[roderickvd]

The latest Ares firmware has digital volume control over USB, if that fits your bill.

 

[gimmeheadroom]

No volume control on Audio-gd standalone DACs, fortunately. But the AIO units like R27 have that.

I would also suggest R-8Mk2. Reserve money for DI-20HE

That's my setup. And even with inexpensive offshore "Klipsch" speakers, my 2 channel setup is more resolving, dynamic, and lifelike than anything my headstation is capable of.

 

[betula]

No volume control on Audio-gd standalone DACs, fortunately. But the AIO units like R27 have that.

My bad, I was actually looking at the R28 and R27...

 

[gimmeheadroom]

My bad, I was actually looking at the R28 and R27...

I hardly consider that bad

If you want to buy the ultimate all-in-one look at the R27HE. Killer!

 

[sajunky]

Few technical comments on the volume control. Soekris done it right by increasing number of bits in the ladder. In this way sound lose digital resolution only with very low volume levels. Ares use a standard ladder, as it is a firmware upgrade. It is a worst solution. It cannot compete against external amp with analogue volume control. R-28 use stepped attenuator for perfect channel balance like in Soekris, but after D/A conversion, it is a proper analogue type.

With a good headphone amp, R-8Mk2 will give the best sound quality and the external clock input gives a room for further improvement.

 

[roderickvd]

What do you mean with increasing number of bits in the ladder versus a standard ladder? Where did you find implementation details of these two units?

I am not using digital volume control on either, just genuinely interested where you got this.

Principally both units could promote any input bit depth to 32 bit by simply padding it with zeroes, then attenuating it would still throw away bits but with loss of dynamic range way above the threshold of human hearing. I don’t know the implementation details of either.

Of course I prefer a separate R-2R attenuation ladder but I don’t think the Soekris has that either, it does it all in the digital domain.

 

[sajunky]

What do you mean with increasing number of bits in the ladder versus a standard ladder? Where did you find implementation details of these two units?

I am not using digital volume control on either, just genuinely interested where you got this.

Principally both units could promote any input bit depth to 32 bit by simply padding it with zeroes, then attenuating it would still throw away bits but with loss of dynamic range way above the threshold of human hearing. I don’t know the implementation details of either.

Of course I prefer a separate R-2R attenuation ladder but I don’t think the Soekris has that either, it does it all in the digital domain.

I know about Soekris implementation from what he has published. It is just mentioned that additional bits are added to compensate loss of a dynamic range. Not sure how it works, I guess these bits provide a dithering required in such operation. Otherwise merely padding some bits with zeros produce distortions, it is why digital volume control has a reputation of being inferior to the analogue. Soekris at least tried to do it right, not sure how is effective.

A proper digital volume control require a high precision mathematic operation, followed by dithering to randomise quantisation errors. It is how it is done on Linux or MAC OS by converting to the 32-bit floating point, but not under Windows. Windows use 24 or 32-bit Integer internally, it is inferior. Foobar 2000 also use 32-bit floating point internally, sound degradation is minimal if a slider is at less than 100%

Digital processing in the DAC is not so powerful as on the PC, therefore it must be inferior, no question about. There is an additional pitfal of having digital volume control inside a DAC, at least when using Foobar, not many people are aware of. When a DAC reports to the PC using USB configuration descriptors a volume control internal capability, Foobar do not apply a volume processing internally, but sends a command to a DAC, so it is done inside a DAC instead. Unfortunately there is no option to disable such 'feature' in Foobar. It is better to have a DAC that do not report to the host about such ability, then volume control in Foobar works as expected.

A digital attenuation (as implemented in R-28) is not related to the R2R ladder, it can be also placed on the output of Delta-Sigma DAC. A series of relays switch to a different resistor sets, giving a required attenuation using a fixed high precision resistors. Everything is static, relays are activated or deactivated only when volume knob is adjusted, it gives a precise balance between channels. A mechanical pot characteristic is not so closely defined. It is the only difference between a pot and a digital attenuator.

