[GoldenSound]

So there is a trade-off between the worse Holo DAC perception of DSD512 rate and the superior behavior of 512+ modulator at this rate as compared to DSD256. Based on the measurements above is it possible to recommend for Holo DAC owners the optimal combination: DSD256 rate + ASDM7EC-Super 512+fs or DSD512 rate + ASDM7EC-Super 512+fs?

This is a behaviour shared by many DACs. Higher rate DSD isn't always better

Here's the T+A DAC 200 for instance.

DSD256:

DSD512:

Noise is pushed further out, as expected, but you can see the distortion products increase, and quite significantly (upto 15dB) under 20khz.

You can still use the 512+fs modulators at DSD256. It's just that because they optimise for dynamic range rather than pushing noise further out, on some DACs (and depending on downstream components) it may not be ideal.
Jussi has mentioned that due to Holo's DSD converter acting as a moving average analog domain FIR filter, this isn't really an issue though so the 512+fs modulators work nearly without penalty at 256

 

[bogi]

Any example of da taking direct modulated stream without using it's own modualation?

See https://audiophilestyle.com/forums/...tal-filtering/?do=findComment&comment=1254269

 

[GoldenSound]

ESS chips don't convert "back" to lower PCM rates when they are fed by DSD data. No resampling is applied to DSD data in ESS chips. Data format is adapted to allow optional volume level adjustment and to fit modulator input, but original DSD data resolution is not lost like in the case of downsampling to PCM rates (which involves so called decimation). Adapted input data is then re-modulated in ESS chip and finally it is fed to D/A stage. ESS modulator runs at incoming DSD data rate. For DACs utilizing this ESS chip functionality it is suitable to feed them with DSD data. That brings still less internal processing in DAC chip than with PCM input since complete hardware oversampling (including the 2nd simplified stage) is skipped with DSD input.

Sometimes companies keep that stuff close to their chest (not necessarily because it's 'secret' but simply because they didn't think it's something that might be important to discuss). So we don't know exactly what it gets converted to internally, could be 32 bit 768khz could be 11Mhz 5 bit, we have no idea. But we do know that it IS resampled/converted.

It is not converted natively. Meaning you cannot benefit from high performance modulators in the same way that you can with an un-processed output. The block diagram for the 9038 for instance shows it going through ASRC

 

[LXMc]

I don't mean to derail the conversation, I just want to be sure my setup is working the way I think it is...I'm a bit new at this type of thing:

I've been searching for a means to utilize HQPlayer with Deezer as that's been my preferred streaming service for a while. I really dislike the UX in Qobuz and Tidal organization is a mess. Unfortunately Deezer is very much plug-and-play and gives very few options such as no output selection. I had been curious about HQPlayer and was messing with it using Roon/Qobuz trials but the cost of those in comparison didn't feel worth it for both when I still don't love Roon's UX/UI for my uses.

I came across a thread on SBAF that discussed using Virtual Audio Cable to route HQPlayer so as to use a better interface like Foobar, so I downloaded the VAC lite version, got it set up by assigning the VAC line-in as the input in HQP and my DDC as the output. My question is, the Lite version of VAC is capped at 16/48000, but since Deezer has everything at 16/44100 that should be more than sufficient right? I would actually set VAC's rate at 16/44100, and then HQP will take that and upsample the crap out of it to 24/192000 which is the cap on my DAC. The only annoying part is I have to make sure I switch PC output back to my DDC and close HQP if I want to do anything other than listen to music since there's a major delay when it all routs through HQP.

The full version of VAC allows up to 32/384000 but since Deezer is only outputting 16/44100 I don't think this makes any difference to me. Does this setup make sense on paper? I haven't had a chance yet to really A/B the differences since figuring out how to set it up (also tricky with the delay on HQPlayer switching sources and all that). Another question I had is - if HQP is outputting at 24/192000 through my DDC, do I need to set my DDC output to that within Windows? or does HQP just override whatever you have set there?

Appreciate any input!

 

[GoldenSound]

I came across a thread on SBAF that discussed using Virtual Audio Cable to route HQPlayer so as to use a better interface like Foobar, so I downloaded the VAC lite version, got it set up by assigning the VAC line-in as the input in HQP and my DDC as the output.

One thing to note is that VAC can be bitperfect but isn't by default. You need to set Stream format limit to 'cable format' and disable channel mixing/volume control to make sure it's bit perfect.

I've no idea if Deezer actually has WASAPI Exclusive output though. Previously it didn't, which means all your audio will be going through windows mixer and won't be bitperfect anyway unfortunately. No idea why a lossless streaming service has no exclusive output option.

if HQP is outputting at 24/192000 through my DDC, do I need to set my DDC output to that within Windows? or does HQP just override whatever you have set there?

Nope, HQPlayer will adjust your DAC/DDC sample rate automatically

 

[LXMc]

One thing to note is that VAC can be bitperfect but isn't by default. You need to set Stream format limit to 'cable format' and disable channel mixing/volume control to make sure it's bit perfect.

