampling their DSD files too for release, https://www.nativedsd.com/news/the-higher-rates-program/)

 

[dlovesmusic]

It's 2024 now. DSD512 DAC has become mainstream consumer products. Budget DACs (less than $100) can support it easily.

How about DSD1024? It seems that only few makers are having products that support it? Is it because no ESS/AKM/CS chip that can support it?

In the past, DSD512 was not that important as there were only few content that are in that format. However, with the readily available of cheap CPU/GPU power together with professional real-time upsampling software (e.g. HQPlayer) for the general public, tons of quality DSD512 (or even DSD1024) contents are availble for us to enjoy (as we can just upsampling the CD quality source to DSD ourselves).

I am looking for a DAC that can support DSD1024 but seems that it would cost you a lot as of today (or did I miss any budget option?).

Btw, I highly suggest to try out HQPlayer for upsampling. It is really amazing. (Note: DSDNative uses it for upsampling their DSD files too for release, https://www.nativedsd.com/news/the-higher-rates-program/)

Except when you get to dsd1024, it measures significantly worse in audible band noise floor compared to dsd512 and even more so to dsd256. At least at this stage there is no benefits at all going up to dsd1024.

My Holo Cyan 2 supports up to dsd1024 but it sounds significantly worse than dsd256 that I wouldn’t bother to use it…

Almost all other measurements for dacs that support dsd1024 like T+A dac200, Holo Dacs… show similar results that there is simply no benefits at all going up to dsd1024 so it’s pretty much a waste of computer resource

Even the head of R&D at T+A recommends feeding dsd256 to the dac200

 

[JSmith]

upsampling the CD quality source to DSD

What does this achieve?


JSmith

 

[Sokel]

What does this achieve?


JSmith

On Khadas Tone board,no ESS hump

(and some others as people suggest)

 

[Sokel]

Except when you get to dsd1024, it measures significantly worse in audible band noise floor compared to dsd512 and even more so to dsd256. At least at this stage there is no benefits at all going up to dsd1024.

My Holo Cyan 2 supports up to dsd1024 but it sounds significantly worse than dsd256 that I wouldn’t bother to use it…

Almost all other measurements for dacs that support dsd1024 like T+A dac200, Holo Dacs… show similar results that there is simply no benefits at all going up to dsd1024 so it’s pretty much a waste of computer resource

Even the head of R&D at T+A recommends feeding dsd256 to the dac200

Here's a comparison where you can see in and out of band noise.
DSD512 is smoother of course out of band but it does take a hit compared to DSD64 in overall noise:

 

[sunjam]

Here's a comparison where you can see in and out of band noise.
DSD512 is smoother of course out of band but it does take a hit compared to DSD64 in overall noise:

View attachment 359241

Hmm... I thought the noise level should be similar (or at least not worse) if the rate goes up. Would that be related to a particulart DAC's implmentation? Or does it something related to the nature of DSD coding?

From the DXD doc provided by Merging Technologies ( https://www.merging.com/uploads/assets/Merging_pdfs/dxd_Resolution_v3.5.pdf ), it suggests that the noise levels are similar (shown on Figure 3 in the doc and I'd attached it here). It is too bad that the figure is quite blur....

 

[antcollinet]

What does this achieve?


JSmith

Utterly pointless.

 

[sunjam]

Except when you get to dsd1024, it measures significantly worse in audible band noise floor compared to dsd512 and even more so to dsd256. At least at this stage there is no benefits at all going up to dsd1024.

My Holo Cyan 2 supports up to dsd1024 but it sounds significantly worse than dsd256 that I wouldn’t bother to use it…

Almost all other measurements for dacs that support dsd1024 like T+A dac200, Holo Dacs… show similar results that there is simply no benefits at all going up to dsd1024 so it’s pretty much a waste of computer resource

Even the head of R&D at T+A recommends feeding dsd256 to the dac200

How about DSD512? My DAC can support up to DSD256. I am currently upsampling all my contents to DSD256 with HQPlayer.

Do you think it is worth to spend on a DAC in order to get DSD512 (for HQPlayer)? Or it would just be similar to DSD256?

The DAC I have are ESS DAC (ESS9018K2M), CS DAC (43131). My understanding is that they do not support direct DSD mode. If I get a AKM DAC that supports DSD512 and DSD direct, would that make a difference when compared with what I have right now?

 

[antcollinet]

I am currently upsampling all my contents to DSD256 with HQPlayer

There is no point to this.

Converting anything to DSD provides no audible benefit.

 

[sunjam]

What does this achieve?


JSmith

From DSDNative web site:

"What is different with higher bitrate DSD releases is the uncorrelated modulation noise content placement in the frequency spectrum. While the spectral shape is the same regardless of the DSD bitrate, its effective start and end points move an octave higher for every doubling of the DSD bitrate. For DSD 64, the uncorrelated modulation noise is about -110dB at 20KHz, rising to about -50dB at 100KHz. For DSD 512, the modulation noise is about -110dB at 160KHz, and -50dB at 800KHz. Higher bitrate DSD releases (including DSD 128, DSD 256, and DSD 512) facilitate the listener’s Digital to Analog Converter (DAC) operation utilizing gentler, less phase distorting reconstruction filters and algorithms in the Digital to Analog process. The result is music with far improved phase response, a sound that is less aggressive, more spacious and airier, and less “digital”.

