Audirvana 3.1

[watercourse]

On 12/25/2017 at 3:21 PM, watercourse said:

Actually, I believe that SoX is handling DSD to PCM conversion slightly better than iZotope, at least it has bit more ease in its presentation. No loss in resolution at the SoX default filter settings (my previous negative SQ comments were based on short listening with a bandwidth setting of 55). And 16/44 FLAC sounds great too.

 

Update: after listening to many locally-stored 16/44 FLAC files, I'm finding that I prefer the ability to fine-tune in iZotope - a bit more texture and depth than SoX. I feel the fixed steepness of SoX doesn't allow the ability to dial in sonics in the same way as iZotope. There is an initial apparent increase in resolution via SoX that I find is more fatiguing than iZotope over time.

 

Will focus on hi rez local files and Tidal streams next.

[buonassi]

On 12/19/2017 at 8:59 PM, RunHomeSlow said:

 

 

Just for information, since i don't use HQPlayer often just because of library look, but with HQPlayer upsampling at DSD128 my mac only peak at 100% or 95  to 105% stable... Why?

 

My question here is: How can a mac works at 300% CPU usage? is that number good in Activity Monitor ??

i don't feel my computer is working harder and slow me down elsewhere while music is playing. i don't hear the fan workin'...

 

 

HQ player may request thread processing differently than A+, where it requires the CPU to run longer to do interpolation, but at a lower intensity CPU.  This is just my guess.

 

A Mac can work at multiples of 100% based on your core count.  For example, if you have 4 cores and are at 300%, that's really only 75% of your entire CPU processing power.  Keep in mind that even virtual cores are counted in this CPU load summing algorithm.  So it's possible you have 2 virtual and 2 physical CPU cores. 

 

edit:  if you are so inclined, you can check out how to throttle CPU by process on Mac.  I did a quick google search and found at least one result that's current.  Though I don't have the guts to try this myself. 

[RunHomeSlow]

As in my signature i have a late 2012 mac mini, put in it 16gb of ram and installed and ssd hard drive just for the system, and i think it has only one core...

 

now i moved in a house and can listen to music at 85db, and with all the clicks, static noise and not finishing song till the end and pops after dsf album finished, i stop upsampling to dsd128.

 

now no upsampling, no direct mode in high sierra, blasting music without worries to hear a high screee noise anywhere without any reason ?

[AudioDoctor]

1 hour ago, RunHomeSlow said:

As in my signature i have a late 2012 mac mini, put in it 16gb of ram and installed and ssd hard drive just for the system, and i think it has only one core...

 

now i moved in a house and can listen to music at 85db, and with all the clicks, static noise and not finishing song till the end and pops after dsf album finished, i stop upsampling to dsd128.

 

now no upsampling, no direct mode in high sierra, blasting music without worries to hear a high screee noise anywhere without any reason ?

 

I have a 2011 Mini with the i5 chip, I have 2 physical cores and 2 virtual cores.  I would guess you do as well.

[RunHomeSlow]

33 minutes ago, AudioDoctor said:

 

I have a 2011 Mini with the i5 chip, I have 2 physical cores and 2 virtual cores.  I would guess you do as well.

 

you are right just checked... don't know about the virtual ones, but indeed 1 processor, 2 cores

so now they are working each at 150% for less time with SoX then at 125% each with Izotope

compared with HQPlayer steady at 50% each if i divide for 2 cores still at the same DSD128...

but library is way better with A+ and iphone remote...

 

[RunHomeSlow]

i was happy at the time A+ went DSD upsampling because i could never use A+ with my DAC upsampling by 2 PCM filesto 352/384 it never locks signal for sure but at 176/192 all was OK... hiccups occured every time or most when switching sample rate. So when DSD upsampling happened all was good... it was playing good without bugs, never checked the CPU then, but never felt computer was overworking, even fan was resting. But with later version, ticks happens in DSD upsampling, sometimes in songs like playing a vinyl... or at the end of an .DSF album having static noise... recently just was putting songs in list with iphone and when sample rate change to DSD big noise happen at the end of song... never fun to have not a steady software always playing like it should... i went to digital because i had an ARCAM CD73 citrus that was never reading perfectly good CD 

I will do i backup before using the next PC/MAC version  A+ for sure

[AudioDoctor]

Well, call me wrong.  It looks like Apple put non hyper threading I5 chips in our Mac minis, so we have only 2 physical cores.