Thanks for this - it does appear that VAC lite doesn't have the option to adjust those cable settings - I had attempted to limit it manually by removing the range at the top but not sure what sort of difference it makes. The cost of a regular license isn't that much so I'll likely switch to that and get those options, but mostly just to support the dev.

I've no idea if Deezer actually has WASAPI Exclusive output though. Previously it didn't, which means all your audio will be going through windows mixer and won't be bitperfect anyway unfortunately. No idea why a lossless streaming service has no exclusive output option.

You're correct that Deezer [still] does not have a WASAPI exclusive mode. My hope was I could get it as close as possible by keeping output through VAC at 16/44100. It's definitely not ideal (and pretty ridiculous it's still not in there), but it may be a concession I'm willing to put up with in the end.

Nope, HQPlayer will adjust your DAC/DDC sample rate automatically

Good to know - thanks again for the info

 

[GoldenSound]

Thanks for this - it does appear that VAC lite doesn't have the option to adjust those cable settings - I had attempted to limit it manually by removing the range at the top but not sure what sort of difference it makes. The cost of a regular license isn't that much so I'll likely switch to that and get those options, but mostly just to support the dev.



You're correct that Deezer [still] does not have a WASAPI exclusive mode. My hope was I could get it as close as possible by keeping output through VAC at 16/44100. It's definitely not ideal (and pretty ridiculous it's still not in there), but it may be a concession I'm willing to put up with in the end.


Good to know - thanks again for the info

Definitely worth supporting the Dev. It's a great piece of software.

Alternatively, VB-Hifi cable (not VB-Audio cable) is free and bitperfect. It just doesn't support automatic sample rate switching, you have to manually set the input and output to the same sample rate/bit depth in windows sound settings.

 

[1laraz]

You can still use the 512+fs modulators at DSD256. It's just that because they optimise for dynamic range rather than pushing noise further out, on some DACs (and depending on downstream components) it may not be ideal.
Jussi has mentioned that due to Holo's DSD converter acting as a moving average analog domain FIR filter, this isn't really an issue though so the 512+fs modulators work nearly without penalty at 256

Thank you a lot for your valuable explanation, GoldenOne!

 

[bogi]

But we do know that it IS resampled/converted.

Converted yes, but according to more sources not downsampled. The block schematics is too simplified.

Since ESS did not provide any official info on this, we rely on alternative sources available.

This one is describing processing with 2.8M input: https://audiophilestyle.com/forums/...volume-control/?do=findComment&comment=327563
~~~~~~~~~~~~~~~~~
Mark Mallinson asked for a few days to see if he could get an answer and later replied with this:
Quote

"Inside the Sabre DAC chip the DSD stream is processed at the DSD rate, but it is extending to have a 32 bit representation so that it can be scaled. This is NOT the same as converting to PCM as that implies decimation filtering. Inside the Sabre, all DSD data is processed in a PCM domain, however its PCM at 64*44.1kHz = 2.8224MHz. This data is then applied to the modulator."

I assume this came from ESS, since there are close ties between both companies.
~~~~~~~~~~~~~~~~~

barrows working for Sonore: https://audiophilestyle.com/forums/...ampling-to-dsd/?do=findComment&comment=801485
~~~~~~~~~~~~~~~~~
My understanding with the ESS chips is that they take incoming DSD, and then make it multibit, but they do not down convert the sample rate. This way they avoid the possibility of creating digital artifacts through the decimation process (good) and retaining the full sample rate of DSD. They basically make the data stream DSD wide, a 32 bits, so they can then apply their (excellent) digital volume control.

I am sure @Miska will correct me if understand this incorrectly. So it is not "DSD direct" in the strictest sense, but is also not DSD-PCM conversion.
~~~~~~~~~~~~~~~~~

Miska's conclusion based also on measurements is also that ESS does not downsample incoming DSD, but processes it further at input rate.
https://audiophilestyle.com/forums/...bit-dsd-debate/?do=findComment&comment=273957
~~~~~~~~~~~~~~~~~
Sabre works similar way as DSD processor mode in CS4398 (non-DirectDSD). So there is digital processing for the DSD stream (filter, volume and remodulator). This diagram of CS4398 illustrates it quite well:

(note on CS4398 there's a configuration option to choose "DSD Processor" path or "Direct DSD" path, controlling those "MUX" switches)

Output of both ESS and Cirrus DSD processing engines is still at the native sampling rate (2.8/5.6 MHz etc) but multi-bit SDM.
~~~~~~~~~~~~~~~~~

 

[GoldenSound]

Converted yes, but according to more sources not downsampled. The block schematics is too simplified.

Since ESS did not provide any official info on this, we rely on alternative sources available.

