Which DACs bypass digital filtering?

[ThenewGearPPK]

I know iBasso portable music player added a selectable NOS filter through a firmware update. This almost seem like it would be a emulated filter instead of bypassing everything.

 

I remember reading Rob Watts of Chord Electronics talking about a lot of DAC's that use NOS are using a digital filter instead.

Something about the DAC needing good quality resistors to be able to do true NOS.

 

[jparvio]

About the list; is there a reason for not include Totaldac? AFAIK all the models are r-2r ladder (with Vishay resistors) and natively NOS by nature but one can activate a software FIR in case of...Well... Fun? 

 

At least that's how it is with my d1 mk2. 

[semente]

On 4/29/2021 at 6:43 AM, jparvio said:

About the list; is there a reason for not include Totaldac? AFAIK all the models are r-2r ladder (with Vishay resistors) and natively NOS by nature but one can activate a software FIR in case of...Well... Fun? 

 

At least that's how it is with my d1 mk2. 

 

I had TotalDAC and a few others before (in lighter gray - here) but ended up removing them some because they wouldn't take high rates and bit depth and/or others because their technical performance is not particularly outstanding.

 

Edit:

This is not a generic list of NOS DACs. As mentioned in the original post, this is a list of high-performance* commercial DACs that allow user to bypass digital filtering (NOS), leaving that task to a software music player such as HQ Player.

 

* - no TDA1543-based DACs, and only equipment which will take 24bit and sampling rates above 176.4/192kHz and perhaps DSD.

[semente]

Can someone please confirm that Musician Audio's DACs are eligible to go on the list?

 

http://www.musician-audio.com/en/

[semente]

Denafrips DACs are disqualified, they're not NOS.

 

 

Updated list:

 

- Aqua DACs

- Armature (some models): Asterion (similar to Holo Spring) and Cronos (similar to Denafrips Ares)

- Audiobyte Vox

- Bricasti M21 and M3 (DSD-only)

- Esoteric (some models?)

- Holo Audio (some models)

- iFi micro (some models)

- Musician Audio

- Metrum Acoustics

- Phasure NOS1 DAC

- Singxer SDA-x family

- RME ADI-2 DAC, ADI-2 Pro ADC/DAC

- Rockna DACs

- T+A (some models): DAC 8 DSD, SD 3100 HV, SDV 3100 HV, HA 200

- Teac UD-50x family

- Topping (some AKM models): E30, T70, T90

[semente]

Any newcomers to be added?

 

 

Updated list:

 

- Aqua DACs

- Armature (some models): Asterion (similar to Holo Spring) and Cronos (similar to Denafrips Ares)

- Audiobyte Vox

- Bricasti M21 and M3 (DSD-only)

- Esoteric (some models?)

- Holo Audio (some models)

- iFi micro (some models)

- Musician Audio

- Metrum Acoustics

- Phasure NOS1 DAC

- Singxer SDA-x family

- RME ADI-2 DAC, ADI-2 Pro ADC/DAC (AKM models only)

- Rockna DACs

- T+A (some models): DAC 8 DSD, SD 3100 HV, SDV 3100 HV, HA 200, DAC 200

- Teac UD-50x family, UD-701

- Topping E30, T70, T90 (AKM models only)

[barrows]

T+A DAC 200 with DSD at least.  this is a separate product from HA 200.

[dericchan1]

With the recent firmware update and the launch of the performance edition, the ifi neo idsd dac should be added to the list of NOS dac, with now the option to go bitperfect NOS for pcm, and it uses a discreet channel for dsd that supports native dsd512

https://ifi-audio.com/products/neo-idsd/

[louawalters]

On 1/7/2022 at 6:05 PM, semente said:

Any newcomers to be added?

 

 

Updated list:

 

- Aqua DACs

- Armature (some models): Asterion (similar to Holo Spring) and Cronos (similar to Denafrips Ares)

- Audiobyte Vox

- Bricasti M21 and M3 (DSD-only)

- Esoteric (some models?)

