Audirvana Plus 3 (official thread)

[Innocent Bystander]

7 hours ago, jimdukey said:

I just drag and drop.

No muss, no fuss.

 

 

Well, drag and drop works, but especially with a headless mac controlled remotely from a tablet or phone (eg using VNC) it is not slick as A+ is supposed to be with its app. Land does not lend itself well to use by a non-computer-savvy partner, so is not a sustainable solution.

[Charente]

It isolates, for sure, but I would say the mR is the renderer and Audirvana is the player (orchestrator) or streamer ... mR is a low-powered, highly optimised (Linux) computer just for rendering the music to the DAC and, in its design, not using consumer grade components, such as exist in your MacMini. A double-edged benefit IMO.

[Innocent Bystander]

44 minutes ago, Charente said:

It isolates, for sure, but I would say the mR is the renderer and Audirvana is the player (orchestrator) or streamer ... mR is a low-powered, highly optimised (Linux) computer just for rendering the music to the DAC and, in its design, not using consumer grade components, such as exist in your MacMini. A double-edged benefit IMO.

 

Now that is more my understanding of mR, but different from what I seemed to read in bmbayer's post. So if this is the case, what is the benefit of A+ As opposed to any other compatible media server? And more importantly, have you compared mR to A+'s rendered output (isolated as necessary), and if so what's your assessment of tge difference?

[Charente]

18 minutes ago, Innocent Bystander said:

 

... what is the benefit of A+ As opposed to any other compatible media server? 

 

I guess that will be down to your preference in terms of sound ... Audirvana produces a signal for the mR to render .. IMO it's very good. If you like the sound Audirvana produces over, say, Roon or HQPlayer etc. then that's what you get. The mR changes nothing ... there is no re-generation of the music going on ... it renders what it receives ... a (clean) dumb terminal if you like !

 

18 minutes ago, Innocent Bystander said:

 

And more importantly, have you compared mR to A+'s rendered output (isolated as necessary), and if so what's your assessment of tge difference?

 

Originally I had Audirvana playing direct (no isolation) to my Schiit Gungnir ... with mR in situ, in that setup, there was a significant uplift in terms of 'noise' elimination (i.e. clarity) ... it was a case of I didn't know I had noise until it was gone. However, it is fair to say that the Gungnir Gen2 USB input is probably best described as 'not the best' ! Maybe not an issue with your Chord Dave.

[bnbayer]

4 hours ago, Innocent Bystander said:

 

Now that is more my understanding of mR, but different from what I seemed to read in bmbayer's post. So if this is the case, what is the benefit of A+ As opposed to any other compatible media server? And more importantly, have you compared mR to A+'s rendered output (isolated as necessary), and if so what's your assessment of tge difference?

Obviously I used some incorrect terminology... music guy, not techie. Here's Damien's answer to more-or-less the same question on this forum 7 months ago: 

"Not serving the files as is, but preparing them as for local playback, serving AIFF or WAV (or raw DSD) to the renderer.

And thus enabling, in addition to remove this decoding workload from the renderer (and this makes a difference), to use all processing available in Audirvana Plus: MQA core decoding, Upsampling with iZotope, Audio Units, ..."

[Innocent Bystander]

 

10 hours ago, bnbayer said:

Obviously I used some incorrect terminology... music guy, not techie. Here's Damien's answer to more-or-less the same question on this forum 7 months ago: 

"Not serving the files as is, but preparing them as for local playback, serving AIFF or WAV (or raw DSD) to the renderer.

And thus enabling, in addition to remove this decoding workload from the renderer (and this makes a difference), to use all processing available in Audirvana Plus: MQA core decoding, Upsampling with iZotope, Audio Units, ..."

 

Ok, that is interesting, and clarifies a little - I obviously missed Damien saying at the time, so thanks for repeating it.

 

However, to me the critical thing is the renderer (which is one thing A+ does very well, and the one reason I am still using it), the sound of the music being key, and that is where I would be interested in alternatives, purely because A+'s library side is a frustration that lets it down.

 

As for the other potential functions of A+ Cited: I understand transcoding can be done very effectively on the fly by many UPnP servers; I am uninterested in lossy formats or online streaming so MQA is of no interest to me; my DAC does not benefit from prior upsampling (in fact Chord's designer advises against upsampling in a computer for any of their DACs, saying it is done better in the DAC); and I don't use audio units. 

