Tu peux également m’écrire en FR; pas de souci. Mais mon commentaire précédent reste pertinent: s’il y a du “bruit”, le logiciel n’est pas bit-perfect.
C’est ça le truc. En théorie, dans le mode Bit perfect, tous les lecteurs auraient dû produire le même son exactement. Puisque tous sont censés transférer les bits du morceau qu’ils jouent au DAC et ne rien faire d’autre. Mais dans la pratique, on constate une différence du son produit par des lecteurs différents. La cause de cela est le bruit que produisent les autres processus qui tournent sur l’ordinateur.
Et cette différence existe également entre des lecteurs différents d’Audirvana (Exemple : il y a une différence en mode Bit perfect entre A3.5 et AS). La raison à cela est que ces deux lecteurs ne procèdent pas exactement de la même manière dans le processus de réduction du bruit de l’ordinateur.
Encore une fois: dans ce cas, les logiciels concernés ne sont pas “bit-perfect” - il doit avoir du DSP dedans - par ailleurs, TOUTE autre différence sonore est située dans le domaine analogique.
Je vais essayer d’être plus clair en le formulant différement.
Les lecteurs ne sont Bit perfect qu’en théorie. Dans la pratique, la “pureté” des bits du morceau qu’ils transfèrent au DAC est “pollué” par le bruit ambiant que produisent les autres processus qui tournent sur l’ordinateur. Donc dans la pratique, le Bit perfect n’existe pas.
Et la raison à cela est la présence du bruit ambiant. Même si aucun DSP n’est appliqué par le lecteur au son, le DAC ne recevra pas exactement les Bits du morceau d’origine, puisque le bruit s’en mêle.
Étant donné que des lecteurs différents appliquent des procédés différents pour réduire le bruit de la machine, on entendra un son légèrement différent d’un lecteur à l’autre, même s’ils sont tous réglé en mode Bit perfect et n’appliquent aucun DSP au son.
Cela est même valable pour des versions différentes des lecteurs Audirvana, puisqu’ils n’appliquent pas forcément les mêmes procédés pour réduire le bruit de l’ordinateur. Le résultat est qu’ils sonnent légèrement différemment.
I am sorry @Doudou.
This does not sound physically possible.
Bits are either on or off. 1 or 0.
Are you saying some of the bits get flipped while being sent via USB?
I understand what you mean; yet I can’t agree with it, as the premise of your comment is simply flawed - in digital transmissions, bits cannot be “polluted” - otherwise a DIFFERENT transmission would take place. In other words, you’re implying that Audirvana is lying when it states that it’s bit-perfect - again, what CAN happen is further degradation of sound in the analog space - not when “bits” are being transmitted unless packets are lost.
That’s the point on which I would have preferred to find the explanation that was made by the guy that I mentioned above, and who elaborated on what Damien said, because this guy did it very well and explained also what the noise provokes in the DAC, and what effect it has on Bit perfect. But unfortunately, I don’t remember exactly all what he said.
It will be worthy to find that discussion that was held in the first or second day of the trial period.
But in summery, instead of receiving only the bits of the song, the DAC receives also noise, which provokes problems in the Bit perfect reproduction of the sound…
I told you that “purer” or “polluted” may not be the best chosen words for it, but I’m not a computing engineer, so I could not find better.
What’s worth to understand is that Bit perfect is impossible because when the bits are sent to the DAC, there are other operations that occur in the computer. Audirvana is not lying, the player sends only the bits of the track to the DAC. But since other processes run at the same time on the computer there will be no Bit perfect transmission to the DAC.
EDIT
This is of course is valid for all the players, not only for Audirvana.
As you say, a 1 is a one, and a 0 is a zero, but as such, they are of absolutely no use to anyone but mathematicians.
I am sure you do not imagine your hard disk or RAM littered with tiny dancing 1s and 0s.
In practice, 1 and 0 are just the status distinguishing electric currents or magnetic states.
We have to consider the transmission of the 1s and the 0s, because that is what music players do, like so many other applications.
We can safely say that nothing man made being perfect, during their transmission or treatment, some 1s and some 0s can go astray, hence the use of redundancy checks, CRC being one of them.
Since those checks are themselves man made, there are occasionally, but at smaller occurence, mistakes that are repaired by other man made routines, with their own occasional failures.
Hence that odd funny pixel while watching a DVD or little click when listening to a CD.
Rare and of little actual consequence, those aberrations make that bit perfect can only mean “extremely low probability of aberration occurence in the treatment and transmission”
Let’s not despair, near perfect usually does it for very many of us, we are humans after all…
Guys, the bit error rate over USB is so low that you’d have to listen to music for a million years to get one flipped bit. Forget the nonsense with “purity” of bits.
3 Likes
Yes I get that but if the transmission via USB is not reliable how can any app possibly have any effect what does the USB controller send to a DAC?
Padraic just stated it quite neatly above, (while risking being flamed by the USB cables afficionados.)
Yes indeed. 
I’m still really curious if Audirvana employs any DSP when in supposedly bit perfect mode. And if not why different versions sound different and different from other apps.
The secret sauce explanation doesn’t cut it.
I understand that Damien is keeping his cards very close to his chest on that subject.
You can check it very easily.
Set any player you choose to Bit perfect mode.
Then launch 20 apps on your computer and listen to the track. Then play the same track, after you quitted all the apps except the player.
You won’t have the same playback in the two cases. So which of them is Bit perfect according to you? They can not be both Bit perfect since the playback defer.
My humble answer is: none! as I probably do not have a hearing of a sufficient quality to hear such differences.
If I follow your train of thought, there is no end of things to consider and get bothered about.
What computer? What power supply? What RAM,? What Type of disk? What type of connection?
What HiFi system?
What God?…
Believe me that with a decent Hi-Fi system you will hear the difference very well.
Try it by yourself.
Do you think people would have bought software such as Fidelizer, if there was no audible difference in the sound quality due to the reduction of the noise of the computer? The difference in the sound quality is important, that’s why they spend money on it.
I am not disputing your point of view, I am just stating that having two HiFi system of decent repute, I am not certain I perceive all the differences I hear about in Audiophile circles.
To this HiFi agnostic, pleasure is the measure.
If you don’t hear the difference in playback between a silent computer and a noisy one, there’s no reason that you’ll hear the difference between an audiophile player and iTunes. So why bather with Audirvana, if you can play your music with iTunes for free?
Running other apps doesn’t introduce “noise”.
For what it’s worth, I ran 3.5 with and without “SysOptimizer” and detected no difference in sound quality.
1 Like