- Holo Audio (some models)

- iFi micro (some models)

- Musician Audio

- Metrum Acoustics

- Phasure NOS1 DAC

- Singxer SDA-x family

- RME ADI-2 DAC, ADI-2 Pro ADC/DAC (AKM models only)

- Rockna DACs

- T+A (some models): DAC 8 DSD, SD 3100 HV, SDV 3100 HV, HA 200, DAC 200

- Teac UD-50x family, UD-701

- Topping E30, T70, T90 (AKM models only)

So @GoldenOne did some analysis on the big Rockna Wavedream.  I have the wavelight, and I wonder if it behaves like its bigger brother when put into NOS mode.  I have recently tried @goldenOnes HQ Player settings, and set the WL to NOS, and I was amazed at the improvement.   I can only send 8X rates, but colour me impressed.   The Wavelight is SUPER sensitive to power cords.  I will share that I had for the past year been using a LessLoss cMark power cord.  No slouch the lessloss is, but before the Christmas holiday I was able to acquire a Synergetic Research Galileo UEF power cord.  Yes this PC is expensive, but the improvement in sound stage depth, width and placement of instruments within that stage was SIGNIFICANT.   So, would be interested in other people's journey with wavelight DACs.   This DAC continues to impress me with its ability to improve.  I.E QSA Purple https://www.tweekgeek.com/quantum-science-audio-6-3-x-30mm-slow-blow-fuse-limited-availability/ fuse in the DAC over the SR Orange is profound!   So I have two thoughts, how much better is the big Rockna?  2) how close in performance is my tweeked wavelight to the bigger brother.   Would the MAY KTE be an improvement - using HQ player, or do I set sights on something more significant.    I am very impressed with the influence the renderer has on the overall sound.   

 

 

[GoldenOne]

On 1/7/2022 at 11:05 PM, semente said:

Any newcomers to be added?

 

 

Updated list:

 

- Aqua DACs

- Armature (some models): Asterion (similar to Holo Spring) and Cronos (similar to Denafrips Ares)

- Audiobyte Vox

- Bricasti M21 and M3 (DSD-only)

- Esoteric (some models?)

- Holo Audio (some models)

- iFi micro (some models)

- Musician Audio

- Metrum Acoustics

- Phasure NOS1 DAC

- Singxer SDA-x family

- RME ADI-2 DAC, ADI-2 Pro ADC/DAC (AKM models only)

- Rockna DACs

- T+A (some models): DAC 8 DSD, SD 3100 HV, SDV 3100 HV, HA 200, DAC 200

- Teac UD-50x family, UD-701

- Topping E30, T70, T90 (AKM models only)

So it's probably worth noting that a fair few of the DACs on this list do not actually bypass digital filtering. For normal PCM playback anyway.
Many of them will offer options that produce an output similar to NOS, but is not actually NOS.

It may be worth having two lists. One of 'Native DSD converters' which can do DSD without alteration, and one for 'Native PCM converters' which can do PCM without alteration/filtering.
As these two are not mutually exclusive. Some DACs like Rockna can do PCM NOS, but must convert DSD to PCM first. And some DACs like AKM can convert DSD natively, but must use oversampling & Delta-sigma modulation to play PCM.

Going down the list:

Audiobyte Vox: This is a 1-bit DAC so could not operate NOS. It could play DSD with no alteration but not PCM.

Ifi micro iDSD: The 'bitperfect' mode on these is using the same approach as the Teac (and I believe esoteric) models that used the burr brown advance-segment dacs. I've described what's happening at the bottom of this post: https://goldensound.audio/2021/12/02/teac-ud501-dac-review-measurements/

but the summary is that it effectively converts at a 6-bit level and still uses delta-sigma modulation at a high rate for the lower 18 bits.
It does not do any filtering for the MSBs but is not NOS.

Musician Audio: These are denafrips based designs so whilst I've not had one in to check, I would bet money that these are using the same approach as denafrips which is to use linear interpolation in NOS mode, not actually NOS and produces a result very different to actual NOS. This is discussed at the bottom of this post: https://goldensound.audio/2021/10/07/denafrips-terminator-plus-with-gaia-measurements/
Termi plus 'NOS' impulse:

 

Singxer SDA-X and RME ADI-2:  These have a 'NOS' mode, but it's not actually NOS (as these chips are not full 16 bit converters and cannot run NOS, they require delta sigma to function). It is just a zero-order-hold oversampling filter.
This produces a result that for the most part resembles NOS, but still, is not actually NOS and dithering & delta-sigma modulation (and oversampling itself) are still occurring. They can convert DSD in true 1-bit fashion with no alteration though.