[Jud]

2 hours ago, Innocent Bystander said:

(in fact Chord's designer advises against upsampling in a computer for any of their DACs, saying it is done better in the DAC)

 

DAC makers often say their secret upsampling sauce is better than doing it elsewhere.  And of course software developers often say their software does a better job.  I would simply say listen for yourself and make up your own mind about it.

[Innocent Bystander]

11 minutes ago, Jud said:

 

DAC makers often say their secret upsampling sauce is better than doing it elsewhere.  And of course software developers often say their software does a better job.  I would simply say listen for yourself and make up your own mind about it.

 

I did - in fact it was actually while in the middle of trying A+ upsampling and not hearing a difference that I came across that note by Chord's designer! The reality presumably depends on the DAC's available processing power, however using A+ to upsample to 768 did produce one dropout. Although only momentary and only the once, that makes it enough to avoid, so it remains up to the DAC.

[copy_of_a]

On 30.10.2017 at 12:30 AM, fanofquality said:

I'm I the only one that thinks Audirvana 2 sound better than version 3?

(...)

version 3 has less bass and the bass is less detailed. Also the general sound is not as open and clear as version 2.

I am under the impression that most users here prefer V3 SQ-wise. Me too. If at all, I am finding the bass more detailed with V3.

But I think it may also depend on the mode you are using ... Integer Mode 1, Integer Mode 2 or MAC's native integer Mode (so A+'s Direct Mode disabled). I am listening through Integer Mode 2 and this is why I am referring to this Mode...

 

[jimdukey]

What or where is the Mac Integer Mode?

I thought I.M. was for PCM playback.

I upsample PCM to DSD 128, so I thought it didn't apply.

TIA!

[copy_of_a]

19 minutes ago, jimdukey said:

What or where is the Mac Integer Mode?

as noted in brackets: when you disable "Direct Mode" in the prefs A+'s own proprietary audio driver is bypassed and you are utilizing MACs native integer mode (provided your DAC supports integer mode ... and provided you are not on Sierra or High Sierra since A+'s Direct Mode doesn't work by default on these new operating systems. So on Sierra / High Sierra you get MACs native integer mode by default ...).

 

[jimdukey]

Thanks. 

 

[Innocent Bystander]

On 11/6/2017 at 5:12 PM, copy_of_a said:

I am under the impression that most users here prefer V3 SQ-wise. Me too. If at all, I am finding the bass more detailed with V3.

But I think it may also depend on the mode you are using ... Integer Mode 1, Integer Mode 2 or MAC's native integer Mode (so A+'s Direct Mode disabled). I am listening through Integer Mode 2 and this is why I am referring to this Mode...

 

 

I teresting, as according to the manual 

Mode 1    brings the highest transparency, and soundstage depth
Mode 2    is more on the warm side

"warm" often means more bass (or less treble), so maybe not surprising if those using M2 don't find the reduced bass an issue. However I took the manual to mean that M1 is the most accurate.

 

It would be informative to know if there is a correlation between  the apparent difference in sound quality between A+ v2 and v3 for people using Integer Modes 1 vs 2.

 

[Innocent Bystander]

On 11/6/2017 at 5:33 PM, copy_of_a said:

as noted in brackets: when you disable "Direct Mode" in the prefs A+'s own proprietary audio driver is bypassed and you are utilizing MACs native integer mode (provided your DAC supports integer mode ... and provided you are not on Sierra or High Sierra since A+'s Direct Mode doesn't work by default on these new operating systems. So on Sierra / High Sierra you get MACs native integer mode by default ...).

 

 

IIUC it is not Audirvana's  audio driver that is bypassed, but the Mac's own audio driver and format converter, thus accessing the lowest level in Mac's Core Audio to maximise sound quality.

[Jud]

6 hours ago, Innocent Bystander said:

 

IIUC it is not Audirvana's  audio driver that is bypassed, but the Mac's own audio driver and format converter, thus accessing the lowest level in Mac's Core Audio to maximise sound quality.

