Review and Measurements of PS Audio PerfectWave DirectStream DAC

[typericey]

May I also suggest a new award? The 'Shattered Panther'. Maybe a single paw left. Or just a bit of a tail. And a pile of rubble. You really do need something better than the 'Decapitated Panther' for this one...

I disagree. #SaveThePanthers. In fact, would be nice if Amir would glue his head back on.

Btw, I now have an idea on @amirm's methodology for his reviews:
Step 1: take measurements
Step 2: write review
Step 3: take product photo with appropriate Panther

 

[audimus]

No. My testing was blind. You can have all the knowledge of measurements and still can't game the system.

That it was blind was not mentioned in the description. How blind? Did someone else control the switching to randomly select one or the other and not just keep alternating between the two? Otherwise, the consecutive tests were not independent and so detection of a problem in one test track would influence the next track.

I am not trying to deny that you heard the difference. Obviously, there is a threshold beyond which one can clearly hear a difference. But this is a very slippery slope, one I do believe we have to take to address the gap between measurement and hearing as I have ranted before. So, I am glad you included that part.

BUT, as described it reads like any flawed test anybody could have done that would have been swiftly dismissed. This is what prompts people like Ted to claim (perhaps incoherently) that you would not hear a qualitative difference even if one exists in some units because it measured well and you heard it here to validate your own measurements.

So, perhaps documenting the listening procedure elsewhere once (so as not to repeat for each device tested) would bring more credibility by being verifiable.

 

[Wolven]

Ted’s argument reminds me of the arguments from climate-change deniers trying to use “science” to argue. Using terms that sounds like science but with no understanding of what they mean.

You have that in reverse.

 

[phoenixdogfan]

Maybe dude is onto something. Then again, I find it strange that the council overseeing fundamentalism here at ASR could be wrong after they all agreed with your findings.

View attachment 34208

What could be more fundamentalist as an audiophile to believe a component performs as it measures? If it does not, isn't it incumbent on the alternative audiophile believers to specify the recipe for their "secret sauce"?

 

[BillG]

Oof. Ouch. The number of super high end DACs shown to have simply flawed engineering is amazing.

And this is why I personally avoid boutique brands, celebrity amp and speaker designers, and the like. They appear to be running a confidence game for the most part...

 

[phoenixdogfan]

And this is why I personally avoid boutique brands, celebrity amp and speaker designers, and the like. They appear to be running a confidence game for the most part...

Also hilarious how they shit on obviously superior, high value stuff like ESS chips, by talking about their "characteristic house sound." Talk about complete horse shit.

 

[Herbert]

BTW, don't ESS Chips throw away the
clock from the source as well and generate it from the datastream?
I might be wrong here...

 

[BillG]

by talking about their "characteristic house sound."

Perhaps this is what Paul is alluding to when he's going on about "musicality" and "voicing". Yeah, I never bought it either...

 

[HammerSandwich]

Also hilarious how they shit on obviously superior, high value stuff like ESS chips, by talking about their "characteristic house sound." Talk about complete horse shit.

Clearly poop if the device sounds so different after a firmware update. Which house sound should I expect from their next product?

 

[Sal1950]

These companies need to hire trained listeners and perform a simple level matched test before spending so much money on engineering and producing subpar products.

I'm not so sure they didn't as opposed to doing things in a purposeful manner to achieve a "house sound". For better or worse it's a good thing when the devices sound signature can be easily determined thru listening. Use of output transformers almost guarantees that.
In "high end" reporting media much of the subjective results will be influenced by the amount spent on advertising. PS spends a lot.
All this yapping over firmware versions misses the point that any manufacturer with a decent engineering department would never let a product out the door with a firmware version performing like this one did. Firmware updates can be expected to make very minor performance improvements and maybe a bug fix or two in the UI, not turn a sows ear into a silk purse (or vise versa )

 

[amirm]

Can you please do a quick FFT at 50Hz like JA did, just for peace of mind.

Good suggestion. Here you go:

So it seems they have improved the product since JA tested it. Wonder if they had some clipping that caused the high harmonic distortion that JA measured.

Imagine the number of limbs the Pink Panther would have lost if I had tested what JA was given!

 

[amirm]

Btw, I now have an idea on @amirm's methodology for his reviews:
Step 1: take measurements
Step 2: write review
Step 3: take product photo with appropriate Panther

Almost. It is 1, 3 and 2.

 

[amirm]

BUT, as described it reads like any flawed test anybody could have done that would have been swiftly dismissed. This is what prompts people like Ted to claim (perhaps incoherently) that you would not hear a qualitative difference even if one exists in some units because it measured well and you heard it here to validate your own measurements.

Ted? You mean the guys who don't use proper testing themselves would ask why I didn't do that? It is like saying an Atheist wants a Christian to swear on a bible before he believes what he says.

