[theora] Indexing Ogg files for faster seeking

Bernard Jungen bjung50169 at euphonynet.be
Thu Jan 21 21:05:42 PST 2010

On Fri, Jan 22, 2010 at 04:17:40PM +1300, Chris Pearce wrote:
> On 22/01/2010 1:50 p.m., Bernard Jungen wrote:
> > On Fri, Jan 22, 2010 at 12:46:46PM +1300, Chris Pearce wrote:
> >    
> >> I previously tried compressing the indexes only with zlib (i.e. not
> >> delta-then-variable-byte-encoding them before zlib deflating them), and
> >> that got us about 50% compression.
> >>      
> > Have you tried delta+deflate, i.e. without variable encoding? And also
> > with each kind of data grouped together?
> >
> >    
> Seems that delta+deflate initially performs about 5-10% worse than 
> delta+variable_length, until it converges at around 2000 keypoints in 
> the media (about 2 hours long), where it's about 1-2% better. I think we 
> should keep it simple.

I agree about simplicity, but that only comes after testing possibilities.

Have you tried packing each kind together? What are the results for time deltas
alone? I ask because I strongly suspect keyframe offset deltas to be very
difficult for deflate to compress due to the tight zeroes/random data mix.

It would also be interesting to compare variable-byte-length vs
constant-bit-length coding IMO.

GPG: D3ED A92F D243 FC07 1881 BE2E E68A A45D A54A DA90

More information about the theora mailing list