PS Audio DS DAC Reviewed in HiFi News - sounds good, measures badly.

[andyv]

There's a very interesting review of the PS Audio DS DAC in this months HiFi News. John Bamford was full of praise for the sound quality, saying, amongst other things, that it was "the best DSD I've heard". But Paul Miller's measurements revealed some issues. For example the realisable frequency response with 24/192 was barely better than that achieved with 24/96 (though it might not matter much in practice), "the impulse response shows extended pre/post ripples in the time domain" and "an increase in noise through the audioband" which reduces the "S/N ratio to a '16-bit' 96dB", and which makes it hard to measure the jitter. Distortion is "20-30dB higher over the top 20dB of it's dynamic range than typical" and "distortion is very high indeed at low (<100Hz) frequencies reaching a full 3% at 20Hz". Paul Miller says that this could cause a sensitivity to the partnering amplifier.

 

These do seem rather poor measurements compared to just about any other DAC Paul Miller has measured recently. The noise figures are poor, and the distortion at low frequencies is appalling. All his results are available online through the HiFi News website, as is the magazine.

[CoolHand]

Those measurements aren't so bad...heck it's almost approaching those of a basic record player, which is perhaps why it sounds half decent after all is hyped and done :^)

[EuroDriver]

This is a often a trait with DAC's that do PCM to DSD conversion. The Playback Designs DAC's sound very nice but do not measure well.

 

what do you want in playback, great measurements, or great sound ?

 

it would be interesting to see how software conversion done with HQ Player and then played by a DSD256 DAC measures compared to DirectStream and others

[ErikM]

Maybe Ted will respond.

 

 

This is a often a trait with DAC's that do PCM to DSD conversion. The Playback Designs DAC's sound very nice but do not measure well.

 

what do you want in playback, great measurements, or great sound ?

 

it would be interesting to see how software conversion done with HQ Player and then played by a DSD256 DAC measures compared to DirectStream and others

[bmoura]

Looked at that web site. Seems pretty dated. The last equipment reviews posted are from November 2013. So it may take awhile to read their review.

[andyv]

Looked at that web site. Seems pretty dated. The last equipment reviews posted are from November 2013. So it may take awhile to read their review.

 

You can read the review I mentioned right now by either buying a hard copy of the magazine, or by buying a download copy through the apple store or through the HiFi News wesbsite. This has half a page of measurements and comment on those measurements, including graphs of distortion (very bad), noise measurements (bad), jitter (hard to tell because of the noise), etc.

 

If you register with the website you can get free access to very detailed data from all tests - pages and pages of measurements for every product. These data are published online a few months after the summary in the magazine.

[christopher3393]

Sorry ,not clear to me, do I need to pay to read this review? Unless this can be made public now at a reasonable cost, I don't see how it is relevant.

[weff]

Sorry ,not clear to me, do I need to pay to read this review? Unless this can be made public now at a reasonable cost, I don't see how it is relevant.

 

It's a case of pay to read either the download copy from Apple Store - or a hardcopy of the magazine.

 

As a non-subscriber the available review data is only from 2012 and earlier.

 

I'm surprised that a detailed test review could even be prepared by this point - what with magazine lead times being 1-2 months.

They certainly can't have tested with the latest DAC firmware - and I would have to wonder what burn in they did prior to the review ? Though what difference either of those factors would make to the measurements ?

[weff]

It's a case of pay to read either the download copy from Apple Store - or a hardcopy of the magazine.

 

As a non-subscriber the available review data is only from 2012 and earlier.

 

I'm surprised that a detailed test review could even be prepared by this point - what with magazine lead times being 1-2 months.

They certainly can't have tested with the latest DAC firmware - and I would have to wonder what burn in they did prior to the review ? Though what difference either of those factors would make to the measurements ?

 

Correction to the above - As a non-subscriber I can see technical data up until Feb 2014.

 

Despite not being able to purchase the Aug 2014 Hi Fi News edition by itself - I did come across a pdf in Google - although not having access to the full data report it is hard to make proper sense of what the limited test data in their review copy means.

[andyv]

Sorry ,not clear to me, do I need to pay to read this review? Unless this can be made public now at a reasonable cost, I don't see how it is relevant.

 

The magazine costs £4.00 an issue.

[pooger]

Since I'm a big believer that testing is only based on current assumptions of how we hear, I've never given any tests more (or even equal) value over what I hear. Those of us old enough to remember "perfect sound forever" --how early CDs tested absolutely distortion free in almost everything (until engineers that trusted their ears began looking into digital domain distortions, that really hadn't been tested before)-- know that testing regimens do have some value, but must be taken with a grain of salt.

[kana813]

Wonder if Hi Fi News had the latest firmware installed.

 

It will be interesting to see JA's measurements of the DirectStream DAC the Art Dudley is reviewing.

[k6davis]

Since I'm a big believer that testing is only based on current assumptions of how we hear, I've never given any tests more (or even equal) value over what I hear. Those of us old enough to remember "perfect sound forever" --how early CDs tested absolutely distortion free in almost everything (until engineers that trusted their ears began looking into digital domain distortions, that really hadn't been tested before)-- know that testing regimens do have some value, but must be taken with a grain of salt.

 

+10