PerfectWave combo vs PWT+Berkeley

[barrows]

Hi Emmo:

 

I think I have also explained my reasons re: degradation of sound quality by ASRCs in previous threads, herewith:

I previously worked for a high end audio company, and during the development process of our new DAC designs we tested many configurations. One of our tests was to determine, what if any effect an ASRC had on sound (as our engineers were suspicious of these devices). Typically ASRCs are employed in DACs to allow for re-clocking of the datastream, thereby rejecting a lot of incoming jitter. We built DACs that allowed us to listen with and without the ASRC with all other things remaining the same-our system also was very low in jitter to begin with, so it did not need the ASRC to combat jitter. In every case during these listening tests the sound was considerably better without the ASRC in the data path, and these differences were not subtle in nature.

The conclusion that we came to was that if one can keep jitter low in the first place, it is much better to not use an extra step of DSP (in this case the ASRC).

It is interesting to consider that many on these forums seem very concerned with the idea of "bit perfect"; it is important to realize that an ASRC resamples the data, and as such its output will not be "bit perfect".

Just to be complete, let me also say that I am not a big fan of NOS designs, as I believe the best digital filters should be used with sample rates much higher than 44.1, I just do not believe that an ASRC is the best way to up (or over) sample.

I am not an engineer, and I am not aware of any measurements which wil show the sonic problems caused by ASRCs, but the listening experience is clear enough for me to be looking for a DAC which does not employ an ASRC, and instead insures that jitter is low to begin with.

If one is using a high jittered source (adaptive USB, toslink output from a MAC, or perhaps even most SPDIF interfaces) one may achieve better sound quality by using a DAC with an ASRC. With a well designed low jitter digital interface (I2S, Firewire, Async USB, or perhaps clock linked SPDIF) Jitter rejection is not an issue, and a DAC without an ASRC is likely to sound best (all things remaining equal, ie everything matters).

 

[barrows]

Eloise,

 

While I have a tremendous amount of respect for Gordon Rankin, I have no understanding of his opinion on I2S for data transmission. In my experience it is clearly superior to SPDIF and AES as long as the length of the cable is 2 meters or less, the cable is good, and the sending and receiving devices are properly engineered. Objective and subjective testing of the PWT/PWD combo clearly demonstrates the superiority of the I2S connection over SPDIF/AES. Now if one is looking to interface with a computer transport, technically I believe Gordon's Async USB is one of the best ways to go, as it allows the DAC clock to become the master.

 

Clay,

 

I agree totally, well designed Firewire interfaces achieve superb performance, and although I have not yet experienced it, I suspect that Gordon Rankin's Async USB approach also perfroms superbly. The use of I2S as described is for digital connection between a disc transport and a DAC. With a computer transport I would expect a well implemented Firewire, Async USB, or networked connection to offer the best performance, as each of these interfaces allows for master clock control via the clock in the DAC.

 

In no way I am attempting to downgrade the perfromance of the BAD Alpha DAC, indeed I have not even heard this product in a system I am familar with and I look forward to doing so. I have also heard very good reports from people I trust on how good it sounds, and I look forward to hearing it with some familar components when I get the opportunity. I do think it is too bad that the only input the Alpha has is SPDIF/AES, as it would likely sound even better with a well implemented Firewire, Async USB, or network interface.

 

[icebreaker]

What were the designs of the prototype non- asynchronous DACs that you compared to the prototype asynchronous DAC?

Details of the sound system used to compare the prototype DACs would be helpful.

What is the name of the DAC your former company ended up designing?

 

Thanks,

 

[davidR]

wslam, Thanks for sharing your experience.

 

[emmodad]

a user named "barrows" was a frequent poster as employee, and referred to by other employees, on the PS Audio Forums until several months back.

 

[Audio_ELF]

The comments re i2s and distance, while stated by Gordon here, I have also read echod by dCS and other companies I don't recall currently. It's also (I think) in Phillips / Sony original spec. There were IIRC attempts to create a universal external i2s spec but I don't think anything came of it. If PS Audio have made it work then great - but it's not a universal panacea. AES and SPDIF have their problems and async USB and FireWire interfaces maybe better, but it's not as flawed as some would have us believe. At least there are methods for recovering errors which (is I understand correcty) i2s doesn't have - if the different line become unsynced (which is possible) you're is BIG trouble wrt to SQ, much more so than jitter with SPDIF.

 

Eloise

 

[barrows]

Eloise,

 

I cannot comment on what problems other products may have had using an I2S interface, I can only comment on the performance gains that PS Audio achieved by their implementation of I2S between the PWT/PWD products-the result of the I2S connection is a not insubstantial increase in sonic performance (listening tests) and an increase in measured performance as well. Anyone who owns these units can test this hypothesis for themselves as they are free to hook them up by single ended SPDIF or balanced AES, and I2S and make comparisons.

I cannot state it any more clearly: I believe that SPDIF is a flawed interface that results in higher levels of jitter. I am not suggesting that anyone should believe me, but the fact that audio designers such as Gordon Rankin, and Charles Hansen also believe that SPDIF is flawed appears to lend some weight to my opinion.

