Review and Measurements of PS Audio PerfectWave DirectStream DAC

[carlosmante]

FPGA vs R2R. So Totaldac definitely sounds more analog. Because I say so. Still, costing twice as cheap, and not showing ground loop, the Perfect Wave has the best value, IMHO. I would say, save money and go for the PerfectWave! But don't forget to burn in your unit 1000 hours, at least.

The right way is to use a wood-fired Adobe oven.

 

[amirm]

I posted Amir's review on Canuck Audio Mart (I'm Hi Fi Nut there); some predictable responses.

https://www.canuckaudiomart.com/forum/viewtopic.php?f=11&t=54647&p=887618#p887618

"Amir is an imbecile. Don't trust anyone's ears but your own. I certainly do not trust 60 year old year men ears who cannot hear above 8khz."

Let him know this is Ted Smith (on the right), the designer of DirectStream DAC:

He and Paul have more gray hairs than I do. If he doesn't trust my ears, he better be very scared of what Ted and Paul's can't hear...

 

[Milt]

Late to the show as always, but did anyone here expect a different reaction to the measurements?
To me, PSAudio is the equivalent of Audioquest.
Pay a shitload for gear that can be demonstrably bettered at 1/4 of the price at best.
Edit: Sorry. Corrected

 

[GrimSurfer]

Looks like Ted Kaczynski has been released on an ADMAX Colorado prison work furlough under the recognizance of Paul McGowan.

If so, we can expect to see future PS Audio components to be released in non ferrous chassis and come with a rambling, hand written, owner's manual featuring all sorts of life tips.

Psst Ted - Jimmy Buffett called. He wants his shirt back, preferably not by mail.

 

[Canuck57]

he sure looks over 60 to me!

 

[mkawa]

3 orders of magnitude off. the 8$ apple usb-c to 3.5mm outperforms it in every way, including output levels

 

[audimus]

Late to the show as always, but did anyone here expect a different reaction to the measurements?
To me, PSAudio is the equivalent of Audioquest.
Pay a shitload for gear that can be demonstrably bettered at 1/4 of the price at best.
Edit: Sorry. Corrected

I think it just confirmed how unprofessional and unethical the principal of this company is in response to valid criticism for people who may not have had an opinion before.

What PS Audio has to fear are not the cultists who are hitting out publicly in those forums but the silent readers that might decide their patronage is misplaced based on the exchanges and go away.

Those reactions are no different from some of the reactions to the NAD measurements. Fortunately, none have come here to troll... as yet.

 

[GrimSurfer]

Word of mouth is crucial in the direct sales business.

 

[Rja4000]

And I still don't have the answer to my question:

Does the unbalanced output also involve the transformer?
I guess yes, but I'd like to be sure.

Or did I miss the answer?

 

[Ron Texas]

That is very true. But it is also true that many recordings just do sound harsh, some because of bad miking/mastering and some because many instruments do sound that way live. Audiophiles gobble up gear that tends to smooth those ruff edges and critize gear that is transparent and accurate.
"They can't handle the truth" LOL

So, what should we do? Buy some expensive crap which cuts off the harshness, but makes well made recordings sound dull? It would be nice if we could press a button and get it right. I suppose we can do it with DSP, although it takes more than one button. Perhaps that's why some DAC's have multiple filters, with some of them producing measured output which is worse.

 

[Ron Texas]

he sure looks over 60 to me!

Oh heck, I am ancient and I look a lot better than either of those guys. OK, they play with soldering guns and pencils are my thing...

 

[amirm]

And I still don't have the answer to my question:

Does the unbalanced output also involve the transformer?

It would have to because the primary of the transformer is the reconstruction filter. I will test to be sure but I suspect unbalanced is derived from the balanced output of the transformer.

 

[Rja4000]

It would have to because the primary of the transformer is the reconstruction filter. I will test to be sure but I suspect unbalanced is derived from the balanced output of the transformer.

Thanks.
I know transformer was a way to achieve balanced out in pro audio, so that's why I ask.
But if they use it to somehow add an effect to the sound, they probably use it in both cases.