T+A: These can avoid any alteration for DSD, but not for PCM. Again these are not native PCM/full 16 bit converters. To my knowledge a lot of these are using burr brown chips such as the DAC 8 DSD so will probably be doing the same approach as ifi/teac.

Teac UD50X family: This is a slightly trickier one as they've changed the DAC chips being used several times and so it may not be exactly the same for all DACs in the family. But the UD501 I've discussed and provided a link above, this is doing the same 'filter bypass' approach as the iDSD products but is not actually NOS.
I've not had a chance to use their newer DACs with the discrete design, but the fact that they are not full native 16 bit converters in itself means it cannot actually be NOS.

Topping DACs: Same as SDA-X and ADI-2

One to add: The Gustard X26 pro also has a 'NOS' mode. Not exactly sure if this is doing anything different to standard filtering internally. But the filter itself is just a very short, minimum phase filter, it's not actually NOS at all and does not resemble NOS.

 

If you want to ACTUALLY do NOS, it is a requirement that the converter itself be a native PCM converter at the bit depth you want to convert. Ie: You need an R2R dac. But even then not all R2R dacs actually do NOS

[Qhwoeprktiyns]

Just out of curiosity, why does it matter what the DAC is doing "internally"? Shouldn't we just be concerned with how it sounds? 

[GoldenOne]

2 minutes ago, hopkins said:

Just out of curiosity, why does it matter what the DAC is doing "internally"? Shouldn't we just be concerned with how it sounds? 

Because what it's doing internally determines how it will sound.

Many people prefer the sound of a true NOS DAC and so will want to know what DACs are NOS.
They may not want to spend money on something that is not actually doing what they want. A lot of these DACs are not easily available to demo and so it's not possible to evaluate/compare in person prior to purchase

[Qhwoeprktiyns]

35 minutes ago, GoldenOne said:

Because what it's doing internally determines how it will sound.

Many people prefer the sound of a true NOS DAC and so will want to know what DACs are NOS.
They may not want to spend money on something that is not actually doing what they want. A lot of these DACs are not easily available to demo and so it's not possible to evaluate/compare in person prior to purchase

 

There are so many factors that will determine the sound quality of a DAC that i think it is a little foolish to make a choice based on any single criteria. You just have to listen. 

 

How a DAC sounds is also obviously dépendant on the equipment that you use it with. Because DACs are not perfect they are susceptible to the source.  Preamps, amplifiers, speakers, cables, also have their own  sound caracteristics (they are not transparent), so no DAC is going to sound the same in two different systems. That's a real when problem comparing audio components, which is unfortunately rarely considered as everyone (manufacturers, reviewers, customers) would like to think otherwise. 

 

DAC A could sound "neutral" in one system and "colored" in another. DAC B could sound "detailed" in one system and "muddy" in another. 

 

That's why people keep changing their DACs all the time and DACs are the topic of endless discussions :) 

 

So whether a DAC is really NOS or not probably "does not amount to a hill of beans"... 

 

But carry on 🙂

[GoldenOne]

10 minutes ago, hopkins said:

 

There are so many factors that will determine the sound quality of a DAC that i think it is a little foolish to make a choice based on any single criteria. You just have to listen. 

 

How a DAC sounds is also obviously dépendant on the equipment that you use it with. Because DACs are not perfect they are susceptible to the source.  Preamps, amplifiers, speakers, cables, also have their own  sound caracteristics (they are not transparent), so no DAC is going to sound the same in two different systems. That's a real when problem comparing audio components, which is unfortunately rarely considered as everyone (manufacturers, reviewers, customers) would like to think otherwise. 

 

DAC A could sound "neutral" in one system and "colored" in another. DAC B could sound "detailed" in one system and "muddy" in another. 

 

That's why people keep changing their DACs all the time and DACs are the topic of endless discussions :) 

 

So whether a DAC is really NOS or not probably "does not amount to a hill of beans"... 

 

But carry on 🙂

True, but the presence of many factors does not mean we should just ignore any of them individually and not even look at all about the design/operation of a DAC before buying.

Otherwise we might as well be throwing darts at a board randomly to pick what to get. 

We cannot build up a perfect picture of how something will sound from specs and measurements, but there are plenty of things which can give a good general indication.
And from personal experience, I'd say that NOS vs OS dacs are one of the biggest single factors tbh. They do sound quite different indeed and I'm not sure I'd want to move to a daily driver DAC that didn't have NOS as an option in future.