 

Direct Mode uses the A+ driver.  Thus not using Direct Mode “bypasses” the A+ driver (or at any rate doesn’t use the driver).

[Innocent Bystander]

3 hours ago, Jud said:

 

Direct Mode uses the A+ driver.  Thus not using Direct Mode “bypasses” the A+ driver (or at any rate doesn’t use the driver).

 

Now, I will admit to not being clear on what A+ is doing, only having the manual to go on:- 

 

From the manual's wording The only conclusion is that Integer Mode is PART OF Direct mode, not an alternative:

 

Direct Mode: In this mode Audirvana Plus bypasses completely CoreAudio, including its low-level layers. This enables a further shorter audio signal path for better sound quality, and gives possibility to enable Integer Mode in Lion and later.

 

So whilst not using Direct Mode indeed means not using the A+ driver, it also means not using Integer Mode

 

Then the description of Integer mode:

Integer mode: In this mode Audirvana Plus directly sends the audio data to the lowest level in CoreAudio, bypassing the driver audio mixer and the format converter. 

 

The only interpretation of this that makes sense is that it is the Mac's audio driver and converter that are being bypassed when Integer Mode is used [under Direct Mode]. 

 

I know the original description will have teanslated from Fremch, so I suppose it is possible that something will have come across incorrectly, but where does information come from to say this is wrong?

 

[Jud]

2 hours ago, Innocent Bystander said:

 

Now, I will admit to not being clear on what A+ is doing, only having the manual to go on:- 

 

From the manual's wording The only conclusion is that Integer Mode is PART OF Direct mode, not an alternative:

 

Direct Mode: In this mode Audirvana Plus bypasses completely CoreAudio, including its low-level layers. This enables a further shorter audio signal path for better sound quality, and gives possibility to enable Integer Mode in Lion and later.

 

So whilst not using Direct Mode indeed means not using the A+ driver, it also means not using Integer Mode

 

Then the description of Integer mode:

Integer mode: In this mode Audirvana Plus directly sends the audio data to the lowest level in CoreAudio, bypassing the driver audio mixer and the format converter. 

 

The only interpretation of this that makes sense is that it is the Mac's audio driver and converter that are being bypassed when Integer Mode is used [under Direct Mode]. 

 

I know the original description will have teanslated from Fremch, so I suppose it is possible that something will have come across incorrectly, but where does information come from to say this is wrong?

 

 

Not trying to say anyone is wrong.

 

Just to be brief and clear, hopefully for the benefit of understanding:

 

- Integer mode can run on top of either Damien’s driver (Direct Mode) or Apple’s driver (Core Audio).

 

- “Driver” just means the interface between the hardware itself and the operating system.

[Innocent Bystander]

15 hours ago, Jud said:

 

Not trying to say anyone is wrong.

 

Just to be brief and clear, hopefully for the benefit of understanding:

 

- Integer mode can run on top of either Damien’s driver (Direct Mode) or Apple’s driver (Core Audio).

 

- “Driver” just means the interface between the hardware itself and the operating system.

 

If that is the case, it means there are two different implementations of integer mode, and people need to bear in mind that any commentary on sound quality needs to be clear which is being used or compared - i.e.

A) Direct Mode Integer Mode (1 or 2) = shortest path if the manual is correct, and minimal reliance on Apple.

B) Core Audio (or Indirect) Integer Mode (1 or 2) = the only option post Sierra unless the Direct Mode fix is applied.

 

This is aside from the manual description under Direct Mode being misleading - something hopefully for @Damien78 to amend some time.

[Jud]

17 minutes ago, Innocent Bystander said:

 

If that is the case, it means there are two different implementations of integer mode, and people need to bear in mind that any commentary on sound quality needs to be clear which is being used or compared - i.e.

A) Direct Mode Integer Mode (1 or 2) = shortest path if the manual is correct, and minimal reliance on Apple.

B) Core Audio (or Indirect) Integer Mode (1 or 2) = the only option post Sierra unless the Direct Mode fix is applied.

 

This is aside from the manual description under Direct Mode being misleading - something hopefully for @Damien78 to amend some time.

 

Don’t make it more complicated than it needs to be.  It’s the same integer mode, but just running with two different interfaces between the DAC and the operating system.

