After more than a decade leading the sound department at the famous Copra Studios (whose clients include TF1, Canal+, Pathé, Gaumont…), Philippe’s quite savvy when it comes to post-production–dailies synchronization, mixing, recording and sound editing.

Always on the lookout for the best solution scratching a particular itch, he has challenged many times Audionamix’ innovative source separation technology. Let’s hear what he has to say about it.

How does it differ from other remastering solutions?

It sticks out because it creates a new material one can use while remixing, whereas the others do nothing.

How does it actually work?

Let’s consider the simple example of a piece of music involving a saxophone and a piano. If the source is mono or stereo, the notes of both instruments are initially mixed together. Never mind: the Audionamix algorithm processing the track has memorized the sound prints of both instruments. Thanks to those, when the sax plays, the program will be able to recognize and pull its score out. The same goes for the piano. A little bit of computing later, both tracks are extracted, producing the expected result: a track per instrument. Thanks to this new material, I can very easily remix the piece and place the instruments in space as if they had been recorded separately.

How does the Audionamix technology interact with your usual professional audio environment?

In the post-production world, magnetic tapes gave way to digital files. Audionamix delivers WAV files which I can import directly into Protools. This way, I can be very reactive. Suppose I need a small demixing change. Thanks to Internet magic, I get the new files in less than one hour and can start working again.

What are the best applications produced by source separation?

The possibilities are endless: we have already mentioned remixing; there is also audio restoration–like removing parasitic noises from old tapes. In this specific case, I use Audionamix’ technology, grouping as a “virtual instrument” all cracks, blastings and clicks. Once extracted in their own track, those noises can then seamlessly be deleted. Another field of use for source separation: new orchestrations, for which I can keep just the voice. Inversely, in the case of karaoke-like applications, I can take the vocal track out.

What about sound dynamics?

In this field, it is precisely the new mix and therefore my work that brings a new dimension to the piece. However, what matters most to my ears is to respect the original work and not spoil it. That’s what I did for “La Môme”. Piaf’s songs and her so particular tone are famous, worldly recognized, and appreciated as such. Changing them just a little bit would have been unwelcome, and would have upset the audience.

Tell me about the residual artifacts on separated sources?

One strength of this technology is to maintain formally the integrity of the original piece. All separated tracks sum up to the input sound.

What kind of catalogs can Audionamix technology address best?

In sound processing, nothing is excluded a priori: the music of a CD, a musical DVD, an original soundtrack, alterated soundtracks from old movies and TV programs remastered or re-issued in Blu-ray. In other words, any media destined to a second sound life.

Audionamix Unmixing Station is presently available as a service. Do you agree that’s the best choice?

The obvious advantage is leveraging the skills and knowledge of a whole team of specialized researchers and developers, endlessly committed to addressing our needs and to improve the product. Downstream, sound engineers like me use it with our know-how. Dialog can sprinkle between both teams, sometimes producing finer solutions.

Sum up Audionamix technology in one word?

Unique!

ITV conducted on March 22nd, 2009, at Copra studios.