[CELT-dev] Just a question on how things are going

Jean-Marc Valin Jean-Marc.Valin at USherbrooke.ca
Fri Aug 6 10:21:10 PDT 2010


Hi John,

John Ridges <jridges at masque.com> a écrit :
> I was just wondering if maybe you would consider a "blog-like" text file
> in the codebase where you just jot down your thoughts about why you are
> making a change, and where you think you are headed.

Actually, I was planning on commenting more on my LiveJournal blog:  
http://jmspeex.livejournal.com/

> For example, I
> noticed today that you completely removed the pitch prediction code. It
> would be nice to know if you did that because you're moving away from
> using it as a speech codec, or you just found that the pitch predictor
> just wasn't making any difference. I realize that such a "blog" might
> elicit endless debate from the peanut gallery on every change you make,
> so I'll understand if you think it's a bad idea.

The reasons for the change are:
1) Pitch has always been useful only on a narrow range of bit-rates
2) It's responsible for a large increase in complexity
3) CELT is going to end up merged in the IETF interactive codec, which  
will also include SILK for voice.

> Also I was hoping you might give us a sense of how close you think you
> are to a 1.0 candidate. I don't want to put you on the spot, but for
> those of us that might be deploying a product with CELT soon, if we
> thought you were just cleaning up we might be willing to wait a little
> while so we could deploy with the frozen bitstream. If, on the other
> hand, you still have a bunch of ideas to try out, we'll work with what
> you have (kindly) given us so far. (I guess one clue is that you're
> quickly running out of 0.x version numbers :-)

As things are now, it's quite likely that CELT will eventually cease  
to exist as a stand-alone codec and instead be included in the IETF  
codec. That codec would basically get the best of SILK and CELT. We're  
progressing quickly on this, but I couldn't really tell you when it  
would be frozen (depends on the working group). On the CELT side, I'm  
working hard on pretty much freezing that part of the bit-stream, so  
there's light at the end of the tunnel. For that reason, I'd like to  
see it being tested a lot to make sure that the final bit-stream is  
good.

Cheers,

      Jean-Marc





More information about the celt-dev mailing list