 

Things like whether a DAC has a linear or switching PSU might not be something you should decide your purchase on, but you'll probably still want to know.

 

If we didn't choose to consider anything and instead choose blindly cause 'there are too many things so why bother' then this forum need not exist.

Plus, even if some things might not be all that important or maybe even audible at all, a lot are still simply just quite interesting to discuss and talk about regardless :)

[Qhwoeprktiyns]

Yes of course it is still interesting to understand the inner workings. 

 

The Terminator, for example, has the two modes, and neither I nor the owner of the model I have heard, could ever notice a significant difference. You may say that the reason is because it is not "really NOS" and that could be true. 

 

Regardless, your own findings in your system may not be "universal". The best way to evaluate a DAC, IMO, is to test it in as many systems as possible. When you do that, you will realize that DAC performance is unfortunately highly variable. 

 

There is more dart throwing happening than we may think. 

[dericchan1]

1 hour ago, GoldenOne said:

Ifi micro iDSD: The 'bitperfect' mode on these is using the same approach as the Teac (and I believe esoteric) models that used the burr brown advance-segment dacs. I've described what's happening at the bottom of this post: https://goldensound.audio/2021/12/02/teac-ud501-dac-review-measurements/

but the summary is that it effectively converts at a 6-bit level and still uses delta-sigma modulation at a high rate for the lower 18 bits.
It does not do any filtering for the MSBs but is not NOS.
 

So do IFI that uses the burr brown chips NOS for DSD NOT PCM?

[GoldenOne]

33 minutes ago, dericchan1 said:

So do IFI that uses the burr brown chips NOS for DSD NOT PCM?

AFAIK yes it does convert in true 1-bit without alteration, though NOS is a bit of a different debate for DSD. As DSD is not intended to have any digital processing done. The intended way to convert DSD is to convert with a true 1-bit converter and then apply an analog low pass filter.

Whereas PCM is intended to have a digital reconstruction filter used to adhere to nyquist.

The effect of something being 'NOS' on PCM and DSD are very different.

[louawalters]

1 hour ago, GoldenOne said:

True, but the presence of many factors does not mean we should just ignore any of them individually and not even look at all about the design/operation of a DAC before buying.

Otherwise we might as well be throwing darts at a board randomly to pick what to get. 

We cannot build up a perfect picture of how something will sound from specs and measurements, but there are plenty of things which can give a good general indication.
And from personal experience, I'd say that NOS vs OS dacs are one of the biggest single factors tbh. They do sound quite different indeed and I'm not sure I'd want to move to a daily driver DAC that didn't have NOS as an option in future.

 

Things like whether a DAC has a linear or switching PSU might not be something you should decide your purchase on, but you'll probably still want to know.

 

If we didn't choose to consider anything and instead choose blindly cause 'there are too many things so why bother' then this forum need not exist.

Plus, even if some things might not be all that important or maybe even audible at all, a lot are still simply just quite interesting to discuss and talk about regardless :)

I agree very interesting discussion - without having an AP to test my wavelight DAC what other markers or indicators would identify that the dac is or is not a true NOS?  I suppose I could ask Nicolae....

 

[GoldenOne]

27 minutes ago, louawalters said:

I agree very interesting discussion - without having an AP to test my wavelight DAC what other markers or indicators would identify that the dac is or is not a true NOS?  I suppose I could ask Nicolae....

 

There are usually two big giveaways:

1) Is the impulse response actually something that NOS could produce. (ie: A square, or analog lowpassed square)
2) Does FFT show clear activity at a common resampling multiple of 44.1khz/48khz?
(Such as 352.8khz, 384khz, 705.6khz, 768khz)

A proper NOS impulse response should look like a square due to how NOS/sample and hold/zero-order-hold behaviour works.



Sample 1: still 0/silence. Dac continues holding at 0 output
Sample 2: Higher level, DAC moves output level up to meet it, then holds
Sample 3: 0 Again, DAC immediately moves down to meet it, then holds
Sample 4: Still silence, DAC will continue holding.

