# [FFmpeg-devel] matroska encoder option -- force new Cluster on keyframe

Michael Niedermayer michaelni at gmx.at
Sun Jul 21 12:40:22 CEST 2013

On Sun, Jul 21, 2013 at 11:51:53AM +0200, Nicolas George wrote:
> Le tridi 3 thermidor, an CCXXI, Bernie Habermeier a écrit :
> > But I explicitly WANT to have an option that doesn't "take it under
> >
> > Specifically, I want to control exact placement of my I-frames, and
> > therefore CuePoints, and I want to a 100% guarantee that I have a new
> > Cluster start for each such CuePoint.  What you're proposing does not do
> > that at all, and has a completely different feel to it.  We can certainly
> > add this kind of control, and I'm fine implementing that, but that is
> > specifically not what I want.
>
> I am not sure the direction of your requirement, but I believe the current
> logic with configurable limits would work for both:
>
> * Positive: if you WANT a new cluster for each key frame, irregardless of
>   the size/duration/age/whatever of the current cluster, -min_cluster_size 0
>   (instead of 4k) should do that.
>
> * Negative: if you WANT NO new cluster except at key frames, then
>   -max_cluster_size \infty -max_cluster_duration \infty (instead of 5M/5000,
>   or 32k/1000 for streams) should do that.
>
>   (lavf does not have \infty, but INT64_MAX should do the trick unless you
>   intend to produce a eight etaoctets file longer than the history of
>   mankind.)
>
> And of course, if you want both, you set all these options like that.
>
> Am I missing something?

user interfaces, convenience and humans

If we assume that the user wants "start a new cluster on every new
cue point"
an option that has a matching name will be found, used and will
make the user happy.

Having to tune 2+ nummeric parameters, will not be found/realized by
99% of users, so 99% of the users who want to use this specific feature
wouldnt be able to.
And the 1% would probably find it inconvenient, but then maybe some
of these 1% would also find it beatifull in some mathematical way
but then still it fails for thr 99% of normal users so i dont think
thats an argument here.

open question is how many people will want to
"start a new cluster on every new cue point"
how many people will want to tune one or more of the nummeric
parameters
and of course
is that cluster start heuristic optimal to begin with ?

[...]

--

If a bugfix only changes things apparently unrelated to the bug with no
further explanation, that is a good sign that the bugfix is wrong.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20130721/ac117949/attachment.asc>