https://bugs.kde.org/show_bug.cgi?id=491074

--- Comment #2 from Uwe Dippel <[email protected]> ---
(In reply to emohr from comment #1)
> To summarize: Your idea is in the render dialog to have a choice, like a
> tick box to render out:
> 1.    Video + audio track 1 (ambient) + audio track 2 (English) + subtitle
> track 1 (English)
> 2.    Video + audio track 1 (ambient) + audio track 2 (German) + subtitle 
> track
> 1 (English)
> 3.    Video + audio track 1 (ambient) + audio track 2 (French) + subtitle 
> track
> 1 (English)
> 
> Now, Kdenlive is so “clever” that it renders number 1. and then just the
> audio of number 2 (German) and 3 (French) and add it together that you have
> a complete number 2 and 3. Then you have a minimized render time with a
> maximum on output file.
> 
> Is that correct?

I'm not so sure I get your comment right. I know that I kind of overloaded the
initial request. Maybe we should separate the subtitles all together, since as
of now I haven't found a way to add existing sub title files at all (in the
time line; not with ffmpeg in the end)?

I think, we should offer the user much more flexibility w.r.t output file(s).
In some project, he'll want a single output file with all audio tracks mixed
down (ambient(A1), background(A2), music(A3), voices(A4) ...). Another one
(like in your example), the user will want a single output file, but with
separate audio tracks; to be selected at reproducing (either English or French
or Japanese). In this case, including one or more sub title tracks to be
selected from, or deselecting all. 

As of now, kdenlive is so "clever" (I hope you like irony as much as I do),
that it mixes down all audio tracks automatically together with video (e.g.
mp4); and on request ("separate audio tracks") all tracks as individual mp4
audio tracks. 
[I have written some utilities based on ffmpeg to handle the resulting streams
for my needs.]
These mp4 audio tracks take a hell of time, because they actually render as
video, just without the video information (which is perfect only for 100%
correct timing, while AC3 would be like "fire and forget" in this respect; and
yet sufficiently precise for most cases).

Yes, for my most frequent purpose, a single output file of V1 plus, including
English(A1), French(A2), ... Japanese(An) as MP4 with AC3 audio would be
perfect. Including the chance to (de)select individual tracks for that unified
output. Which isn't possible yet. -> I have to go to the "Render" window and
remove the tracks that I don't want as individual output files. 
Therefore, the "Render" dialogue might be a good choice, though I think it
would be easier and clearer, to mark/tick the tracks that one wants in a SINGLE
output file, why not in the time line? Like
V2 I
V1 -
A1 I
A2 -
A3 I
will render V2 together A1 and A3 into  a single file. 

Now it will be a design decision: What to do for an extra file: either go back
to the timeline, and select another combination, and go to "Render" again, and
so forth. 
Or, to use a matrix, with the tracks as lines, and the columns for output
files. So 3 columns would mean three output files. 
(My personal preference would be the first method, though.) 

The last decision then is on the eventually desired mixing down to single
tracks. I propose to do similar, only with another identifier. Like
V2 M
V1 M
A1 -
A2 M
A3 M
will mix down V1 and V2 into a single video stream, and A2 and A3 into a single
audio stream.

For the first user as above, who uses kdenlive to assemble a clip, he'd select
V2 -
V1 I
A1 M
A2 M
A3 M
A4 M
A5 M

I THINK that this is a method with maximal flexibility for all possible uses.
Kind of a construction kit for one or more output files for any combination of
tracks (streams). Including eventual subtitle tracks.

Sorry, this was pretty long, and hopefully still comprehensible. If not, just
keep asking! I'll be glad to explain it further!

-- 
You are receiving this mail because:
You are watching all bug changes.

Reply via email to