[asterisk-dev] Opus and VP8

[Olle E. Johansson]

Matt,
Thank you very much for your clarification. I have a lot of understanding
for Digium's position in this case, even though I don't like that this is happening.

One small comment:

29 maj 2013 kl. 20:55 skrev Matthew Jordan <mjordan at digium.com>:

> VP8 is the easier of the two to clarify. A codec for VP8 is probably not
> appropriate, regardless of any patent or IPR issues. Asterisk doesn't
> perform video transcoding. Video transcoding is an intensive operation
> that performs poorly without hardware augmentation. We've always taken
> the stance that software video transcoding in Asterisk would cause more
> problems then it would solve; as such, VP8 as a codec is best left
> outside of Asterisk.

I don't agree here. We need video codecs not only for transcoding, but 
also for reformatting between different frame rates, different frame sizes.
Inserting frames for IVRs - and possibly a lot of other innovation.

When Mark created Asterisk people thought he was a fool to believe
that Asterisk could transcode on a normal CPU without any DSPs.
He did prove them wrong and the stuff we're doing today is amazing.

In many cases today, I install Asterisk in 4*quad core servers and
use less than 5% of the raw cpu power. Early experiments with WebRTC
has shown a lot of cool stuff on normal laptops - merging video with
images, text or just merging multiple videos.

I do believe that having video codecs in Asterisk makes a lot of sense
and will open up for a lot of new applications. If we can't have it in
software, I hope that we can get licensed codecs in hardware cards
with standardized API's to enable video codecs in Asterisk.

I do want this innovation to happen in Asterisk.

/O