 

[roderickvd]

I know about Soekris implementation from what he has published. It is just mentioned that additional bits are added to compensate loss of a dynamic range. Not sure how it works, I guess these bits provide a dithering required in such operation. Otherwise merely padding some bits with zeros produce distortions, it is why digital volume control has a reputation of being inferior to the analogue. Soekris at least tried to do it right, not sure how is effective.

This is not the case with attenuation. You would be right with zero-hold interpolation when upsampling, which indeed does increase distortion. I know this because I personally wrote this audio code in librespot, the open source Spotify client you all know and love from your favourite audio distribution. Which indeed has a 64-bit floating point internal pipeline, then can convert it to any bit depth. Following the arithmetic you can see that "adding bits" is for all intents and purposes the same as "zero padding" by scaling to the bounds of the integer.

Not entirely by accident I also wrote the dithering code, and without going in depth I can assure that this only alters (not adds) 1 or 2 MSB -- it's another matter.

Digital processing in the DAC is not so powerful as on the PC, therefore it must be inferior, no question about.

Indeed I don't know of any DACs that does floating point in hardware. I think most of them do 32 bit integer (thus padding for inputs with lower bit depth). This gives 192 dB of DNR. It's true that floating point gives even higher DNR, but at 192 dB you can still attenuate by 100 dB and be beyond the threshold of human hearing. So technically inferior yes, but unnoticeable in the case attenuation. For DSP filter chains or upsampling it's a different matter, that indeed takes a PC of some FPGA's or you'll lose precision.

A digital attenuation (as implemented in R-28) is not related to the R2R ladder, it can be also placed on the output of Delta-Sigma DAC. A series of relays switch to a different resistor sets, giving a required attenuation using a fixed high precision resistors. Everything is static, relays are activated or deactivated only when volume knob is adjusted, it gives a precise balance between channels. A mechanical pot characteristic is not so closely defined. It is the only difference between a pot and a digital attenuator.

Yes, this is what I mean. It's also called a R2R attenuator (not to be mistaken with an R2R DAC) which can be stepped or with relays. Circling back that's why I asked what you meant with a "standard ladder" and why you believe the Ares implementation is the worst.

 

[hpamdr]

Soekris 2541 R2R DAC vs. RME ADI-2 FS (AKM4493 version)


I guess the journey never ends, but this DAC is a pleasant stop. My aim is probably to rebuy a Chord TT2, but I was hoping I can get away with something more affordable. It seems I prefer R2R tone to D/S by a big margin, but I miss a bit of resolution, bass clarity and bass impact. I see, Audio GD products (R28) are getting some love, even in this thread. Would the R28 be a better match for my personal taste? R2R sound with slightly clearer and more impactful bass?

To me 2451 is a bit darker with black filter this should increase a bit bass tone, also it likes to be hot to gives it best so do not hesitate to turn it on 15 minutes before listening.

 

[betula]

To me 2451 is a bit darker with black filter this should increase a bit bass tone, also it likes to be hot to gives it best so do not hesitate to turn it on 15 minutes before listening.

Yes, I have noticed it needs a bit of warm up time. Perhaps even more than 15 mins.

 

[gimmeheadroom]

Yes, I have noticed it needs a bit of warm up time. Perhaps even more than 15 mins.

It has been around 39 here the last few days. Send it over

 

[lostrockets]

Soekris 2541 R2R DAC vs. RME ADI-2 FS (AKM4493 version)

First impressions:

This write up is not a review or a proper comparison, just a selection of random personal notes and subjective impressions.

In a nutshell, I wish it was possible to combine some of the characteristics of the two DACs. Just for the record, I made this comparison with an external amp, with my Burson Soloist 3XP in power amp mode in this case. The internal amps of these DACs are not bad but also not great. Perhaps the Soekris internal amp is a little better than the RME's, but nothing to write home about. Versus the 3XP the internal amps just sound a bit closed in and congested with a loss of headspace, resolution and refinement. I would say the amp section on the 2541 performs somewhere in the $300-400 range, when it comes to standalone external amps.