As an example this is the impulse from the Phasure NOS1A DAC:

Little bit of overshoot which is likely something to do with impedance matching. But otherwise the square is pretty obvious.
Some DACs might have an analog lowpass filter to eliminate unnecessary ultrasonic frequencies (such as those above 100khz or something) even though the converter itself is NOS (Holo Dacs for example) and so you'll see a slight curve to this shape:



But if you see something which bears absolutely no resemblance to this and couldn't be explained by an analog filter, then you can pretty much immediately discount it as being NOS.
Such as the terminator plus or gustard x26 pro 'NOS' impulses:

BUT, even if you do have an IR that looks like that, you still can't necessarily conclude it is NOS.
For example the ADI-2's 'NOS' filter gives a fairly NOS looking impulse response:


In this case, you can do a few other checks. One of the easiest is looking to see if there is signal content at a common resampling multiple of the input sample rate. (Though this requires a high bandwidth analyzer as you'd need to be sampling at over 1.5mhz in most cases) For example if you're playing 44.1khz content and the DAC has a spike at 705.6khz then that's a good indicator that oversampling is occurring as there would be no reason for that to be there.
Especially if it then changes to 768khz when you swap to a 48khz input. As that shows it's tied to the sample rate and not just some component happening to oscillate at exactly that fixed frequency.

For example on the Ares 2 in 'NOS' mode (or OS mode). You can see a large spike at 705.6khz.

This also changes to 768khz if you play something that is in 48khz. 

With the wavedream specifically (I've not yet tested wavelight unfortunately but would like to), we see the same thing at 705.6khz when in OS mode:

But when changing to NOS mode it looks quite different:


Still quite a bit of activity which is a bit unusual but the main clear spike is gone.
Looking at the NOS impulse:


There is some pre/post ringing evident. HOWEVER, this ringing is riiiight at the edge of the bandwidth of my analyzer so in this particular case I can't say for certain if this is ringing introduced by the DAC itself not being NOS, or if it's because the rise/fall time exceeds a 1.2mhz bandwidth which would mean the ADC's own filter is causing this.
I didn't have an oscilloscope to check at the time so would need to check again at some point if I'm able to get a wavelight in or a wavedream again for some reason :P

(Or if anyone has a wavedream/wavelight and an oscilloscope and would be willing to check)



PS: Just to clarify, I'm not at all saying that any particular reconstruction method is 'best'. This is simply a description of a couple ways to check if something is indeed actually NOS or just pretending to be.
Not whether being NOS is preferable or not, that part is upto you.

[louawalters]

@GoldenOne would this do the trick :

https://www.robotshop.com/ca/en/seeedstudio-dso-nano-v3.html?gclid=CjwKCAiA0KmPBhBqEiwAJqKK41HPs50shhek3nE1n8AkM3QvphMUBRljARGqxxw-zNP_ounYmzfsNxoCHX4QAvD_BwE

 

[GoldenOne]

18 minutes ago, louawalters said:

@GoldenOne would this do the trick :

https://www.robotshop.com/ca/en/seeedstudio-dso-nano-v3.html?gclid=CjwKCAiA0KmPBhBqEiwAJqKK41HPs50shhek3nE1n8AkM3QvphMUBRljARGqxxw-zNP_ounYmzfsNxoCHX4QAvD_BwE

 

Unfortunately not. That only has a bandwidth of 1mhz which is lower than the analyzer. 

You'd ideally want a 100mhz oscilloscope or higher

[semente]

8 hours ago, GoldenOne said:

So it's probably worth noting that a fair few of the DACs on this list do not actually bypass digital filtering. For normal PCM playback anyway.
Many of them will offer options that produce an output similar to NOS, but is not actually NOS.

It may be worth having two lists. One of 'Native DSD converters' which can do DSD without alteration, and one for 'Native PCM converters' which can do PCM without alteration/filtering.
As these two are not mutually exclusive. Some DACs like Rockna can do PCM NOS, but must convert DSD to PCM first. And some DACs like AKM can convert DSD natively, but must use oversampling & Delta-sigma modulation to play PCM.

Going down the list:

Audiobyte Vox: This is a 1-bit DAC so could not operate NOS. It could play DSD with no alteration but not PCM.

Ifi micro iDSD: The 'bitperfect' mode on these is using the same approach as the Teac (and I believe esoteric) models that used the burr brown advance-segment dacs. I've described what's happening at the bottom of this post: https://goldensound.audio/2021/12/02/teac-ud501-dac-review-measurements/

but the summary is that it effectively converts at a 6-bit level and still uses delta-sigma modulation at a high rate for the lower 18 bits.
It does not do any filtering for the MSBs but is not NOS.

