Review and Measurements of Holo Audio May --- Probably the best discrete R2R DAC

[barrows]

Yes, @majingotan indeed! I would suggest that you have a listen to the Tambaqui before disparaging it though, you might be surprised. Just because it is expensive is no reason to disparage something, as no one is forcing anyone to purchase it. I think looking at the measurements anyone with even the smallest understanding of engineering would be impressed, regardless of pricing strategy.
Regarding the May, looking at what the KTE version is, there could be actual measured improvements resulting for the claim that the modules are hand picked, at least for the PCM section, as the precision of the resistors will make a difference: if they measure twenty modules, and just use the top ten percent in the KTE version, then, perhaps there could be a measured improvement.
I am interested in the DSD performance though and for DSD the resistor precision is not really an issue. I am talking about the Tambaqui here as there is a similarity in how the Tambaqui conversion stage operates and how the May (for DSD) conversion stage operates: If one oversamples in software to high rate DSD, and sends that to the the May's DSD stage in NOS/DSD mode, performance (and sound quality) might be similar to the Tambaqui.

 

[EuropeanSwallow]

Thank you for this thread. I placed my order for the Holo May KTE DAC a few days ago. I think I might just remove the covers and stare at it mindlessly while playing tunes! I've been waiting patiently for an R2R DAC that measured well and wasnt priced above 10k+. Currently, I own the Matrix XSabre Pro MQA DAC which is a great measuring and sounding DAC as well, but I am so enamored by the engineering that went into the May that I've just got to own it if for no other reason than I can spend hours oggling it!

(I'm well aware that it may not sound any different than what I have presently, but it's a beautiful work of art and masterfully engineered, and I must possess it )

 

[solderdude]

@EuropeanSwallow : What is the airspeed velocity of an unladen swallow?

 

[EuropeanSwallow]

@EuropeanSwallow : What is the airspeed velocity of an unladen swallow?

10 Meters per second according to Google... Ni! Lol

 

[Jimbob54]

10 Meters per second according to Google... Ni! Lol

Please tell me your name is Martin

 

[EuropeanSwallow]

Please tell me your name is Martin

Jonathan, the last time I checked.

 

[Jimbob54]

Jonathan, the last time I checked.

You picked the wrong bird species then ;-)

 

[EuropeanSwallow]

You picked the wrong bird species then ;-)

View attachment 83532

ROFL

 

[solderdude]

10 Meters per second according to Google... Ni! Lol

You can cross the bridge safely now but watch out for the rabbit.

 

[EuropeanSwallow]

You can cross the bridge safely now but watch out for the rabbit.

Viscious creatures! Lol

 

[Ravenasec]

@EuropeanSwallow

Hello,
I have owned the same Matrix dac as you.
Did you get any audible benefit from Holo May KTE in comparison? Currently, I am using RME ADI2 DAC FS (second version with AKM 4493 inside).

 

[barrows]

@EuropeanSwallow

Hello,
I have owned the same Matrix dac as you.
Did you get any audible benefit from Holo May KTE in comparison? Currently, I am using RME ADI2 DAC FS (second version with AKM 4493 inside).

Careful, you are asking for a subjective opinion, which appears to be somewhat frowned upon here. I would also mention that best performance with the Holo Audio DACs (measured, and subjectively) occurs when one uses NOS mode and oversamples in software to high rates, especially to DSD 256 using HQPlayer. It is very likely that anyone listening to the May is not hearing it at its best without oversampling in software.

 

[Veri]

Careful, you are asking for a subjective opinion, which appears to be somewhat frowned upon here. I would also mention that best performance with the Holo Audio DACs (measured, and subjectively) occurs when one uses NOS mode and oversamples in software to high rates, especially to DSD 256 using HQPlayer. It is very likely that anyone listening to the May is not hearing it at its best without oversampling in software.

Why is that? I had a holo spring 2 at home for quite some time and only fed it normal PCM, sounded just fine to me. I feel like HQPlayer upsampling is just another nervosa in this audiophile scene.

 

[sq225917]

Upsampling is just a cludge to ameliorate crap 45khz filtrts

 

[barrows]

Why is that? I had a holo spring 2 at home for quite some time and only fed it normal PCM, sounded just fine to me. I feel like HQPlayer upsampling is just another nervosa in this audiophile scene.

