New JKDAC32 USB DAC

[jkeny]

Great stuff guys and appreciation to Ciaran for allowing the comparison. If the JKDAC was best it would be game over so the development continues. As Buzz Lightyear says "to infinity and beyond".
John, what Jplay settings were you using?

KM

I used kernel streaming with a 1 sample buffer (directLink)
Started with -6dB vol but this was the one that lacked dynamics & I changed it to no attenuation which sounded much better. This attenuation wasn't reducing any resolution, AFAIK - I was running 16 bit audio into 24bit DAC using a 32bit channel (Young driver).

Surprisingly, the Hiface driver was preferred over the Young driver so I'm scratching my head on this but think it might have to do with the very high resolution of Ciaran's playback system - being so clean, it doesn't sound sharp or strident in any way & so doesn't need the smoother treatment that the Young DAc seems to offer? This is my guess at what's going on.

[Ciaran]

I'm afraid I have to apologise for the absence of a coffee break! As so often happens, we were to caught up in the audiogeekery to have time for a break, and the afternoon seemed to fly by very quickly. Thank you all for coming along: it was a great pleasure!

John had to cope with my classical-only rule: I'd ripped a lot of tracks (losslessly to WAV files) and we were able to compare them with CD replay on the dCS stack.

We also had some hi-rez files we were able to compare with SACD.

For me the performance of the JKDAC got much closer to the dCS stack once it was fed hi rez files (24/88.2 or above). We didn't compare any 16/44.1 files via the JKDAC and dCS after John had fixed the 6dB attenuation in his player, so we don't know if that would have made replay at that resolution better or not. We should probably have been more methodical and checked that.

Did you feed the DCS using Jplay from PC?

No, the dCS was not fed from John's laptop: only from the Verdi la Scala transport or in one case from the Squeezebox Touch, streaming from my computer. The comparisons with the Touch were JKDAC playing the same 24/88.2 file as the Touch and the stack playing the equivalent track from a standard CD. In this case I thought honours were divided about evenly between the JKDAC and the stack, with the Touch objectively fine but subjectively less involving.

Very interesting! I'm sure there are big strides to be made in computer audio, probably in the short term, but my investigations highlighted for me the dearth of available hi-rez material. I would want to see more of this before I got serious about computer audio.

[Fran]

Yeah - I can't ever see widespread availability of hi-res material. Maybe, just maybe for the odd special release, but for anything prior to the last few years, the only way something like this would become available is if they got their hands on the master tapes and transferred from them in hi-res.

So if using a PC of source is to become the way of the future, then it must be able to do 16/44 well. Now, yesterdays experiment (and previous experience) suggests to me that the PC route equals or beats the higher end players out there, only falling short when compared to the very best available. It seems likely that if a few small improvements more are made, then it could be very close indeed.

Something else in the bigger picture that needs to be considered: how long will CDs be available? There are already numerous rumours spreading around the web that the "majors" are planning on dropping CD releases by the end of next year..... So will we the choice between the download route or vinyl route be the only one open in the future?!

Anyway, I've drifted OT a bit......

Ken/John/Sligolad - what settings in jplay are you using?


Fran

[jkeny]

I'm afraid I have to apologise for the absence of a coffee break! As so often happens, we were to caught up in the audiogeekery to have time for a break, and the afternoon seemed to fly by very quickly. Thank you all for coming along: it was a great pleasure!

Don't worry, Ciaran, as you said the time flew by without notice.

John had to cope with my classical-only rule: I'd ripped a lot of tracks (losslessly to WAV files) and we were able to compare them with CD replay on the dCS stack.

We also had some hi-rez files we were able to compare with SACD.

Thanks for doing this - classical is not my area of listening as evidenced when asked what track I was playing, all I could say was it was Christmas Mass :)

For me the performance of the JKDAC got much closer to the dCS stack once it was fed hi rez files (24/88.2 or above). We didn't compare any 16/44.1 files via the JKDAC and dCS after John had fixed the 6dB attenuation in his player, so we don't know if that would have made replay at that resolution better or not. We should probably have been more methodical and checked that.

Yes, the volume issue was a real slip up on my part but the day threw up a number of conundrums & the biggest one revolves around why the vol change could make such a difference - it doesn't do this on any other system, I've heard? Ciaran, does your system show this characteristic normally i.e if the system volume is turned down a bit that it loses a lot of attack, dynamics & ambience? This vol issue also seemed to mask the difference between Foobar & Jplay so again it wasn't just a matter of the sound pressure being lower it was masking something! Just to explain again, I had the digital vol set at -6dB which is dropping 1bit off the signal. AFAIK, this shouldn't eat into any resolution of the sound as if playing back 16 bit material using a 24 bit DAC, in theory I should be able to drop 8 bits before I'm effecting the information carrying bits. I need to investigate this further but as I said it does not do this on other systems. Any ideas?

]Did you feed the DCS using Jplay from PC?