 

Direct Mode was made as a leaner interface than Core Audio, so there may potentially be a sonic advantage to it.  (Subjectively I prefer it.)

[Adacadac]

Currently on the trial license. If I decline to purchase V3, will I revert to 2.6 automatically? I'm running an already ultra-revealing rig (Metrum>Krell>Theil) and feel this version has taken away the 'magic' of 2.6. This is too clinical for my taste. It just doesn't 'flow'. Thanks

 

 

[Innocent Bystander]

4 hours ago, Jud said:

 

Don’t make it more complicated than it needs to be.  It’s the same integer mode, but just running with two different interfaces between the DAC and the operating system.

 

Direct Mode was made as a leaner interface than Core Audio, so there may potentially be a sonic advantage to it.  (Subjectively I prefer it.)

 

Not sure what you're saying here -- agreed Direct Mode is leaner than Core Audio (neither with Integer Mode, but are you saying Direct Mode -Integer Mode is the same as NonDirect (i.e. core Audio) Integer Mode? If so, Direct Mode has no point, despite Damien saying in the Manual that the best Sound Quality is Integer Mode enabled by Direct Mode.

 

maybe it would be helpful if @damien78could step in here and clarify...

[bnbayer]

40 minutes ago, Innocent Bystander said:

maybe it would be helpful if @damien78could step in here and clarify...

 

It's not really that unclear, and Damien really needs to use his time programming.

 

Don't worry about descriptions in the manual: Listen to the various combinations - your own music with your own ears on your own system. Which one do you like better? Use that one.

 

[Jud]

1 hour ago, Innocent Bystander said:

 

Not sure what you're saying here -- agreed Direct Mode is leaner than Core Audio (neither with Integer Mode, but are you saying Direct Mode -Integer Mode is the same as NonDirect (i.e. core Audio) Integer Mode? If so, Direct Mode has no point, despite Damien saying in the Manual that the best Sound Quality is Integer Mode enabled by Direct Mode.

 

maybe it would be helpful if @damien78could step in here and clarify...

 

No.  I'm saying Integer Mode is the same thing (it's a way of sending data to the DAC in the DAC's native format) no matter whether it runs on top of Direct Mode or Core Audio, but that the fact you are running Direct Mode instead of Core Audio may make a difference in the sound.  **Direct Mode plus Integer Mode may sound different than Core Audio plus Integer Mode.**  Not sure what's unclear about this.

 

Anyway, as bnbayer suggested, perhaps you might try simply listening for yourself and see which you prefer rather than worrying about manuals and definitions.

[Innocent Bystander]

1 hour ago, bnbayer said:

 

It's not really that unclear, and Damien really needs to use his time programming.

 

Don't worry about descriptions in the manual: Listen to the various combinations - your own music with your own ears on your own system. Which one do you like better? Use that one.

 

 

59 minutes ago, Jud said:

 

No.  I'm saying Integer Mode is the same thing (it's a way of sending data to the DAC in the DAC's native format) no matter whether it runs on top of Direct Mode or Core Audio, but that the fact you are running Direct Mode instead of Core Audio may make a difference in the sound.  **Direct Mode plus Integer Mode may sound different than Core Audio plus Integer Mode.**  Not sure what's unclear about this.

 

Anyway, as bnbayer suggested, perhaps you might try simply listening for yourself and see which you prefer rather than worrying about manuals and definitions.

 

I'm quite content with the sound quality how I have it (v2.6.8 under Sierra, Direct Mode, Integer Mode 1). Aside from trying to understand exactly what the options are, my point was that anyone making observations about relative sound quality should clarify what setup they are comparing on each.

 

As for Damien's time, indeed I wish he'd attend to the things some of us have been requesting for years, but he is clearly uninterested, however if the manual is incorrect or confusing then it would be as well for him to be aware of that and clarify at next revision..

[copy_of_a]

15 hours ago, Innocent Bystander said:

I'm quite content with the sound quality how I have it (v2.6.8 under Sierra, Direct Mode, Integer Mode 1).

Unless you‘ve „hacked“ your OS you are listening through Core Audio (in integer mode), not through A+‘s „Direct Mode“. Due to limitations (security ...) of the OS „Direct Mode“ does not work by default on Sierra (and High Sierra).