Here I will evaluate the DAC sections only. Cutting straight to the chase, overall the Soekris 2541 produces the better sound, but not without compromises. The 2541 sounds a lot more lifelike, alive and real. It is also warmer and softer, but not overly warm or overly soft. Soren's DAC has without a question a more pleasing sound, especially when it comes to acoustic instruments or vocals. Vocals are a lot richer in tonality and sound more lifelike, more organic.

The ADI-2 in comparison has the technical edge when it comes to clarity, resolution and dynamics. This is the most obvious in bass reproduction, especially with modern or electronic music. Bass on the RME is just more defined, clearer with bigger impact, better dynamics and focus. Dynamics (especially macro) on the RME are overall better. The real issue on the German DAC is with tonality and then mids and treble compared to the Soekris R2R. The RME just sounds a bit dead and boring in comparison. Like a cold, lifeless recording. Sure, the clarity and that bass impact are attractive and entertaining, but the rest of the sound is just dry and lacks soul. This is true for D/S DACs in general, not unique to RME. RME is actually a hint warmer and smoother than many delta/sigma DACs. Still, versus the Soekris R2R these differences are very apparent.

To me personally, bass rendering is one of the top priorities, as a good portion of my music is modern EDM (ambient). So to me this compromise on the Soekris is a little bit painful. At the same time, I find it nearly impossible to go back to RME (or any delta/sigma for that matter) after the Soekris due to its richer, more lifelike and more pleasing tonality.

Soundstage on the RME is wider, but on the Soekris it is a little deeper and instruments are better positioned in a 3D space. Not by an awful lot as I was expecting it, but still better.

I guess the journey never ends, but this DAC is a pleasant stop. My aim is probably to rebuy a Chord TT2, but I was hoping I can get away with something more affordable. It seems I prefer R2R tone to D/S by a big margin, but I miss a bit of resolution, bass clarity and bass impact. I see, Audio GD products (R28) are getting some love, even in this thread. Would the R28 be a better match for my personal taste? R2R sound with slightly clearer and more impactful bass?

I am happy to experiment further as buying and selling is part of the hobby for me, but I start to think I should probably just save up and get the TT2 as I personally found that unit the most satisfying so far, but it is hard to justify the cost even second-hand.

In the meantime I think I will probably keep the Soekris and sell the RME, despite the softer and less impactful bass. The other advantages, like a richer sound, are simply too good to let go.

Objectively the RME and the 2541 are both good DACs but do things differently. Ultimately the choice should be based on personal sonic priorities and musical taste: clarity & impact but a less engaging and less lifelike sound versus a soulful and musically engaging sound but with a softer and slightly less technical presentation. Tough choice, as they emphasise different treats.

I quite like the slightly retro look of the dac2541.

Further impressions after some more listening as I got used to the Soekris sound even more:

Sound reproduction versus (live) music listening.
Studio clarity/sterility versus live performance.
Details are actually on a very similar level, just the RME is slightly more open and brighter in the treble which contributes to the perception of higher resolution. As well as the speedier (PRaT), more precise sound of the D/S DAC.
RME definitely comes across as colder, Soekris definitely feels richer and more organic. With acoustic music it is not even a contest. With studio produced electronic music the RME has the edge because of the added dynamics and punch but the Soekris actually does a decent and enjoyable job as well, just in a little softer way. To be clear, the difference is not in bass quantity or body, it is about focus and impact.

Orchestral or live performance acoustic music brings you a lot closer to the real experience with the Soekris. Also, electronic music with real sounds and samples in it (Amon Tobin for example) actually sounds better on the Soekris. It elevates the experience. The RME's advantage really is only with hardcore electronic stuff like Boris Brejcha. I also have to mention that RME is a studio audio company, just audiophiles picked up their ADI-2 due to its clear sound, versatility and compact form factor. RME is not a personal boutique home audio company like most companies we usually buy from.

Even though on some of my tracks I will miss the bass from the ADI-2, the vast majority of my music is more engaging, more enjoyable on the 2541. My only complaint about the 2541 really is that the low end could be a little more impactful. The rest is pretty delicious. I am enjoying it quite a lot actually.