This one is describing processing with 2.8M input: https://audiophilestyle.com/forums/...volume-control/?do=findComment&comment=327563
~~~~~~~~~~~~~~~~~
Mark Mallinson asked for a few days to see if he could get an answer and later replied with this:
Quote


I assume this came from ESS, since there are close ties between both companies.
~~~~~~~~~~~~~~~~~

barrows working for Sonore: https://audiophilestyle.com/forums/...ampling-to-dsd/?do=findComment&comment=801485
~~~~~~~~~~~~~~~~~
My understanding with the ESS chips is that they take incoming DSD, and then make it multibit, but they do not down convert the sample rate. This way they avoid the possibility of creating digital artifacts through the decimation process (good) and retaining the full sample rate of DSD. They basically make the data stream DSD wide, a 32 bits, so they can then apply their (excellent) digital volume control.

I am sure @Miska will correct me if understand this incorrectly. So it is not "DSD direct" in the strictest sense, but is also not DSD-PCM conversion.
~~~~~~~~~~~~~~~~~

Miska's conclusion based also on measurements is also that ESS does not downsample incoming DSD, but processes it further at input rate.
https://audiophilestyle.com/forums/...bit-dsd-debate/?do=findComment&comment=273957
~~~~~~~~~~~~~~~~~
Sabre works similar way as DSD processor mode in CS4398 (non-DirectDSD). So there is digital processing for the DSD stream (filter, volume and remodulator). This diagram of CS4398 illustrates it quite well:

(note on CS4398 there's a configuration option to choose "DSD Processor" path or "Direct DSD" path, controlling those "MUX" switches)

Output of both ESS and Cirrus DSD processing engines is still at the native sampling rate (2.8/5.6 MHz etc) but multi-bit SDM.
~~~~~~~~~~~~~~~~~

That seems to align with what I've been told by a few manufacturers. (Though funnily enough manufacturers themselves have been seemingly given different answers or no answers by ESS in some cases. Apparently ESS is a lot more difficult to communicate with than AKM.)

But the point still stands, it's not being converted natively.
If it's being converted to 32 bit, that's PCM. It might be very high rate, but it's still PCM. And the point I was making about native DSD is that then you can take advantage of a particularly high performance noise shaper, but you cannot when it is being processed such as ESS is doing.

How much it matters is another question. The reason I mention it is simply that if you are wanting to get an improvement via HQP DSD modulators 'replacing' those in your DAC, you cannot with ESS.

If one feels that high performance modulators make no difference, then they can just use the PCM input anyway and not worry about it.

 

[rkt31]

See https://audiophilestyle.com/forums/...tal-filtering/?do=findComment&comment=1254269

I meant dac

 

[bogi]

But the point still stands, it's not being converted natively.

The reason I mention it is simply that if you are wanting to get an improvement via HQP DSD modulators 'replacing' those in your DAC, you cannot with ESS.

I never told it is converted natively. I pointed to a fact that internal oversampling is fully skipped and that's the main difference against case when DSD is converted to lower rate PCM.

There is a difference between
- downsampling to usual PCM rates, which except of resolution loss caused by decimation inolves also anti-alias filtering and thus removing upper part of original audio band
and
- format conversion from unary to binary data representation. To call it PCM is only a semantics. The term 'binary data representation' is more general and existed long time before PCM audio format was introduced. Under DSD to PCM conversion we usually understand downsampling to usual PCM rates and bit depths with anti aliasing and decimation involved.
So once again: No anti alias filtering, no decimation, no hardware oversampling with ESS chip fed by DSD data.

You don't reflect the difference between these 2 cases. You present too black and white vision.

you can take advantage of a particularly high performance noise shaper, but you cannot when it is being processed such as ESS is doing

That's again simplified black and white vision and what's worse, you present a purely subjective opinion as a truth. "You cannot" is only your subjective statement and like any other statements about sound it is only an opinion, equally true as opinion of other people. I was using ESS DACs with HQPlayer set to DSD output for many years. Introduction of EC-super modulators led me to order HQPlayer v5 - with ESS chip based DAC, so much I liked the change in sound.

I compared PCM and DSD input with more ESS based DACs in the last 10 years, maybe 6 or 7 pieces and I always came to the same conclusion. For my ears it is always a benefit to avoid ESS oversampling filters by feeding an ESS based DAC by DSD data stream. I am using also a direct DSD capable DAC so I can compare how much of the direct DSD benefit one can perceive with an ESS chip based DAC.

Of course, people's taste is individual subjective thing, so other persons may have different personal opinion.

 

[bogi]

I meant dac

Yes, I understood. On my link you can find a list of direct DSD capable DACs (the bottom part).

 

[rkt31]

Yes, I understood. On my link you can find a list of direct DSD capable DACs (the bottom part).

Thanks. Do those smsl and topping ones deal dsd natively ie they will treat hq player modulated stream natively, without any modulation of their own?

 

[bogi]

Thanks. Do those smsl and topping ones deal dsd natively ie they will treat hq player modulated stream natively, without any modulation of their own?

Yes. Some models may contain extra setting to get direct DSD, or it may be set when device is switched from preamp mode to fixed volume mode.

Generally, digital volume control is not available with direct DSD mode. Of course, digital volume control in available in HQPlayer - both for PCM and DSD content. For DSD content it works (as all DSP) without intermediate conversion to PCM domain.