Audirvana Plus 2.5.1 with PCM->DSD Conversion

[Jud]

So this is interesting: If I change the three parameters as suggested, no more noise issues. A+ just crashes immediately when I hit Play. If I change back to my previous settings, it's back to the same old noise problems.

 

--David

 

Glad I could help??

[orgel]

What is the DSD filter setting on your Mytek? 50, 60 or 70kHz LPF ?

 

It's set at 60. I'm going to leave it that way for the time being, but let me know if you want me to experiment with that.

 

What do you mean by "A+ would crash about every two filter changes"? Is it an application crash (in this case, can you send me by email or PM the crash report)?

 

I'd select a DSD filter in the prefs, close the Prefs dialog, hit Play, listen long enough to hear some problems, hit Stop, open prefs to change the filter to a different setting (say, A0 -> A4). Yesterday, the second or third time I'd do this, A+ would crash when I hit Play after selecting a new filter setting.

 

Unfortunately, I couldn't reproduce this behavior today, but I could get A+ to crash reliably when I changed my iZotope settings per Jud's suggestions. I'll email you a crash log.

 

--David

[TheDuke]

If you're using DoP, you need the 5.1A firmware for DSD256.

 

Thanks for the reply mansr. IFI were recommending staying with the Gelato version of 5.1. Downloaded 5.1A and upsampling to DSD256 is working fine.

[huaklom]

Damien nice work. I was shocked to see this thread as I wasn't expecting to see DSD conversion functionality in A+. I assume the latest remote app isn't built to work with these beta's?

 

A+remote app can be used with this beta smoothly. Very impressive.

[matthias]

Damien,

this is great news ! Congratulations !

Now we have a second front which we can open to persuade the PCM to DSD non believers

Best Regards

Edward

 

+1

 

KR

 

Matt

[craighartley]

This A+ version of PCM>DSD has really jumped up my sound to the best I've experienced yet. It's more revealing, has more depth, and yet is relaxed and non-fatiguing. Very, very nice. Have noticed a very few dropouts and also one louder "click" when changing from track on one album to a track on another album, but mostly just nice, involving music. This is upsampling to DSD 256 thru my Exosound e-22 and my 2008 Mac Pro with only one dual core processor (vs. the usual quad core). This may put HQ Player in my rear view mirror because now I have great sound with a much more friendly GUI (I don't have to search and search for where HQP put my classical tracks).

 

DSD256 is greyed out in Audirvana settings with my e22. What other settings are you using to get this?

[OldBigEars]

Am I correct in thinking that Schiit DAC owners won't be able to enjoy their music up-sampled to DSD? If so, that's unfortunate.

[Jud]

Am I correct in thinking that Schiit DAC owners won't be able to enjoy their music up-sampled to DSD? If so, that's unfortunate.

 

The DAC has to accept DSD input, which none of Schiit's current DACs do.

[Miska]

The interesting/puzzling thing for me is that I got stuttering with HQPlayer at a CPU load about equal to a load where I did not get stuttering with A+. Is the load an average and stuttering occurs at momentary higher loads? Or is there another explanation?

 

If any of your cores get exhausted, you get stuttering as a result. So looking at total load is misleading, instead what matters is per-core load.

 

It also takes just one load spike that takes a bit too long and misses a deadline and you get a drop-out. So average load is just that - average. What determines possible drop-outs are worst case scenarios.

[Jud]

If any of your cores get exhausted, you get stuttering as a result. So looking at total load is misleading, instead what matters is per-core load.

 

It also takes just one load spike that takes a bit too long and misses a deadline and you get a drop-out. So average load is just that - average. What determines possible drop-outs are worst case scenarios.

 

Thanks, Miska. Two cores, so I'm guessing there's not a lot of difference between 180-190% total utilization with HQPlayer vs. A+. If that's correct, it would be a matter of HQPlayer's modulator producing more spikes. As you say HQPlayer's modulators are more complex, this makes some sense.

[Miska]

Thanks, Miska. Two cores, so I'm guessing there's not a lot of difference between 180-190% total utilization with HQPlayer vs. A+.

 

For dual-core, 180-190% is 10 - 5 % away from full load. So getting stable audio from such on a general purpose OS is hard. It takes just one OS service deciding to do some extensive graphics, or maintenance work and things get broken...

 

Generally, I'd say that it is good to keep about 25% reserve for the OS. Especially if you run something graphically heavy like Roon as a front-end. Best if you have one or two extra CPU cores for the OS, etc.

[EdmontonCanuck]

I think I might just jump in and try this. What kind of performance can I expect running this on a late-2012 quad-core i7 Mac Mini? Should I easily be able to upsample to DSD128? DSD256? (running an exaSound e20 DAC here).

