Audirvana Plus 2.5.1 with PCM->DSD Conversion

[damien78]

Just released new beta 2.5.0.9

Fixing the Mytek not playing (and crashing) at DSD128, and the noise at 5s issues

[Le Concombre Masqué]

unfortunately, my speakers which were top references for studio monitoring and top of the line at Radio France until the 90's have no 20k plus region

Real life experience on various systems and setups. Apparently there is spatial information in the 20k plus region. Even for people with medior or senior ears.. This is also the theory behind super tweeters.

[skatbelt]

unfortunately, my speakers which were top references for studio monitoring and top of the line at Radio France until the 90's have no 20k plus region[ATTACH=CONFIG]25745[/ATTACH]

 

Hence your advice of 0,95. Actually, if I read the diagrams correctly, your speakers are doing 30k at -10dB.

[Le Concombre Masqué]

Damien, I will elaborate later but I don't feel the DSD 64 playing is on par with the PCM to DSD 128 conversion and that you should consider upsampling DSD 64 to higher resolutions as well

[Jud]

Real life experience on various systems and setups. Apparently there is spatial information in the 20k plus region. Even for people with medior or senior ears.. This is also the theory behind super tweeters.

 

Many so-called "apodizing" filters that many people like begin rolloff well short of Nyquist.

 

In my own system and with my hearing, I find the A+ adjustment to which I am most sensitive in terms of spatial cues is phase (called "pre-ringing").

 

This is not to say in any way you are somehow not correct in hearing what you do (after all, you hear it), but simply to suggest that different ears and different systems can react so differently to various settings that it may be The Masked Cucumber's setting of .95 Nyquist doesn't mess up his spatial cues, while it certainly would for you. (For me .95 Nyquist would roll off the top end too much. I can hear up to about 16KHz, so I guess the cut must start at a frequency below that - actually, substantially below that, since most musical content is well below 16KHz.)

[YashN]

The last beta and the one before it five me a beach ball when changing tracks, so it's hard to test for long.

 

However, for the little we heard doing realtime PCM to DSD128, it sounded very good indeed.

 

Not sure why it freezes every time I change tracks - this is on Mac OS X Mavericks WITH 4Gb of RAM.

[damien78]

The last beta and the one before it five me a beach ball when changing tracks, so it's hard to test for long.

 

However, for the little we heard doing realtime PCM to DSD128, it sounded very good indeed.

 

Not sure why it freezes every time I change tracks - this is on Mac OS X Mavericks WITH 4Gb of RAM.

 

With 4GB RAM, be sure to have the audio buffers pre-allocated memory be no more than 1.2GB

[Le Concombre Masqué]

no roll off before 20k

Many so-called "apodizing" filters that many people like begin rolloff well short of Nyquist.

 

In my own system and with my hearing, I find the A+ adjustment to which I am most sensitive in terms of spatial cues is phase (called "pre-ringing").

 

This is not to say in any way you are somehow not correct in hearing what you do (after all, you hear it), but simply to suggest that different ears and different systems can react so differently to various settings that it may be The Masked Cucumber's setting of .95 Nyquist doesn't mess up his spatial cues, while it certainly would for you. (For me .95 Nyquist would roll off the top end too much. I can hear up to about 16KHz, so I guess the cut must start at a frequency below that - actually, substantially below that, since most musical content is well below 16KHz.)

[Rondark]

I have run the last two betas on Wyred4sound DAC 2 DSDse and everything is working fine at DSD128. Pretty much plug and play.

[Lee Henley]

Tested this afternoon with a friends Lampizator Golden Gate, my Mac is the old type mini from 2009 and only has 2gb ram memory, I had to turn the memory allocation down to the minimum amount and turned off direct mode, once this was done the Mac played PCM to DSD without any issues at all.

 

Know onto the important part, how did it sound? Bloody awesome, so much so it's got my friend thinking of buying a 2012 mini and installing A+, I've heard his system many times and to date this was the best it's sounded by a country mile

 

As for me, well I'm gonna have to save my Pennines and get myself a Lampizator Dac to playback native DSD, probably a Lite 7 model or maybe push to a Big 7

 

A big thumbs up for Damien for providing this feature in A+ it's a remarkable achievement

 

Lee

[orgel]

Just released new beta 2.5.0.9

Fixing the Mytek not playing (and crashing) at DSD128, and the noise at 5s issues

 

Thanks so much for this. Everything seems to be working well, and I'm happily experimenting with the different filter types.

 

--David

[bobbmd]

i have tried the trial of a+ and will buy it

do i understand correctly it will take TIDAL and Qobuz streaming flac from pcm to dsd/same goes for my itunes library of 29000+ songs in AIFF ? or is this just for recordings i have purchased and downloaded? will my pre-multibit Gungnir DAC still handle this 'upsampled' music? or am i misunderstanding the application's processing entirely

thanks bobbmd

[Jud]

i have tried the trial of a+ and will buy it

do i understand correctly it will take TIDAL and Qobuz streaming flac from pcm to dsd/same goes for my itunes library of 29000+ songs in AIFF ? or is this just for recordings i have purchased and downloaded? will my pre-multibit Gungnir DAC still handle this 'upsampled' music? or am i misunderstanding the application's processing entirely

thanks bobbmd

 

I don't think the Gungnir, delta-sigma or multibit, accepts DSD input.

[YashN]

With 4GB RAM, be sure to have the audio buffers pre-allocated memory be no more than 1.2GB

 

Salut Damien,

 

Yes that must be it, a memory issue with such low RAM to begin with. I re-tested after closing Firefox and then could change tracks without any of the previous problem.