[som4ew]

I'm on A+ 3.2.3 Mem. Buffer set at 3.072 mb /OS 10.12.3/ MAC Macmini6,2, 16 mb Ram, Intel Core i7. When up sampling in SoX to DSD 128 or even down to power of 2, I ALWAYS got hicup, static nois, stop playing in middle of songs, the bugging most of all is hash screech hiss noise so that I have to stop playing, now I'm forced to go for no up sampling to avoid all the above. Hope A+ will be sort out in the next version. I don't have these problems when play Iriver 23 up sampling in SoX at DSD 128.

[bazzvid]

i never up-sample to my Metrum Nos dac ,but even in normal mode SOX initially sounds impressive then it quickly wears off, realising it sounds hard,forced,the music dosent swing or groove anymore,and the decay or reverb is clipped like its running through a badly set up Dolby unit,maybe it just dosent suit Nos or it needs a bit of tweeking ,its not a problem happy with izotope.  

[buonassi]

On 12/31/2017 at 9:20 PM, RunHomeSlow said:

i was happy at the time A+ went DSD upsampling because i could never use A+ with my DAC upsampling by 2 PCM filesto 352/384 it never locks signal for sure but at 176/192 all was OK... hiccups occured every time or most when switching sample rate. So when DSD upsampling happened all was good... it was playing good without bugs, never checked the CPU then, but never felt computer was overworking, even fan was resting. But with later version, ticks happens in DSD upsampling, sometimes in songs like playing a vinyl... or at the end of an .DSF album having static noise... recently just was putting songs in list with iphone and when sample rate change to DSD big noise happen at the end of song... never fun to have not a steady software always playing like it should... i went to digital because i had an ARCAM CD73 citrus that was never reading perfectly good CD 

I will do i backup before using the next PC/MAC version  A+ for sure

 

My theory.... Damien wanted to make upconversion to DSD sound so much like analogue, he coded in all the sonic impurities that accompany vinyl.  Hyuk Hyuk.   I couldn't resist.

[tobes]

On 1/1/2018 at 2:20 PM, RunHomeSlow said:

i was happy at the time A+ went DSD upsampling because i could never use A+ with my DAC upsampling by 2 PCM filesto 352/384 it never locks signal for sure but at 176/192 all was OK... hiccups occured every time or most when switching sample rate. So when DSD upsampling happened all was good... it was playing good without bugs, never checked the CPU then, but never felt computer was overworking, even fan was resting. But with later version, ticks happens in DSD upsampling, sometimes in songs like playing a vinyl... or at the end of an .DSF album having static noise... recently just was putting songs in list with iphone and when sample rate change to DSD big noise happen at the end of song... never fun to have not a steady software always playing like it should... i went to digital because i had an ARCAM CD73 citrus that was never reading perfectly good CD 

I will do i backup before using the next PC/MAC version  A+ for sure

 

I think SoX sounds really good up sampling to DSD, but I too am getting ticks during playback. 

I'm up sampling to DSD64 - using 2011 Mac Mini (2.3Ghz i5 with 16GB ram, El-Capitan). Dac is the PSAudio Directstream Jnr.

After some experimentation I realised I only get the ticks when using files with a multiple of 48K sampling rate (i.e. 48/96/192). I get no ticks with 44.1/88.2/176.4.

No ticks using Izotope with the 48K based rates.

I wonder what's going on?

[mazuly]

Hello 

SoX sounds harsh to me as well.  iZotope is just smother and more musical.  I upsample to DSD128.

 

But

 

I updated to High Sierra yesterday and now iZotope crashed mid song or at the end of Play Queue.  I have send the crash reports to Damien and am waiting for the next update to hopefully fix the issue.  

Maybe I should just go back to El Capitan since I think the sound is also degraded with this upgrade.

 

[rruffin]

On December 24, 2017 at 2:40 AM, vicento said:

Hi, I use a Digione with DietPi operating system, no problem with the volume controler.

So you are not using a preamp (going directly from dac to amp) and using software to control the volume?

[AudioDoctor]

On 12/31/2017 at 10:35 PM, AudioDoctor said:

Well, call me wrong.  It looks like Apple put non hyper threading I5 chips in our Mac minis, so we have only 2 physical cores.

 

Now I don't know what to think...

