PS Audio DirectStream DAC Mk.2 measurements

[Matias]

-80 dB ultrasonic noise floor, amazing... For 8 k usd.

PS Audio PerfectWave DirectStream MK2 D/A processor Measurements | Stereophile.com

Sidebar 3: Measurements I measured the DirectStream MK2 first using my Audio Precision SYS2722 system, then with the magazine's higher-performance APx500. The review sample was fitted with the v2.3.5 FPGA firmware, but after I experienced some anomalous measured results, I installed the v2.3.3...

www.stereophile.com

 

[amirm]

Good grief. On top of the many bugs, it is not linear to beyond -80 dB or 13 bits! These people really need to start measuring their gear....

 

[amirm]

This is supposed to look like a sine wave:

So much noise pollution.

 

[Martini]

This is supposed to look like a sine wave:

So much noise pollution.

What, you can't see it?

 

[Ricardus]

I saw some narcissistic audiophools sitting around a table on some youtube channel talking about how this latest DAC from PS Audio BLOWS THE OLD ONE OUT OF THE WATER, and how when you hear the sound, the soundscape is so real "you can walk through it."

 

[restorer-john]

This is supposed to look like a sine wave:

Actually not. That plot is undithered 16 bit data and it should show three distinct levels like hundreds of players and DACs JA has previously reviewed. But yeah, I've never seen a Stereophile plot with that much noise.

 

[DonR]

According to the review, it is "musical" so fear not and dump your cash. Paul is waiting.

 

[Matias]

Actually not. That plot is undithered 16 bit data and it should show three distinct levels like hundreds of players and DACs JA has previously reviewed. But yeah, I've never seen a Stereophile plot with that much noise.

View attachment 289863

Yes, this is how it should be below, from Mola Mola. After all these years one would assume the competition would catch up, not distance themselves even further...

Gramophone Dreams #55: Mola Mola Tambaqui D/A processor Measurements | Stereophile.com

Measurements, from January 2022 (Vol.45 No.1): When Herb Reichert reviewed the Mola Mola Tambaqui in his December 2021 Gramophone Dreams column, he very much liked what he heard from this Bruno Putzeys–designed D/A processor. "During my first days of listening, the Mola Mola's most conspicuous...

www.stereophile.com

 

[Sokel]

I don't know if I did it right but even poor E-MU looks better.

(1Khz,16bits,undithered at 44100Hz)

 

[restorer-john]

Yes, this is how it should be below, from Mola Mola. After all these years one would assume the competition would catch up, not distance themselves even further...

Gramophone Dreams #55: Mola Mola Tambaqui D/A processor Measurements | Stereophile.com

Measurements, from January 2022 (Vol.45 No.1): When Herb Reichert reviewed the Mola Mola Tambaqui in his December 2021 Gramophone Dreams column, he very much liked what he heard from this Bruno Putzeys–designed D/A processor. "During my first days of listening, the Mola Mola's most conspicuous...

www.stereophile.com

View attachment 289918

Notice also the levels, at 250-300uV vs the autoscaling on the AP to 12-25mV to account for the broadband noise on the PS unit. There also appears to be an asymmetric offset. Perhaps there's an issue with the tested unit?

 

[restorer-john]

I don't know if I did it right but even poor E-MU looks better.

That's a really good effort. I can't comment on absolute levels, but it looks pretty much right.

 

[IAtaman]

Actually not. That plot is undithered 16 bit data and it should show three distinct levels like hundreds of players and DACs JA has previously reviewed. But yeah, I've never seen a Stereophile plot with that much noise.

View attachment 289863

Ha! It is supposed to look like a band-limited square wave but there is so much noise it looks like it is supposed to look like a sine wave.

 

[Hayabusa]

Yes, this is how it should be below, from Mola Mola. After all these years one would assume the competition would catch up, not distance themselves even further...

Gramophone Dreams #55: Mola Mola Tambaqui D/A processor Measurements | Stereophile.com

Measurements, from January 2022 (Vol.45 No.1): When Herb Reichert reviewed the Mola Mola Tambaqui in his December 2021 Gramophone Dreams column, he very much liked what he heard from this Bruno Putzeys–designed D/A processor. "During my first days of listening, the Mola Mola's most conspicuous...

www.stereophile.com

View attachment 289918

-90db should be ~60 uVolt rms...