Oversampling allows all the artifacts in playback to be shifted to higher frequencies where they are then removed by the analog filter stage, without oversampling there are a lot of artifacts which obscure detail and reduce resolution. The onboard oversampling chip in the Holo DACs offers fairly poor performance compared to what software based oversampling is capable of: consider the processing power of a full fledged computer.
Jussi Laasko of Signalyst has published many measurements showing the output of the Holo Audio Spring, and the improvements in the output when oversampling to much higher rates in software. The measurements at stereophile.com of the Holo Audio May clearly show the problems when the DAC is in NOS mode (as is true of any NOS DAC running at 44.1 PCM). Jussi has also published measurements comparing the results of the onboard oversampling chip vs. oversampling in software, and the difference is clear for anyone to see.
Additionally, many owners of both the Spring and the May have reported best sound in playback with these DACs when the DAC is in NOS mode and using high rates of PCM (352.8/384) or DSD 256/512, oversampled in software.
I would recommend trying HQPlayer based oversampling with your Spring, it is free to try (HQPlayer will run for 30 minutes in demo mode). Run 44.1 PCM up to 352.8, and see what you think. Also consider trying it with all input converted to DSD 256 in HQPlayer, instead of "feel(ing)" like this is just "nervosa", at least then you would have some actual experience to base your opinion on. Additionally HQPlayer offers many, many different filter options which allow one to fine tune the presentation to their liking-the eis no "correct" choice necessarily with 44.1 rate input files, as every digital filter is a compromise with CD rates due to their inherent limitations.

 

[Veri]

I would recommend trying HQPlayer based oversampling with your Spring, it is free to try (HQPlayer will run for 30 minutes in demo mode). Run 44.1 PCM up to 352.8, and see what you think. Also consider trying it with all input converted to DSD 256 in HQPlayer, instead of "feel(ing)" like this is just "nervosa", at least then you would have some actual experience to base your opinion on. Additionally HQPlayer offers many, many different filter options which allow one to fine tune the presentation to their liking-the eis no "correct" choice necessarily with 44.1 rate input files, as every digital filter is a compromise with CD rates due to their inherent limitations.

I did install and trial it I didn't play with the DSD much but did try the 1.536k PCM upsampling as well as dithering to improve linearity. I was of the opinion that I could not really hear the improvement, should there be one. But as mentioned I did not exhaust the DSD options.

 

[barrows]

My favorite approach with HQPlayer is to oversample to DSD 256. using the ASDMEC7 modulator and the poly-sinc ext2 filter. This approach has sounded absolutely fantastic on every DSD specific DAC I have used here. I feel the EC modulator is really a breakthrough in terms of hearing a natural sound. I have both a DSC-2 style DSD DAC (quite similar approach to Holo Audio) and a Bricasti M3, which has a different design discrete, DSD converter. I have not used HQPlayer much for PCM-PCM oversampling, and do not have an R2R NOS DAC here to try it with so i do not feel i have heard the best of what HQP can do with PCM-PCM oversampling, but my preference for realistic natural sound lies with DSD anyway.

 

[Veri]

Never heard of that DSC2 dac. It looks kinda cool

 

[mocenigo]

Upsampling is just a cludge to ameliorate crap 45khz filtrts

Especially for those that for totally unscientific reasons prefer NOS DACs.

 

[Extreme_Boky]

It's a good way to market that custom DAC chip design for audiophools and charge 20x more, just make sure it gets >130 dB SNR and >120 dB SINAD at full scale. That pretty much sounds like Mola Mola Tambaqui to me. Then again, this is the Holo May DAC thread, but I'm super hoping that Amir or Wolf can get the Kitsune Edition for measurements as it's 1K USD more for some fancy wiring, fancy caps and transformers

The Kitsune edition also has hand-selected DAC boards, apparently...

The silly thing that is really putting me off from getting the May DAC in general, is the Kitsune label they plonked on Talema transformers... really?
I have nice annealed pure silver ribbons for the hookup wiring between IEC connector and power supply PCB; the capacitors, film and electrolytes, can be picked to suit individual needs/rest of the system. Despite voltage-controlled oscillators and jitter reduction, I do not see anything there that would markedly reduce the phase noise at < 0.1 Hz spectrum; the Terminator II has oven-controlled oscillators.

Kitsune edition should have better linearity if it does have the hand-selected DAC boards. So, yes, it would be nice to see its measured specifications