It is our experience that listening to an album from NativeDSD at the highest DSD bitrate that your DAC supports is the best sounding. NativeDSD provides these purchasing choices to our listeners to bring you the best musical experience that you can experience with today’s recording technologies. We hope you continue to enjoy these releases and look forward to your thoughts and continuing support."

You can download and try HQPlayer free here (https://www.signalyst.com/consumer.html). It is very different from the normal upsampling programs. For me, it opens a new door for me to enjoy my exisitng contents. The reasons I want DSD512/1024 DAC is to push it to the limit

 

[sunjam]

There is no point to this.

Converting anything to DSD provides no audible benefit.

Just wondering did you try HQPlayer and upsample your content to DSD256? If not, I highly suggest you to give it a try. The whole upsampling game is not to insert new stuffs into the content, it is to reduce the artifacts that are introduced in the digital to analog process in a DAC.

 

[JSmith]

Utterly pointless.

Yes the question was rather rhetoric.


JSmith

 

[Sokel]

Hmm... I thought the noise level should be similar (or at least not worse) if the rate goes up. Would that be related to a particulart DAC's implmentation? Or does it something related to the nature of DSD coding?

From the DXD doc provided by Merging Technologies ( https://www.merging.com/uploads/assets/Merging_pdfs/dxd_Resolution_v3.5.pdf ), it suggests that the noise levels are similar (shown on Figure 3 in the doc and I'd attached it here). It is too bad that the figure is quite blur....

Could be the DAC,here's another similar one with a Cosmos ADC measured by @pkane :

It looks like it's the same

(My measurements are very limited by the ADC and also in real world conditions,with their own cables,on the rack as they sit,etc)

 

[dlovesmusic]

How about DSD512? My DAC can support up to DSD256. I am currently upsampling all my contents to DSD256 with HQPlayer.

Do you think it is worth to spend on a DAC in order to get DSD512 (for HQPlayer)? Or it would just be similar to DSD256?

The DAC I have are ESS DAC (ESS9018K2M), CS DAC (43131). My understanding is that they do not support direct DSD mode. If I get a AKM DAC that supports DSD512 and DSD direct, would that make a difference when compared with what I have right now?

Actually for most dacs - native dsd delta sigma dacs or nos discrete dacs, dsd256 are usually the “sweet spot”, you get a decent audible band noise floor with substantially most ultrasonic noise being pushed further out.

Dsd512 is usually a trade off compared to dsd256, you are a few db worse in audible band noise compared to dsd256 but ultrasonic noise being pushed even further out…

Dsd1024 is where there is pretty much nil benefits because ultrasonic noise is pretty much already non existent at 512 or even 256 but audible band noise floor is measured worst compared to even dsd512. Even Miska who designed hqplayer does not use dsd1024 and never recommend it

 

[BDWoody]

Do you think it is worth to spend on a DAC in order to get DSD512 (for HQPlayer)?

No.

Even Miska who designed hqplayer does not use dsd1024 and never recommend it.

He's also never actually provided evidence there is an audible benefit to using it.

People like to buy stories.

Just wondering did you try HQPlayer and upsample your content to DSD256? If not, I highly suggest you to give it a try.

Understanding is all that is needed.

 

[pkane]

Could be the DAC,here's another similar one with a Cosmos ADC measured by @pkane :

It looks like it's the same

(My measurements are very limited by the ADC and also in real world conditions,with their own cables,on the rack as they sit,etc)


View attachment 359257

Higher rate DSD allows quantization noise to be moved further out to higher frequencies. Type and order of the modulator can be used to push out more or less of the quantization noise from audible range to ultrasonic. Latest Multitone gives you the choice of modulator order, from 4 to 7, with 6 being the default (as it always was).

 

[Sokel]

He's also never actually provided evidence there is an audible benefit to using it.

I don't use HQPlayer but I had a look around and to be fair it seems that he provides measurements where objective (but probably inaudible) results show some improvements
(some less,some more,some DACs benefit more) .

But it seems to be device-specific,similar to the one I measured in the other thread.
I think that we need to measure a lot to conclude which ones it benefits objectively .

'till now it was difficult even to find decent DSD test files so measurements were extremely rare but Multitone solved that big time.

 

[BDWoody]

I don't use HQPlayer but I had a look around and to be fair it seems that he provides measurements where objective (but probably inaudible) results show some improvements
(some less,some more,some DACs benefit more) .

He does indeed, but doesn't support his claim of audible improvement. That's when the discussions here always went south. Lots of graphs, but no evidence of an audible improvement or difference.

 

[Sokel]

He does indeed, but doesn't support his claim of audible improvement. That's when the discussions here always went south. Lots of graphs, but no evidence of an audible improvement or difference.

There may be some cases (but again,DAC-specific) that it could be.
Some example is here where at a given level (around -30db) the difference is some 20db and approaches audible area.

Someone must make some serious measurements of course,with serious gear and a large sample,the above is just hints.

 

[antcollinet]

If not, I highly suggest you to give it a try.

Have you tried comparing level matched, blind and statistically relevant?

If not, I highly suggest you give it a try. You could save yourself a hell of a lot of effort - and probably storage space - if you do.