No, the dCS was not fed from John's laptop: only from the Verdi la Scala transport or in one case from the Squeezebox Touch, streaming from my computer. The comparisons with the Touch were JKDAC playing the same 24/88.2 file as the Touch and the stack playing the equivalent track from a standard CD. In this case I thought honours were divided about evenly between the JKDAC and the stack, with the Touch objectively fine but subjectively less involving.

Yes, this is the elusive involvement factor - what is it in the sound that gives this? Anyway, I felt the JKDAC & Touch needed A/B listening to sort out, they were that close.

Very interesting! I'm sure there are big strides to be made in computer audio, probably in the short term, but my investigations highlighted for me the dearth of available hi-rez material. I would want to see more of this before I got serious about computer audio.

Yes but I don't think it's only in high-res that computer audio has a role - it is approaching the best of spinning disk playback at any resolution. What the dCs system showed is that 16/44 playback can sound as good as high-res if done with great care & that it really comes down to the recording!

Thanks again Ciaran for affording me the opportunity to have a listen to your wonderful system - it gives me a great benchmark.

[DaveF]

So if using a PC of source is to become the way of the future, then it must be able to do 16/44 well. Now, yesterdays experiment (and previous experience) suggests to me that the PC route equals or beats the higher end players out there, only falling short when compared to the very best available. It seems likely that if a few small improvements more are made, then it could be very close indeed.

I think the PC front end of things is just a bit too messy. You have the OS, drivers and a lot of other stuff in there that isnt needed nor is optimised for audio playback. I'd like to see the current CDP units being the same but with the optical disc mechanism being replaced with hard disk or internet capability. Perhaps with the SW management of tracks being moved onto an ipad tablet type thing.

[jkeny]

So if using a PC of source is to become the way of the future, then it must be able to do 16/44 well. Now, yesterdays experiment (and previous experience) suggests to me that the PC route equals or beats the higher end players out there, only falling short when compared to the very best available. It seems likely that if a few small improvements more are made, then it could be very close indeed.

I think the PC front end of things is just a bit too messy. You have the OS, drivers and a lot of other stuff in there that isnt needed nor is optimised for audio playback. I'd like to see the current CDP units being the same but with the optical disc mechanism being replaced with hard disk or internet capability. Perhaps with the SW management of tracks being moved onto an ipad tablet type thing.

I agree somewhat with Dave & we talked briefly about this yesterday - there are many variables in a general purpose machine like a computer that need consideration. Are there too many? Time will tell? It is a testament to the ingenuity of many that the PC can provide such high levels of audio reproduction given that it has not been designed for this purpose.

At the moment the use of the PC for audio playback is an attractive offering - it avoids the expense & inconvenience of buying another single purpose device solely for audio playback. I believe there are probably two camps - one that wants the almost limitless flexibility of the PC which affords them the potential to tweak & improve their audio AND the other camp who want a fixed solution without any required tweaking needed. Of course the second camp can just buy a pre-built solution that somebody has already tweaked :)

[jkeny]

Fran,
I used kernel streaming, 1 sample buffer (DirectLink), throttle mode on & I usually use -6dB vol attenuation (but eventually not yesterday)

Edit: One other thing I noticed in my Jplay set-up - I was using the River engine yesterday!

[Ken Moreland]

Here are my Jplay settings,

Version 34
Filecache 4000 4Gb of RAM secured for Jplay
Buffer 1 Directlink, this is important for best sound
Clock 0 Fastest clock
Engine 1 The engine is Beach , reputed to be more resolving and better than River
Interface 1 KS streaming using Young driver (can also use Wasapi)
CDBitstream 24
HiRezBitstream 24
Overdrive 1 Pushes the processor
DedicatedCore 1
USBmode 0
Volume 0
PlaybackOrder 0
Hibernate on PC shuts everything else down , this makes a major difference.
ThrottleMode 1 Throttle on
IMHO key settings are Lowest buffer, Hibernate on , Beach.,
I prefer the Young driver but if you disconnect the Hiface ,on reconnection it will revert to the Hiface driver and you need to set it up with the Young driver again.


Personally I find an overlap in sound quality between CD and Hirez with not much to choose between the best of both and a well recorded CD as good as any Hirez.

HDtracks have a good selection of studio masters available and 24bit96khz.org(if you have strong morals look away now) have over 200 studio masters for free download including a lot of classical.

I bought a 2Tb HDD last week, Maplins had put prices up by 100% because of the floods in Thailand but Harvey Normans next door in sunny Blanchardstown had reduced the price from 179euro to 99euro!! Go figure as they say.



KM

[Fran]

Thanks guys.... Have either of you compared beach and river directly?

[Ken Moreland]

For me Beach is clear , detailed, precise but River is warmer, softer but it's not all cut and dried, even one of the developers prefers River. Here's a link from the Jplay forum http://jplay.eu/forum/jplay/beach-vs-river/
Similarly with KS v Wasapi it's down to personal preferences and equipment.

KM