[Jud]

I think I might just jump in and try this. What kind of performance can I expect running this on a late-2012 quad-core i7 Mac Mini? Should I easily be able to upsample to DSD128? DSD256? (running an exaSound e20 DAC here).

 

Since I can easily do DSD128 DoP with a mid-2009 Core 2 Duo laptop, I imagine you should be able to do at least that.

[Dandaman]

DSD256 is greyed out in Audirvana settings with my e22. What other settings are you using to get this?

 

I've attached two screenshots for you with my settings. As an aside, note that A+ thinks the e22 can yield a 768 rate - which of course it cant'.

[mansr]

I've attached two screenshots for you with my settings. As an aside, note that A+ thinks the e22 can yield a 768 rate - which of course it cant'.

 

The fake 768 rate is for DSD256 DoP.

[Dandaman]

The fake 768 rate is for DSD256 DoP.

 

Oh. I'd asked George at Exasound and Damien also but had never heard a good answer. Thanks.

[qdwieertteyrujpo]

Some info on DAC, iZotope settings, SDM "filter," etc.?

 

Jus,

thanks for asking; here some more info:

DAC: IFi Nano; firmware 5.1

DoP standard 1.1

Steepness: 7

Filter: 1,300,000 (tried 10,000; no change)

Cutoff: 1.02 (tried 1; no change)

Anti-aliasing: 200

Pre-ringing: 0.86 (tried 1; no change)

DSD to 128; Filter B0

 

Regards,

Dieter

[mansr]

Jus,

thanks for asking; here some more info:

DAC: IFi Nano; firmware 5.1

DoP standard 1.1

Steepness: 7

Filter: 1,300,000 (tried 10,000; no change)

Cutoff: 1.02 (tried 1; no change)

Anti-aliasing: 200

Pre-ringing: 0.86 (tried 1; no change)

DSD to 128; Filter B0

 

Regards,

Dieter

Do you notice any difference between the A and B filters, especially the higher numbered ones? In my testing, the A filters seem to be preferable with iFi DACs.

[Gonzbull]

Since I can easily do DSD128 DoP with a mid-2009 Core 2 Duo laptop, I imagine you should be able to do at least that.

 

I've got the same comp as you Jud and was feeling a bit sceptical about up sampling to DSD. I'm going to wait for the proper release but eagerly hoping it copes. I've got an SSD in mine and the full 8G RAM. Music stored on external FireWire 800 drive.

[qdwieertteyrujpo]

Jus,

thanks for asking; here some more info:

DAC: IFi Nano; firmware 5.1

DoP standard 1.1

Steepness: 7

Filter: 1,300,000 (tried 10,000; no change)

Cutoff: 1.02 (tried 1; no change)

Anti-aliasing: 200

Pre-ringing: 0.86 (tried 1; no change)

DSD to 128; Filter B0

 

Regards,

Dieter

 

Jud,

apologies for misspelling your name.

[OldBigEars]

When using DSD, should you reduce the maximum volume setting on A+ ? I think I read somewhere you should reduce the volume ~ 6%. Anybody have any information about that?

[mansr]

When using DSD, should you reduce the maximum volume setting on A+ ? I think I read somewhere you should reduce the volume ~ 6%. Anybody have any information about that?

It may help if the modulator destabilises on some input (you'd notice if it did). If this happens, please let us know.

[craighartley]

I've attached two screenshots for you with my settings. As an aside, note that A+ thinks the e22 can yield a 768 rate - which of course it cant'.[ATTACH=CONFIG]25702[/ATTACH][ATTACH=CONFIG]25703[/ATTACH]

 

Thanks. Now working. I think the problem was that I had inadvertently checked 'Direct'

[qdwieertteyrujpo]

Do you notice any difference between the A and B filters, especially the higher numbered ones? In my testing, the A filters seem to be preferable with iFi DACs.

The B0 filter comes closest to what I have been listening to w/o converting to DSD, i.e. w/o conversion, I use the Voxengo Warmifier with A+. For me, especially the A filters sound too harsh. I don't know whether that answers your question, though.

[Jud]

I've got the same comp as you Jud and was feeling a bit sceptical about up sampling to DSD. I'm going to wait for the proper release but eagerly hoping it copes. I've got an SSD in mine and the full 8G RAM. Music stored on external FireWire 800 drive.

 

Same here (SSD, 8GB RAM, ext FW HDD - though I'm now storing the A+ library on a network drive connected to the laptop via Ethernet cable from the computer to a Wi-Fi router).