Tue Feb 19 11:56:14 CET 2008
On 2/19/08, kenye <kenyeby at yahoo.com.tw> wrote:
> We have a project about the multi-video-codec project.
> It is a hardware and software co-design project.
> We try to provide hardware accelerater for all video-codecs.
Which API will you use? Hopefully it will the decoding part use an
open source API like the one from the VaAPI (Video Acceleration API)
What video decoding processes exactly will it be able to accelerate effectively?
- Motion compensation (mocomp)?
- Inverse discrete cosine transform (iDCT)?
- Inverse modified discrete cosine transform (iMDCT)?
- In-loop deblocking filter?
- Intra-frame prediction?
- Inverse quantization (IQ)?
- Variable-Length Decoding (VLD)?
- Spatial-temporal deinterlacing?
- Bitstream processing (CAVLC and/or CABAC)?
I do not speak for FFmpeg nor its developers but my guess is that next
question from an actually FFmpeg developer is; will ASUSTek provide
some free hardware samples of this hardware accelerator (in a PCI/PCIe
card form-factor?) to any volunteering FFmpeg developer?, and if so
would it then be best for a FFmpeg team-member to collect the shipping
addresses of those volunteers and send them to you in a private
e-mail?, and/or will it be simpler for you to send a bunch of cards to
one FFmpeg team-member who volunteer to distribute within the FFmpeg
team? Also, how detailed documentation will you be able to provide for
those developers with and without a NDA (some might be willing to sign
a NDA, others may not but would still be willing to work on the
development of it)?
PS! Maybe a good Google Summer of Code 2008 project if no one
volunteer before then?
Best regards / Andreas Setterlind (a.k.a. Gamester17)
Project Manager of XBMC (a media center/player software which uses FFmpeg)
More information about the ffmpeg-devel