 

Earlier tonight I updated my 10.13.3 to the Beta 3 version, because of that I had to go through the steps to re-enable Direct Mode in A+.  While doing that, specifically repairing permissions I saw this in Terminal.  It clearly says I have 4 processors, meaning I must have 2 virtual cores as well as my 2 physical cores, right?

 

 

[damien78]

I've just released 3.2.4 with fixes for crashes some of you encountered upsampling with SoX.

Thanks again for the reports that have been very helpful to reproduce, and fix the issue.

[copy_of_a]

3 hours ago, damien78 said:

I've just released 3.2.4 with fixes for crashes some of you encountered upsampling with SoX.

Thanks again for the reports that have been very helpful to reproduce, and fix the issue.

Thanks Damien!!

 

I'm still wondering whether or not the implementation of SoX is really correct...

 

Attached 2 screenshots comparing the frequencies of SoX and iZotope when upsampling 44.1kHz to 176.4kHz.

The screenshots show the interface of an EQ plugin in realtime mode while playing music.

SoX shows extremely weak filter quality producing a remarkable amount of mirrored aliasing (~45db).

I can't imagine this is how the SoX resampler is designed originally...?

 

 

SoX settings:

Bandwidth: 95% Nyquist
Anti-Aliasing: 100
Pre-Ringing: 68%

 

 

iZotope Settings:
Steepness: 28
Cutoff: 0.95 Nyquist
Anti-Aliasing: 100
Pre-Ringing: 0.68

 

 

[Dandaman]

@damien78  I'm not sure if Damien is following the Audirvana 3 Bug & Issue reports (though probably) so I will post here as well.  Still getting crashes now with 3.2.4 using Izotope, not Sox.  Songs play for about 4 minutes then the whole program crashes.  Crash report is attached.  Thanks for any help.  Version 3.1.8 does not crash on me.

crash version 3.2.4 using Izotope.rtf

[Chiler54]

Is there any news when the Integer Mode will officially be available again in Apple's High Sierra?

[AudioDoctor]

5 hours ago, Chiler54 said:

Is there any news when the Integer Mode will officially be available again in Apple's High Sierra?

 

Is Integer mode specific to the OS, or to the Device you are sending the bits to?  I have it on my High Sierra implementation, and the modes 1 and 2 do sound noticeably different, so I assume it is working.

 

Perhaps you have to go through the steps to enable direct mode to get integer mode as well.

 

 

edit, never mind.  I am apparently brain dead and completely missed the word "Officially" in your post.

[Chiler54]

Sorry my english is not so good....i can not change the integer mode 1 or 2. This is a High Sierra Apple problem. And replacing the missing kext file is too tricky for me....

[AudioDoctor]

1 hour ago, Chiler54 said:

Sorry my english is not so good....i can not change the integer mode 1 or 2. This is a High Sierra Apple problem. And replacing the missing kext file is too tricky for me....

 

Your English is fine, my reading skills were lacking.

[mazuly]

Hello Damien

Just updated to A+ 3.2.4.  I also did the hack for the direct mode for High Sierra so now I have Direct Mode.

But A+ still crashing when using iZotope to upsample to DSD128.  

Crash report attached.

Please let me know if you need any other information for this issue.

Crash 01 04 2018.rtf

[Chiler54]

How much better is the Direct Mode?

[quark]

 It seems after a while listening, although initially sounding pretty good and revealing much detail, SoX is a little harsh and quite tiring after long periods.

 I always personally find direct mode and exclusive mode much better than upscaling in my system, does anyone know if  you can still use direct mode and exclusive mode when using a plug-in?

 I haven’t bothered programming OS X to accept integer mode, does this really improve on direct and exclusive mode does anyone know? 

And if so, does anybody know any excellent plug-in that can soften sound a little without removing any of the detail? 

 I have absolutely no experience with audio plug-ins :-) 

 Thanks for any help 

[mazuly]

6 hours ago, Chiler54 said:

How much better is the Direct Mode?

 

I think it would be DAC related.

 

In my setup Direct Mode gives me better localization of the isnstunents in the soundstage image.  For example the singer becomes more localized and less hazy.  It’s like a vail which was making sound hazy is lifted and everything is more where it should be and does not move around.  I have tried going without it couple of times but finally decided that something is  amiss when I don’t have it so I finally did the hack.

 

Sorry it’s hard for me to explain since I am really not an audio magazine writer.

 

So yes it does sound better to me in my system.

 

Question

Does anyone know what Use Large Core Audio I/O Buffer does and should I have checked or not with Diret Mode?