samedi 15 mars 2008

ESt-ce que la maximisation apporte plus de résolution?

Une petite référence dans notre discussion intermittente sur les effets de la compression sonore dans la qualité des disques laser...

Un des arguments de Nick Davies dans sa compression du son dans la série des SACD de Genesis est d'utiliser pleinement les 16-bits du format; sous prétexte qu'un volume trop bas nous éloigne de cette pleine résolution.

Barry Diament, qu'on aime parce qu'il fait sonner Genesis comme pas un, réfute...

Voici son explication:

While 6 dB does add up to 1 bit of resolution, I believe this bit ( ) of information is often taken out of context and misrepresented.

The logical extension of the common argument would say that unless we compress the dynamic range to 6 dB or less, we're not using all of the available resolution on a CD. (Oh, wait a minute. The majors are compressing the dynamic range to 6 dB or less, aren't they? Actually, with many current releases, 6 dB would be an expansion. )

As you can see, it becomes a silly argument rather quickly, especially if you want to have a dynamic recording.

As to digital masters after noise shaping being "effectively 14-bit", here again, I must disagree. This too is what I call a "reductionist" perspective on a picture that is much larger.

If an original recording is made in high resolution (say 24-bit) and one wants to make 16-bit CDs, the options are:
1. to simply truncate (i.e. "throw away" bits 17-24), which will wreck the tonality and soundstaging
2. to dither (adding about 4 dB of noise overall), which if done properly will preserve much of the low level information that would otherwise be lost
3. to dither and noise shape, which if done properly will keep the benefits of good dithering while making the added noise considerably less noticeable.

To suggest a reduction in resolution with noise shaping is to imply one will not suffer any loss by avoiding noise shaping. In my experience, it is the opposite that is true.

I hear resolution as being primarily a function of the quality of the original recording (i.e. what leaves the microphones), the quality of the A-D conversion and how carefully everyone in the chain works to preserve what was recorded.

Many of the best sounding recordings I own do not spend much time near the top of the meters and all of them involved some kind of noise shaping for reduction of the original to 16-bit for the CD.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

Aucun commentaire:

Publier un commentaire