I don't document everything in the world in my reviews. You should have noticed that. If there is ambiguity, just ask. Don't make assumptions and then run with them saying results are not valid, etc. This is an interactive endeavour. As I type this, I have been testing another product for review that I will post tomorrow. I can't write a bible for you all.

Also, I am not positioning my listening tests as solid evidence for people in our camp. It is there to appease the people from the other side who say "he doesn't listen." As far as they are concerned, I could have done the test any which way and it would still been valid! I can't bring myself to do such test that poorly so I match levels, and perform blind AB tests. The number of trials is low so doesn't have statistical rigor. Nor does it have a control, etc. So if you are in our camp, take the listening tests as minor supporting fact.

 

[amirm]

Perhaps this is what Paul is alluding to when he's going on about "musicality" and "voicing". Yeah, I never bought it either...

There is no voicing going on. No house sound. A transformer was put in there which added distortion, exaggerated highs, etc. They liked it without performing a proper, controlled AB test. Happens all the time. Faulty experiments result in expensive audio products being sold from cables to boxes with dirt in them....

 

[amirm]

BTW, don't ESS Chips throw away the
clock from the source as well and generate it from the datastream?
I might be wrong here...

They do. They have an asynchronous sample rate converter.

 

[pozz]

That it was blind was not mentioned in the description. How blind? Did someone else control the switching to randomly select one or the other and not just keep alternating between the two? Otherwise, the consecutive tests were not independent and so detection of a problem in one test track would influence the next track.

I am not trying to deny that you heard the difference. Obviously, there is a threshold beyond which one can clearly hear a difference. But this is a very slippery slope, one I do believe we have to take to address the gap between measurement and hearing as I have ranted before. So, I am glad you included that part.

BUT, as described it reads like any flawed test anybody could have done that would have been swiftly dismissed. This is what prompts people like Ted to claim (perhaps incoherently) that you would not hear a qualitative difference even if one exists in some units because it measured well and you heard it here to validate your own measurements.

So, perhaps documenting the listening procedure elsewhere once (so as not to repeat for each device tested) would bring more credibility by being verifiable.

The most frustrating thing here is that the manufacturers are dodging the confrontation. The listening test doesn't rank high in the reason why they've taken notice, though.

 

[maxxevv]

I'm fairly sure than anything from dCS will measure superbly, like the top Chord stuff, but the issue ia absurd price for something not audibly better than good products in the low $hundreds range (for humans anyway).

I guess the real problem with Chord is the pricing bracket they target and their choice of cozying up with subjectivist mysticism.

But as previous Chord devices measured here have shown, they all seem to measure and perform quite well at least for their generation of technical implementation. The current generation product, the Qutest actually still ranks quite high up in the range of DAC's measured here.

How good are they "audibly" is subjective too since we all hear them differently. Not forgetting that dynamic loads and static measurement loads can yield different results, be it slight or dramatic. It definitely bears out for my area of work with regards to mechanical implementations, so I'm confident they are analogous with electronics too.
There was a graph from @amirm for the THX789 that somehow showed the THD+N to swing once connected to a headphone or something.
Its something that's quite hard to consistently duplicate or measure, so perhaps its one aspect of the testing that ASR could put everyone's expertise together to devise a suitable test rig for such dynamic loads.

I have heard mumblings from subjectivists with regards to the THX 789's sound signature being pretty 'dull / dry / thin' but yet gush profusely over the JDS Atom's. Both being excellent, if not exceptional measuring headphone amps here on ASR. Which is quite puzzling if you ask me and the only postulation I can attribute it to is their varied response to dynamic versus static loads.

I do like the Atom which I have but have yet to do a A/B with the THX 789.

I have tried both the Qutest, Dave+Blu MkII and the Hugo2. They sound really nice to my ears. Only issue I have with them is the price they are asking wrt to the various other well measuring devices that have been uncovered here on ASR.

 

[Sal1950]

There is no voicing going on. No house sound. A transformer was put in there which added distortion, exaggerated highs, etc. They liked it without performing a proper, controlled AB test.

Isn't that generating a house sound, something they like?
All a manner of perspective I guess.

 

[amirm]

Isn't that generating a house sound, something they like?

No. It is a terrible accident.

They don't know what that sound is. The only way to know is to conduct an AB test and then, they could tell me if they like the sound. I suspect they would say no as no one wants to advertise liking more highs and screachiness, less bass impact, etc.

 

[majingotan]

DAVE could reach APx555 limit.

We'll see. Hope it is since Chord touts their measurements so much in their marketing. Chances are it might not even beat the current king Matrix X-Sabre MQA DAC since Chord is custom FPGA implemented. Based on empirical data, those who know how to truly maximize off-the-shelf DAC chips are the ones who measure extremely well. Any further complications/implementation just deviate from the absolute transparency. That's also why I'm very curious with the Linn Klimax DSM which probably costs about 1.5x DAVE but it uses an off-the-shelf AK4497EQ DAC chip