In any case, there are quite a few DACs around that sound very good connected by SPDIF, and I would expect those DACs to sound even better if they implemented an interface that allowed for single fixed frequency clock to be the master for the whole system.

Because this is a Computer Audiophile forum, I think it is important to point out that computer transports have a big advantage here: they can be connected to DACs that do not have to use an SPDIF interface, but instead can use Firewire, Async USB, or a network interface which when properly implemented is virtually jitter free. I think it is important for those who use computers as transports to realize that using these other interfaces is a big advantage for computer transports over typical CD/DVD transports.

Certainly, a poorly implemented I2S interface could have problems, just as a poorly implemented interface of any type could have problems. I have tried to be quite careful in my posts to use the term "properly implemented" when discussing the nature of different interfaces.

 

[barrows]

As it may be apparent at this point, I previously worked for PS Audio. I have not worked for PS Audio since February of this year, and I am not currently employed by any audio company. I am posting here as a lover of music and high end audio reproduction, and as an ordinary consumer as well. I do have some experience with audio product design that the average consumer does not have, and I am happy to share what I have learned here in the interest of trying to find out what might sound good, and what does sound good.

I have no interest in supporting any one audio companies products, anymore than audiophile would (of course I have my favorites, as anyone here might!). Chris and I have met on occasion, and I think he can vouch for my enthusiasm for audio, if he has any reason to suspect that my posting here is not appropriate in some way I trust that he will let me know.

 

[barrows]

I do not think I made any reference to testing any Async DACs? Am I mistaken? The testing I was referring to was I2S interfaces vs. SPDIF. I am hoping to get a QB-9 to try in my system soon, but Ayre is selling as many of these as they can make at this point, so it may be awhile.

The I2S vs. SPDIF, and ASRC testing was done via PS Audio's reference system in PS Audio's listening room with prototype Genesis loudspeakers, G Series and various prototype amplifiers, PS Audio cabling and power conditioning, and Boston Audio TuneBlocks isolation footers. Additional ASRC listening tests were done with show demo systems, at RMAF and CES, with the same components as above, excepting Avalon Indra speakers at RMAF, and Focus Audio Master 2.5 speakers at CES.

The differences I am referring to in my posts were large, and I would expect them to be evident on any well sorted high end system.

 

[icebreaker]

Which Asynchronous Sample Rate Converting DAC did you listen to and what implementation did you compare it to? Right now I'm guessing that it was the Burr-Brown asynchronous SRC 4192 that they use in the PS Audio Dig Link III vs. the Cullen mod?

From the details I can ascertain from your post it sounds like you only tested the BB SRC 4192 from the DL III.

It does not necessarily follow that if you don't like the sound of the Burr Brown SRC 4192 in the PS Audio Digital link III that all ASRC implementations are flawed as you contend.

 

[kana813]

"whatever the theoretical superiority the IS2 has over SPDIF, imho, it's overshadowed by the far superior algorithm of the Berkeley in this case."

 

I think all we know from your testing is filter 1 on the BADA, sounded better than filter 1 on the PWD.

 

The PWD has 5 filters:

1. Linear phase 'soft knee filter'

2. Minimum phase 'soft knee filter'

3. Linear phase Brickwall filter

4. Minimum phase apodizing filter

5. linear phase apodizing filter

 

Filters 2-5 might have produced a different result and saved your friend $2K.

 

 

 

 

[barrows]

We tested the 4192, and the new higher performance model from TI, it is their new top of the line ASRC that also has a built in SPDIF receiver, I cannot remember the number (maybe 4293?). We also listened to some Analog Devices parts as I recall. In any case, certainly some ASRCs are going to sound better than others, and we did not listen to every available option, not to mention that a clever designer could do their own math on a FPGA to make their own ASRC. In any case, an ASRC resamples the data, this means that the outgoing data is only an interpolation of the original data. Most of our engineers believed that no matter how good the math is, an interpolation is never going to be as good as the original data: as long as the ASRC is not needed for jitter reduction, their is no reason to use one. PS Audio used other methods to keep jitter low, so the PWD has the option to bypass the ASRC.

It is pretty easy for me, as an audiophile, to conclude that if one does not need an ASRC, there is no good reason to have one in the circuit, as it can only harm the signal, it cannot do anything to improve the original data.

 

 

[cfmsp]

 

 

Thanks for your balanced viewpoint, as always.

 

I really appreciate the thought that goes into your posts.

 

well said, on multiple posts in this thread.

 

clay

 

 

 

[Masis]

kana813,

that is not really true. We did play with each filter on the PS audio, but at the end, we settled on AUTO and NATIVE. We didn't manually use Filter 1. I clarified that already.

The 2k could not be saved. the difference was obvious.

 

[kana813]

"we settled on AUTO and NATIVE"

 

Than you didn't listen to PS Audio's favorite system filter.

 

From the PWD manual:

 

"The filter choices:

1. AUTO. This will automatically choose filter MP Soft for 44.1kHz and LP Soft for any higher sample rates. These are the choices we believe work best.

