To FLAC or not to FLAC

Avec la prolifération des serveurs musicaux ces jours-ci, et les myriades de configurations possibles, il devient de plus en plus intéressant d'obtenir l'opinion de gens ayant une acuité sonore et une expérience professionnelle du son avant de créer notre station. Surtout avec l'investissement considérable de temps requis pour le ripping, le tagging et la compression...

Pas de doute, le FLAC s'est imposé depuis quelques mois comme étant la solution de compression la plus prometteuse: application ouverte, gratuite, largement supportée, théoriquement sans aucune perte sonore. Mais est-ce que le FLAC est réellement transparent?

Citation de Barry Diament, un des ingénieurs au mastering les plus connus et respectés, et un collaborateur régulier et apprécié au forum SteveHoffman.tv.

In another thread, I mentioned recent listening tests to some FLAC files I made (using different applications and different settings). I listened with other ears I trust and in a blind comparison, it took only seconds to note the hardening and brightening of the sound that results when a so-called "lossless" compressed file, like a FLAC, is expanded in real-time.

The differences were relatively subtle but quite easily heard by all. With hard disks dropping in price every day, I see no reason to use any form of data compression. Besides, I want to hear things as good as I possibly can and for me, that means raw, uncompressed AIF (or WAV).

All that said, I did confirm that the FLAC can be expanded to AIF or WAV in an off-line process (i.e. NOT while listening) without any loss. But as I understand it, most folks who use FLAC listen directly to the FLAC (i.e. have the software expand it in real time, while they're listening). Under these circumstances, I don't like what happens to the sound.

This is consistent with my experience with other real-time processes, like sample rate conversion. I can take my favorite src algorithm and use it off-line to create transparent results. And I can use it in real time and hear the subtle hardening and brightening creep in.

So, for my own listening, it is always raw (uncompressed) AIF or WAV, played at its native sample rate.

(...) The tests used many different files, including recordings I made myself and the results were consistent, regardless of the recording, the software used or the settings.(...)

Again, not a night and day difference but an easily audible one, sufficient to deter me from using it for my own listening or from anything Soundkeeper does.

Just my perspective.

Best regards,

Barry

www.soundkeeperrecordings.com

www.barrydiamentaudio.com