As far as I know, several settings within Audirvana Studio will alter the ‘bits’ of an audio stream so the playback is no longer bit perfect. These settings include:
- Enable volume levelling (replay gain)
- Enable upsampling (PCM > PCM and PCM > DSD)
- Enable software volume control
- Enable VST or Audio Units plug-in effects
- Convert (downsample) DSD to PCM
Also, when AS loads any PCM audio file/stream, it decodes it into a 64 bit space to process, then outputs it to the bit depth accepted by the DAC (usually 24 bit) . Does this alter the bits or keep them bit perfect? (The 64 bit depth appears in a pop-up window when you click on the album you are playing in the bottom left corner)
Does anyone know the answer, or if there are also other processes within AS that can cause it to be ‘not bit perfect’?
It plays bit perfect unless some processing is applied in the settings.
With all processing, upsampling etc off Audirvana 2.x, 3.x have a very pleasant tone. AS also has a nice tone.
I did a lot of research that shows that for Jazz, Classical and Orchestral recording v3.5 has a more precise sound, that is more open and clear. AS has a little but darker, pushed back tone.
The sound difference between these versions is much less apparent with Rock or Pop music.
Where I am going with this, is that Audirvana does not stream the exact bit stream from your FLAC or DSF files to the DAC, it does some sort of cleanup/correction, even with everything off.
It is up to you if you want to call this bit perfect.
If you have a Qobuz subscription, find an album called WHAT’S IT ALL ABOUT, by PAUL METHENY.
1st track: The Sound of silence, listen the first 35 seconds.
You will hear the bigger imaging of v3.5, you will hear the clear overtones of the guitar strings, that’s exactly how they sound and feel if you play guitar. With Studio the sound is somehow modified, you still hear the overtones, but they feel like overtaken by a slight delay effect.
10th track: And I Love Her, sounds like he’s playing a nylon string classical, the into bass line is played with thumb flesh (no nail by the sound of it), With 3.5 the thumping bass line sounds very realistic and natural. With Studio the sound is muffled, like someone put a towel on the guitar. The strings are mushed together.
I’m using Sennheiser HD800S headphones and a Marantz HD-DAC1. Your results may vary depending on hardware and ears.
Are you using the PC version to get that muffled sound?
According to what criterion ? Why Audirvana would do this ?
I also found AS “muffled” (talking of very little difference but real) with Big Sur, and I supposed that this was due to the absence of Direct Mode.
Installing both 3.5 and AS on a Mac Mini/MOJAVE and having Direct Mode on both, shows no difference for me. Of course all processing is off, an only Direct Mode and Integer Mode are On. (Integer set on Mode 1. I do not like Mode 2)
And then I think this should be very close to so called “Bit Perfect” if no jitter at all,
Furthermore, this does not necessarily involve any perrfection , considering that perfection is to find in sound pickup. A poor sound recording would give a poor “Bit Perfect” reproduction of course, beeing worse with hight rate sampling, and that why some SD file sound much better than some HD one.
I will test your examples this evening and let you know.
I’m also on Big Sur.
My trial period is over, but I may subscribe to AS for a single month to test what you’ve noticed. A friend is ready to lend me his A3.5 activation key, so I could do that.
I am stupid !! My period is also over
I can’t test anything and V1.4.5 just arrived now !
I may suscribe also lol !
Where do you see V1.4.5? I see nothing alike on the site…
I got the option to upgrade when I checked for upgrade. I.e. under the Account option in settings.
There a new update today for Mac and PC.
What about the fact that all PCM tracks are converted to 64 bit within Audirvana Plus, regardless of settings? I don’t know, I am just wondering.
No the tracks are not converted to 64 bit. The (16, 24 and 32) bits are put in a 64 bit container and padded with zero’s.
So basically the bits stay unchanged but are simply put in a bigger ‘box’. If you get them out of the ‘box’ they are exactly the same as they where before. Lots of music players do this, because it is convenient to have one box that fits them all. It still is bit perfect though because no conversion is done.
and computers are 64bits…
You have a point.
I set both AS 1.4.5 and A3.5 the same.
No upsampling, no sound leveling, nothing… Only Integrer mode is set on both.
I compared some rock and jazz tracks between AS and A3.5. The impression is that AS is a bit muffled, but it’s not so clear with electrical music as it is with acoustic tracks of classical music.
And I can hear better the difference between the two players with Hi Res, rather than with 16/44.1 tracks.
I played some 24-96 tracks from the album “Voice of Hope” by the cellist Camille Thomas (Ed. Deutsche Grammophone 2020).
It’s a very good sounding album that I purchased from Qubuz, so Qobuz subscribers can listen to it and check what I’m saying.
Track 11, William’s “Schindler’s List”, is a cello solo with the orchestra playing in the background.
There’s a clear difference between the two players.
Compared to AS, the cello in A3.5 has more presence, the sound in general has more depth, and it sounds much better.
AS shrills a bit, from time to time, and its sound is a bit aggressive.
I listened with iFi micro iDSD Black Label DAC and Audeze LCD-2 Headphones
I would say that the difference between the two players is more obvious in the second half of the track.
I compared the same track with A2.5 as well.
A2.5 seems quite close to A3.5, and sounds better than AS.
I listened also with others headphones, the Beyerdynamic DT880. They are of a lesser quality than the Audeze, but still good sounding ones. And I heard again what I described in the post above.
Hey Guys I agree with both of you !!
My configuration beeing very different of yours, it confirm that it is not equipment relevant.
Looks like loosing some audio transients, which means nothing in the digital domain but it is the way my analog ears feel it. It is still a small difference but It’s there. The cello of Camille Thomas do not show plainly its beautyfull vibration with AS, and all VoyagerDude observations are completly true.
I did an other test on Yo-Yo Ma Plays Ennio Morricone. Very beginning of track 1 there are several transient noises due to bow inversion on the string and very audible because the orchestra is quiet enough then. The different is clearly audible.
I am going to ask support why. There is no reason to change the basic part of the audio core unless to improve it.
My configuration for info:
- Mac mini late 2012 with MOJAVE
- Preamp / DAC CP-800 Classè Audio
- 4 CA-M300 Classè Audio (for Bi-Amp)
- 2 B&W 800D2
Your test proves also that it’s not OS dependent, since you’re using Mojave, nor mode dependent, since you use Direct mode that is not available to us.
I find it very surprising, because if there is a difference, it should have been an improvement rather than a regression.
I listen now carefully to some DSD tracks, and I find that the difference between the players is even more perceptible with DSD. I’ll post a report about it, later tonight or tomorrow morning.
Exactly what I am saying. I am going to deal that with support directly.
Have you seen that things are going better on UI concerns with 1.4.5. This is encouraging.