 

[charleski]

Obviously the DirectWave has an issue with far-ultrasonic noise as a result of the design resampling everything to DSD (JA measured 210mV of noise at 380kHz). But what I thought was interesting was how this managed to mess up the AP's linearity measurements. Fig 6 shows a small deviation from linear at -90dBFS, but this is far smaller than the ~+9dB indicated by the graph. I wonder if this explains why other DACs with badly-controlled ultrasonic output also produce wildly poor linearity graphs, because he found similarly ghastly behaviour with the Mojo Mystique.

 

[Mart68]

I was quite surprised to see this from John Atkinson in the subjective review:

''I burned in the cables with a CD on repeat overnight before continuing my critical auditioning''

PS Audio PerfectWave DirectStream MK2 D/A processor | Stereophile.com

When Art Dudley reviewed the original PS Audio PerfectWave DirectStream D/A processor in Stereophile's September 2014 issue, he very much liked what he heard. "For those who've waited for a computer-friendly DAC that offers, with every type of music file, the best musicality of which DSD is...

www.stereophile.com

I appreciate that there's a line to be trodden with subjective reviewing (some playing to the gallery is mandatory) but surely he does not believe in cable burn in?

Eight thousand dollars!

 

[ObjectiveSubjectivist]

This thing looks like a noise generator.

 

[kevin1969]

Obviously the DirectWave has an issue with far-ultrasonic noise as a result of the design resampling everything to DSD (JA measured 210mV of noise at 380kHz). But what I thought was interesting was how this managed to mess up the AP's linearity measurements. Fig 6 shows a small deviation from linear at -90dBFS, but this is far smaller than the ~+9dB indicated by the graph. I wonder if this explains why other DACs with badly-controlled ultrasonic output also produce wildly poor linearity graphs, because he found similarly ghastly behaviour with the Mojo Mystique.

I'm not an audio engineer but one of the fascinating things I saw in that review was the mention about ultrasonic noise affecting amplifiers that don't compensate for it

Could you explain what that means? It sounds pretty significant.

 

[egellings]

I saw some narcissistic audiophools sitting around a table on some youtube channel talking about how this latest DAC from PS Audio BLOWS THE OLD ONE OUT OF THE WATER, and how when you hear the sound, the soundscape is so real "you can walk through it."

Just watch out for all the punji stakes, though.

 

[VQR]

I was quite surprised to see this from John Atkinson in the subjective review:

''I burned in the cables with a CD on repeat overnight before continuing my critical auditioning''

PS Audio PerfectWave DirectStream MK2 D/A processor | Stereophile.com

When Art Dudley reviewed the original PS Audio PerfectWave DirectStream D/A processor in Stereophile's September 2014 issue, he very much liked what he heard. "For those who've waited for a computer-friendly DAC that offers, with every type of music file, the best musicality of which DSD is...

www.stereophile.com

I appreciate that there's a line to be trodden with subjective reviewing (some playing to the gallery is mandatory) but surely he does not believe in cable burn in?

Eight thousand dollars!

Eh, some people if paid enough will write whatever nonsense you want. I wouldn't be surprised if he genuinely didn't believe in cable burn-in but wrote that to please the editor and readership.

 

[charleski]

I'm not an audio engineer but one of the fascinating things I saw in that review was the mention about ultrasonic noise affecting amplifiers that don't compensate for it

Could you explain what that means? It sounds pretty significant.

A properly designed audio amp will handle ultrasonic noise without any problem. But over in hifi-land it's not uncommon to find designers who will casually discard basic design principles because they think their amp 'sounds better' without them. Virtually all amps have some form of feedback (the designs boasting 'No NFB' usually still have some local feedback around the individual stages), and almost all speakers feature some element of capacitative load, so they can be driven into ultrasonic instability if not properly compensated. There's a discussion here of how a common chip amp can become unstable and fry your tweeters if the circuit isn't properly designed.