To wrap up, let me leave a few notes here for folks who say all DACs sound the same:

I recently talked to a person, who told me, to his ears all good quality headphones above $200 sound the same. Then we have a group of audio enthusiasts who say it is absolutely pointless to spend more than $500 on a headphone amplifier as all they do is make the audio signal stronger/louder. So effectively they say every good amplifier above $500 sounds the same.
Then you have the group of people who say all DACs sound the same. Then people who say audio cables make a difference but not USB cables. Or not everyone can hear the difference between DAC filters and so on...

It is true that headphones make the biggest difference, then amps, then DACs and then cables. One should keep these priorities in mind. At the same time saying "there is no difference in sound between this and that because I can't hear it" is simply silly. It is like saying "it is impossible to run 100m quicker than 18 seconds, because I can't run it quicker". Well, hearing abilities differ as well.
And please do not start a cable debate.

Thanks for this detailed post, i also have the adi rme2 and a soekris 1321. Have listened to both ab testing and they sound very close on my stax setup. Ill listen more to acoustic and live and some of my favorite electronica but my first impressions were not that night and day different.

Curious if the differences to you were as obvious on speakers as headphones?

 

[betula]

Thanks for this detailed post, i also have the adi rme2 and a soekris 1321. Have listened to both ab testing and they sound very close on my stax setup. Ill listen more to acoustic and live and some of my favorite electronica but my first impressions were not that night and day different.

Curious if the differences to you were as obvious on speakers as headphones?

I only tried them on my Meze Elite. My current circumstances don't allow to have speakers around. I presume there also should be quite a difference between the 1321 and 2541.

That said, although some of the differences (softness, lifelikeness versus sterility and dynamism) were obvious immediately, it took me 2-3 days to really understand the new sound of Soekris. I have to say, I love it a lot more now than initially, I simply can't go back to the RME.

I presume, system chain matters as well, on some systems (or as you say speakers vs headphones) the differences might be more obvious. The Meze Elite is sophisticated enough to bring out link changes in your audio chain in a pretty obvious manner.

Edit: I also have to add, that I just picked up a Bryston BHA-1 amplifier and this adds a lot more clarity, resolution and dynamic impact to the sound than the Burson Soloist 3XP. Another proof, how much pairing and audio chain matters. Synergy can be everything.

 

[technisummer]

I recently had the luck to try out the 1541 DAC with a pair of HD600's at an audio shop. This was my first experience with an R2R DAC, so my impressions are entirely subjective.

To me, the music felt a bit softer and less impactful than on a DS chip DAC, but it sounded more full and rich. Separation seemed more effortless too. But the higher frequencies did feel a tad muffled, and the bass felt softer in impact as well.

The first idea that popped into my head was - is it possible to retain the positive qualities I mentioned earlier while adding a better sense of impact if I use the DAC in pair with a proper headphone amplifier?

 

[sajunky]

I recently had the luck to try out the 1541 DAC with a pair of HD600's at an audio shop. This was my first experience with an R2R DAC, so my impressions are entirely subjective.

To me, the music felt a bit softer and less impactful than on a DS chip DAC, but it sounded more full and rich. Separation seemed more effortless too. But the higher frequencies did feel a tad muffled, and the bass felt softer in impact as well.

The first idea that popped into my head was - is it possible to retain the positive qualities I mentioned earlier while adding a better sense of impact if I use the DAC in pair with a proper headphone amplifier?

Yes. For R2R DAC you need a good class A non-feedback amplifier. And not a hybrid (opamps based) like one made by Burson, i.e. @betula found Bryston to sound completely different. It is expensive, but there are less expensive, by example Audio GD Master 19. Soekris has something wrong on its side, as it use opamps, but still a good class A amp makes a difference. If you want more impactful sound without losing details, Audio GD DAC is worth to try. The same character as Rockna Wavelight (I haven't heard WL, just based on reviews).

You can try A-GD R-1, it is ~$900, while R-8Mk2 is the current model I would recommend.