2. Filter 1 MP APOD. Minimum phase Apodising filter. Low pre-ringing, minimal group delay, minimized post ringing, good phase vs. frequency, sharper filter cutoff.

3. Filter 2 MP SOFT. Minimum phase soft knee filter. Low pre-ringing, minimal group delay, minimized post ringing, good phase vs. frequency, soft cutoff.

4. Filter 3 LP APOD. Linear phase Apodising filter. No group delay, perfect phase vs. frequency, minimal post ringing, some pre-ringing, sharper filter cutoff.

5. Filter 4 LP SOFT. Linear phase soft knee filter. No group delay, perfect phase vs. frequency, minimal post ringing, some pre-ringing, softer filter cutoff.

6. Filter 5 MP 1/2B. Minimum phase recursive Half Band symmetrical filter. Minimized pre and post ringing, good group delay, good phase vs. frequency response, sharp cutoff.

 

Our favorite on the PS system is Minimum Phase Apodising “MP Apod” on the touch screen."

 

[Masis]

let me repeat. we tried ALL filters. but for the most part, we used auto. none of the filter settings provide the same lushness, richness and smoothness of the Berkeley. it's a subjective call on our part for sure, but at least we did a AB test. Better than any beliefs with no justification.

 

[riderforever]

in your opinion, since you do have the Alpha DAC, how much is the sound quality outgoing from this device related to the transport/input chosen?

 

Do you think that a combo like Macbook --> toslink --> Alpha DAC will still be better than PC -> usb -> PWD/QB-9?

 

I mean, does the Alpha DAC still shine also when toslink and SPDIF connections are used, or do these harm considerably the sound?

 

Thanks for sharing

 

[Masis]

hi riderforever,

 

my ans is really as good as anyone's... it will still be a speculation. berkeley's manual did state that for the best result, the AES/EBU connection should be used.

 

I think there are two topics here. 1) which is the best interface/connection between a transport and a dac. 2) which dac inherently has the better sound.

 

I can only subjectively provide an answer for #2... which in this case is the Berkeley. But whether the difference between the connection will 'cost' berkeley the crown is not something i am in a position to say.

 

[riderforever]

with your opinion. Anyway from a purely theoretical point of view a DAC should have consistent level of performance, going from AES to coax or optical using the same transport, since these connections share the same transmission protocol and only the medium and connectors are different.

 

If the Berkeley sounds right with the AES connection I assume the designers have somehow addressed the flaws (if any) of the SPDIF protocol, also for the other inputs available.

 

A comparison of the inputs using the some transport would be surely more reliable, I don't know if Chris or anyone else has ever done it.

 

[kana813]

Let me repeat, here's what you said about the PWD's filters in two of your eariler posts:

 

"I wonder if the filter settings make a big difference on the PS DAC (we didn't try)."

 

"We used Native and Auto on the PS devices. I believe Native referred to the sampling rate, and Auto refers to the filter. I also tried Filter 1. Did not try the rest. "

 

Different subject- I'd love to see a screenshot of a TacT measurement of your Magico Mini IIs. I see you're a member of:

 

http://tech.groups.yahoo.com/group/TheRealTacTHackers/

 

You could post it there.

 

[Masis]

kana813,

we did settle on auto, but we did cycle through the filters when the sales was showing us the transport. we didn't spend a lot of time on any of the filer in a meaningful way except for Auto. speaking of which, why would PS not use their fav filter as the auto setting? i wonder what kind of rationale is behind that.

i moved recently and now have a dedicated room designed by rives. the tact has not been connected at all... i may not need to. i do have RTA measurements of the Mini II in my new room using FuzzMeasure.

 

 

[quid ergo es]

I happen to be an owner of both the PWT and PWD, also the Red Wine Audio Isabellina (a DAC designed by Vinnie Rossi). Very interestingly, the synergy of the PWT with the PWD (connected via I2S with a run-of-the-mill HDMI cable) does not better the PWT paired with the Isabellina (via optical cable). In fact, the signature of the two DACs is quite similar. The PS Audio piece is slightly more detailed and "drier". The Red Wine Audio piece is characterized by (subtly) warmer, denser sound. Both are equally enjoyable.

 

Of note, the Isabellina costs half what the PWD does. On the other hand, the Isabellina will not handle high-definition input, and it has zero capability to become a networked music server, as will the PWD in a few months.

 

[Fyper]

Thanks to vslam and the others for this interesting thread.

 

The PWT and PWD are still on my radar, the PWT more than the PWD, especially now since it seems that there are better alternatives to the combo.

Now the next step will be to compare the Transporter to the PWT regarding network streaming and interface when this is available.

 

[kana813]

wslam-

 

Look foward to seeing measurements when you get the TacT back in the system.

 

Spoke to a friend yesterday that's been listening to a Berkeley Alpha DAC for the last few weeks. He agrees that it's an excellent unit, but still feels the Nagra is the best DAC he's ever heard.

 

[Purite Audio]

I owned the Nagra dac a few years back, it wasn't that good, strident and a little mechanical imho, haven't heard the Berkeley yet though.