Musician Audio: These are denafrips based designs so whilst I've not had one in to check, I would bet money that these are using the same approach as denafrips which is to use linear interpolation in NOS mode, not actually NOS and produces a result very different to actual NOS. This is discussed at the bottom of this post: https://goldensound.audio/2021/10/07/denafrips-terminator-plus-with-gaia-measurements/
Termi plus 'NOS' impulse:

 

Singxer SDA-X and RME ADI-2:  These have a 'NOS' mode, but it's not actually NOS (as these chips are not full 16 bit converters and cannot run NOS, they require delta sigma to function). It is just a zero-order-hold oversampling filter.
This produces a result that for the most part resembles NOS, but still, is not actually NOS and dithering & delta-sigma modulation (and oversampling itself) are still occurring. They can convert DSD in true 1-bit fashion with no alteration though.

T+A: These can avoid any alteration for DSD, but not for PCM. Again these are not native PCM/full 16 bit converters. To my knowledge a lot of these are using burr brown chips such as the DAC 8 DSD so will probably be doing the same approach as ifi/teac.

Teac UD50X family: This is a slightly trickier one as they've changed the DAC chips being used several times and so it may not be exactly the same for all DACs in the family. But the UD501 I've discussed and provided a link above, this is doing the same 'filter bypass' approach as the iDSD products but is not actually NOS.
I've not had a chance to use their newer DACs with the discrete design, but the fact that they are not full native 16 bit converters in itself means it cannot actually be NOS.

Topping DACs: Same as SDA-X and ADI-2

One to add: The Gustard X26 pro also has a 'NOS' mode. Not exactly sure if this is doing anything different to standard filtering internally. But the filter itself is just a very short, minimum phase filter, it's not actually NOS at all and does not resemble NOS.

 

If you want to ACTUALLY do NOS, it is a requirement that the converter itself be a native PCM converter at the bit depth you want to convert. Ie: You need an R2R dac. But even then not all R2R dacs actually do NOS

 

Thanks for the input.

 

Feel free to edit my list, splitting it into PCM filter-bypass DACs and DSD filter-bypass DACs.

We can even ask Chris to rename the topic.

 

I think I made it reasonably clear in my first post what kinds of DACs should go on the list:

 

I have been unsuccessfully searching for a list of high-performance* commercial DACs that allow user to bypass digital filtering (NOS), leaving that task to a software music player such as HQ Player.

It would be interesting if we could compile such a list in this thread.

 

* - no TDA1543-based DACs please, only equipment that can take 24bit and sampling rates above 176.4/192kHz

[semente]

5 hours ago, GoldenOne said:

As an example this is the impulse from the Phasure NOS1A DAC:

Little bit of overshoot which is likely something to do with impedance matching.

 

A bit off-topic, but could you tell us how you're connecting the Phasure to the AP? Can you use USB?

[barrows]

@hopkins:  To my mind, the purpose of this list is to have a verified source of DACs which will natively convert an incoming sample rate.  The reason an audiophile might want to have this information is for those audiophiles who might want to take control of the aspect of digital filtering/OS through software (like HQPlayer) rather than being "stuck" with what the DAC manufacturer, or even just the DAC chip manufacturer provides.  Many have had the experience that HQPlayers much more sophisticated oversampling algorithms are superior to that offered onboard in most (perhaps all) DACs.

Certainly my experience, listening, suggests that HQPlayer's approach to oversampling in demonstrably superior to most DACs of my acquaintance.  This is not really that surprising, considering the processing horsepower required by HQP oversampling.  The I9-9900K processor I use in my server for this is many magnitudes more power fun than any onboard processor in a DAC, and HQPlayer takes advantage of this additional computational power to allow for markedly better sound quality in playback.

So far the only DAC with its own onboard oversampling which has come close is the Mola Mola Tambaqui, which uses multiple AD SHARC chips to apply its oversampling.

[GoldenOne]

1 hour ago, semente said:

 

Thanks for the input.

 

Feel free to edit my list, splitting it into PCM filter-bypass DACs and DSD filter-bypass DACs.

We can even ask Chris to rename the topic.

 

I think I made it reasonably clear in my first post what kinds of DACs should go on the list:

 

I have been unsuccessfully searching for a list of high-performance* commercial DACs that allow user to bypass digital filtering (NOS), leaving that task to a software music player such as HQ Player.