 

[Stephensank]

As much more of an expert on signal transformers than most, I'd like to weigh in here.
What the measurements & audible results show is not *endemic* to using a transformer. Rather, it is primarily a result of using a VERY poorly selected and/or shockingly bad transformer, as far as specs and/or quality. It is clear to me from the graphs that there is a very bad mismatch between the transformer chosen and the capabilities of the circuit driving it. Even if it were a Jensen transformer(which I consider *second* best current maker, behind Cinemag), if it had same or similar specs to whatever cheaper unit was used, the results would be similarly shockingly bad. Problems can be further exacerbated by mismatch on secondary side. So, if, say, the secondary Z is 600ohms, it would be quite meaningful to run tests with various loads(assuming I am correctly reading that the transformer is the final output of the dac), from 600ohms to several steps up to 10 or 15kohms. But I do think this would simply confirm that the main issue is a drastic mismatch between driving circuitry & transformer primary.
The reviewer gives a perhaps unintended impression that transformers are always bad in a signal path. IMO, this is very, very untrue. When done properly, using a transformer is quite often the superior way to go, e.g., for balanced input of preamp/amp, it's rather hard to design a transformerless differential input that doesn't color the sound more than a well-selected transformer. Opinions may vary wildly, but, in my >40 years of pro experience, this has been well demonstrated in practice.

 

[Cahudson42]

Let's put an Apple dongle driving a Dynaco ST70. Then setup a blind A/B test against the DirectStream - level matched with an 'objectivist' amp.. On a pair of AR-3s. Ask the DS proponents/buyers/designers to pick....Bet it's 50/50..

These guys certainly look like they grew up with ST70s. And were trained to love it..

 

[solderdude]

the primary of the transformer is the reconstruction filter.

The reconstruction filter is a very, very steep and well designed linear phase filter it would seem from the test results.
I suspect you meant the transformer is part of the post filtering circuit.

 

[maxxevv]

"They can't handle the truth" LOL

Ooh... that sounds so much like Jack Nicholas's famous line in A Few Good Men.

 

[Guermantes]

There is a tendency here to jump to a conclusion that because there is a measurement that is made here that isn’t aligned with some engineering goal, that it must be necessarily bad sounding or that people who like its sound are either misguided or delusional from various biases. This is true at the extremes but gets fuzzier in the middle.

This is an invalid inference from the measurements not because there aren’t such misguided or delusional people (there are plenty) or that markedly bad measuring equipment will sound bad (obviously true) but because, as I have pointed out, the measurement is unable to authoritatively distinguish between components of the metric that will result in sound that will be judged bad, components (of the metric) that will be unheard by most and components (of the metric) that will sound pleasing or otherwise compensate for people’s own uneven hearing characteristics that deviate from a machine analysis. One measurement metric may include all of that in a cumulative way.

While I am sympathetic to your points regarding our lack of understanding about how the measurements relate to psychological outcomes, the last point regarding deviations of hearing characteristics is difficult to engage with. How would any engineer (or reviewer for that matter) be able to predict, measure or compensate for an unknown like that? Better to provide an interactive tool such as EQ and let the listener adjust it for their own abnormalities or idiosyncrasies.

But here is Rupert Neve, an engineer I have a lot of respect for, describing the resonance produced by his transformer-based designs as a "sweet" and "musical" sound (17:40). He also discusses the subjective "musicality" of his EQ designs and how he prefers the addition of a small amount of 2nd and 3rd harmonic distortion (20:00). And, yes, he's over 60

 

[Thomas savage]

"Amir is an imbecile. Don't trust anyone's ears but your own. I certainly do not trust 60 year old year men ears who cannot hear above 8khz."

Let him know this is Ted Smith (on the right), the designer of DirectStream DAC:

He and Paul have more gray hairs than I do. If he doesn't trust my ears, he better be very scared of what Ted and Paul's can't hear...

But your a very handsome imbecile Amirm!

I don't trust the guy on the left he looks like a insurence salesman but the guy on the right ,, well he obviously shops in the same shirt store as Bruno Putzeys but I fear that's where the similarities end with him.

 

[JJB70]

The standard response of boutique audio companies or companies with pretensions to be part of the high end micro-bubble is to shoot the messenger if anyone dares point to the emperors lack of clothes. You can't argue with measurement unless you are going to argue about how the measurements were taken or in some cases the implications of uncertainty, repeatability etc (and I've done enough measurement in my own field to know that there can be entirely legitimate arguments about those aspects), in the case of audio it is very clear and measurements do tell the truth. Given that the whole viability of these companies depends on maintaining the voodoo mysticism around audio, built on the usual florid language and group think any attempt to introduce measurement and objectivity is an existential threat. Hence the behaviours of certain companies when faced with evidence that they're asking people to pay $$$$$$$$$$$$$$$$$$$s for crap (eg. PS, TotalDAC, Schiit). In some ways it can be entertaining and good fun in that watching a train wreck way to see such people grab a rope and use it to hang themselves but sometimes they can get rather nasty.