 

Hence, I managed to do some additional tests, including putting the preferences on the Filter tab to up-convert to DSD256. While HQ Player so far with similar settings give me a choppy sounds - most probably a combination of low RAM and older CPU so the issue has a lot to do with my side as the recommended system needs higher specs - Audirvana+ did work well.

 

I listened to Peter Gabriel - Mercy Street, some Tears For Fears, Laurent Voulzy's 'Never More' on the excellent 'Cache Derriere' and more - very good sound, clean, dynamic and spacious!

 

On some tracks though, I got the glitch at 5 seconds (it's a similar glitching sound as when the iFi changes sample rate). The glitch also appears to occur at multiples of 5s intervals (like 10s, 25s or something like that) if that helps to diagnose it. Sometimes, re-starting the same track doesn't show the glitch, and I am not sure but I thought I heard the glitch more when I am scrolling through the album cover browser.

 

I haven't yet spent much time comparing the various filter types yet, so the default is being used for now.

 

This is very encouraging.

[qdwieertteyrujpo]

Thanks so much for this. Everything seems to be working well, and I'm happily experimenting with the different filter types.

 

--David

Damien,

 

thank you very much! Playing through the latest Beta: the fan noise interestingly enough still appears about 30 secs. into each track, but subdues after another minute. So you are definitely on the right track; of course, not to mention the great sound.

 

Kind regards,

Dieter

[jhwalker]

2.5.0.9 working *very* well on my system. No more gaps at 5s, with Direct mode disabled, no stalls loading first track, etc.

 

mansr, any chance you might give us some details about the filters we're listening to. FWIW, now that all are working, I find I prefer "C", but all the "B" variants are nice, as well - particularly B3 and B4

[Jud]

[ATTACH=CONFIG]25756[/ATTACH]no roll off before 20k

 

Not with your filter parameters, no. To minimize ringing, my steepness is set at 5, so the cut would begin at lower frequency.

[OldBigEars]

Sadly I'm at the point of bailing out of the Great DSD Adventure. I can find no way to reduce CPU usage below peaks of 250-260%, which seems extraordinarily high to me. The SQ is certainly excellent, but the marginal superiority is not worth potential damage to my computer (if indeed running at sustained high CPU's would have that affect). Hopefully at some point someone will figure out how to achieve the end-result of superb DSD quality without this downside.

[Le Concombre Masqué]

do you mean the graph showing (almost) no roll off before 20K means nothing ?

 

with a steepness of 5 looks like a cutoff of 1.14 is required to get no roll off before 20 and generates a lot of aliasing : guess I don't read right or there is more to the picture... could you please elaborate ?

Not with your filter parameters, no. To minimize ringing, my steepness is set at 5, so the cut would begin at lower frequency.

[tipunch]

For correct sound staging you want the Cutoff te be 1,00 or very close to that north for a little added air and depth but less focus (1,01 - 1,04 max). If 0,95 works good for you I expect something to be wrong in speaker placement.

 

I tried some other settings…

 

 

 

But I find the settings "Le Concombre Masqué" more pleasant, more organic in my ears and on my system with also loudspeakers of the 90s (See signature).

Maybe that the loudspeakers of the 90s were a little better balanced ;-)

 

Best regards

[damien78]

Sadly I'm at the point of bailing out of the Great DSD Adventure. I can find no way to reduce CPU usage below peaks of 250-260%, which seems extraordinarily high to me. The SQ is certainly excellent, but the marginal superiority is not worth potential damage to my computer (if indeed running at sustained high CPU's would have that affect). Hopefully at some point someone will figure out how to achieve the end-result of superb DSD quality without this downside.

 

Just to clarify on the 250-260% figure you see in activity monitor. The 100% value means a full utilization of a single CPU core. Our modern CPUs have multiple cores, and thus a value above 100% means that the computation is spanned accross multiple cores to make better use of the CPU hardware.

In this case, this means the equivalent of 2.5 cores of the CPU are used.

As you can see in the Activity Monitor screenshot below, the total use of my CPU (Core i7 2.3GHz) is only of 37% while upsampling to DSD256:

[damien78]

Salut Damien,

 

Yes that must be it, a memory issue with such low RAM to begin with. I re-tested after closing Firefox and then could change tracks without any of the previous problem.

 

Hence, I managed to do some additional tests, including putting the preferences on the Filter tab to up-convert to DSD256. While HQ Player so far with similar settings give me a choppy sounds - most probably a combination of low RAM and older CPU so the issue has a lot to do with my side as the recommended system needs higher specs - Audirvana+ did work well.

 

I listened to Peter Gabriel - Mercy Street, some Tears For Fears, Laurent Voulzy's 'Never More' on the excellent 'Cache Derriere' and more - very good sound, clean, dynamic and spacious!

 

On some tracks though, I got the glitch at 5 seconds (it's a similar glitching sound as when the iFi changes sample rate). The glitch also appears to occur at multiples of 5s intervals (like 10s, 25s or something like that) if that helps to diagnose it. Sometimes, re-starting the same track doesn't show the glitch, and I am not sure but I thought I heard the glitch more when I am scrolling through the album cover browser.

 

I haven't yet spent much time comparing the various filter types yet, so the default is being used for now.

 

This is very encouraging.

 

Is this happening with the 2.5.0.9? Is it happening when you are not scrolling nor doing anything with the UI?

If yes, can you send me (by email) your iZotope settings, SDM filter used, DSD rate, and the file that shows the issue? And any other details you may think of to help me reproduce it.

[Kugel]

Has anyone tried to Dac Ayre QB9DSD ? As configured ?

Thank you

[emuzaman]

Any chance to get "replay gain" in upcoming betas or 2.5.1?

[Hazard]

Is there a Linux version of this software?