It would be interesting if we could compile such a list in this thread.

 

* - no TDA1543-based DACs please, only equipment that can take 24bit and sampling rates above 176.4/192kHz

No prob!

With your original definition I think that no changes to the thread are needed. Just the clarification that PCM NOS and Native DSD conversion are separate things.

'NOS' is technically a correct description of native DSD conversion, but also a less than ideal one as most dacs that cannot convert DSD natively would be 'downsampling' and converting to PCM. So there probably aren't any dacs that 'oversample' DSD as such.
Though 'native DSD support' also gets messy cause it could be describing both the actual ability to natively convert DSD, or just the ability to receive a 'native DSD' stream over ASIO.

Not sure what the best term would be tbh but anyway :P

Splitting the list into two:

NOS PCM Capable DACs:

- Aqua

- Armature Asterion

- Holo Audio (Excluding Cyan DSD)

- Metrum Acoustics

- Phasure NOS1 

- Rockna

- Sonnet Morpheus
- Audio GD

Native DSD Capable DACs:

- Audiobyte Vox

- Bricasti M21 and M3

- Esoteric (some models?)

- Holo Audio Spring 3, May, Cyan DSD

- iFi iDSD lineup

- Singxer SDA-x family with AKM chips

- RME ADI-2 DAC, ADI-2 Pro ADC/DAC (AKM models only, must enable 'DSD Direct')

- T+A (some models): DAC 8 DSD, SD 3100 HV, SDV 3100 HV, HA 200, DAC 200

- Teac UD-50x family, UD-701

- Topping E30, T70, T90 (AKM models only, must enable 'DAC mode', not 'Pre Mode')

- Many other AKM and Burr-Brown based DACs

Though worth noting that some of the above are assumptions, unfortunately there's a lot of manufacturers that claim NOS capability but do not actually have it.
Also there are many DACs (such as all ESS dacs for example) that have 'Native DSD' support, ie: they can accept the native DSD transmission over ASIO, but cannot actually CONVERT natively and must convert to PCM first.

57 minutes ago, semente said:

 

A bit off-topic, but could you tell us how you're connecting the Phasure to the AP? Can you use USB?

I unfortunately had nothing but trouble getting the Phasure to work, and never got it to work properly. I only ended up being able to play stuff at 48khz to it, and even then had various issues that prevented me from being able to actually listen to it.
That measurement (or recording I suppose) was taken using an ADI-2 Pro ADC at 768khz, it was before I'd gotten the analyzer.
 

39 minutes ago, barrows said:

@hopkins:  To my mind, the purpose of this list is to have a verified source of DACs which will natively convert an incoming sample rate.  The reason an audiophile might want to have this information is for those audiophiles who might want to take control of the aspect of digital filtering/OS through software (like HQPlayer) rather than being "stuck" with what the DAC manufacturer, or even just the DAC chip manufacturer provides.  Many have had the experience that HQPlayers much more sophisticated oversampling algorithms are superior to that offered onboard in most (perhaps all) DACs.

Certainly my experience, listening, suggests that HQPlayer's approach to oversampling in demonstrably superior to most DACs of my acquaintance.  This is not really that surprising, considering the processing horsepower required by HQP oversampling.  The I9-9900K processor I use in my server for this is many magnitudes more power fun than any onboard processor in a DAC, and HQPlayer takes advantage of this additional computational power to allow for markedly better sound quality in playback.

So far the only DAC with its own onboard oversampling which has come close is the Mola Mola Tambaqui, which uses multiple AD SHARC chips to apply its oversampling.

That'd be my understanding too.
If I'm looking explicitly for a NOS DAC it's for one of two reasons.

1) I like the sound of NOS and would like it as an option
2) I want to be able to use external tools like HQPlayer to the fullest extent and guarantee that no filtering or noise shaping work will be undone/masked/ruined by further processing of any sort done in the DAC.

In regards to the first one, this would require a NOS PCM capable DAC.
But in regards to HQPlayer, it could be either NOS PCM or Native DSD capability depending on user preferences.

Though if you wanted both then could go for the Holo DACs. Afaik those are currently the only DACs available that can do both NOS PCM and Native DSD conversion as they have both a PCM R2R ladder and a discrete 1-bit DSD converter. Plus a further bonus for HQPlayer users in the form of 1.536mhz PCM and DSD1024 support.