Hi,

0n 03/08/[EMAIL PROTECTED]:53 Steven M. Schultz told me:

> 
> On Sat, 16 Aug 2003, Maik Holtkamp wrote:
> 
> > I already had a discussion on the dvdauthor list
> > (http://sourceforge.net/mailarchive/forum.php?thread_id=2942599&forum_id=13261)
> > on this topic, so sorry if doubled for someones.
> 
>       I'll give the same advice here that I've given in the past on
>       the dvdauthor list:
> 
>               Use a single file.

Yes, I will do so.

>       More about why this is necessary later on.
> 
> > I tried to make dvd5 from my dvd9. It is no problem when I use one
> > big file, but it will facilitate things here, if I could just use
> > chaptering:
>       
>       That is not the way DVDs work ;(
> 
>       DVDs are _track_ oriented.  Each file goes into a track (sometimes
>       called a "title or titleset").   Chapters within a title/chapter
>       play seamlessly as you have discovered - but transitioning across
>       a track boundary is not seamless.

I understand, hopefully ;), but I used:

dvdauthor -a.. -v.. 1.mpg 2.mpg ... <last>.mpg

and as I understand the readme of dvdauthor I am in one titleset,
when using this command:

---cut---
2. Create your titlesets
...
To create 1 chapter per mpeg, simply do
 dvdauthor [-o dir] [a/v/s options] chap1.mpg chap2.mpg chap3.mpg...
...
To add chapters every fifteen minutes, do
dvdauthor [-o dir] [a/v/s options] -c 0,15:00,30:00,45:00,1:00:00,
1:15:00... longvideo.mpg

Call dvdauthor for each titleset you want to create.
---cut---

As I understand both calls (-c <times> big or 1 2 3) should create
the same result: One titleset/track with chapters. 

Using the first one, I receive these freezes the sencond works.

> > [1] In spite of really understanding what -c is for, man page sounds
> >     promissing for me.
> 
>       Are you creating DVDs with multiple angles in the video stream?

No. I just read the new manpage and thought -c is probably a good
idea. I will skip in future, THX.

> > [2] Is there any chance to roughly estimate how the resulting
> >     file-size will be reduced when:
> >     decreasing -b for 100
> >     increasing -q for 1
> >     in variable bitrate encoding? Or is there a other option I sould
> >     use in order to get more predictable results?
> 
>       -b sets the maximum (+/- a couple percent of course) rate
>       (filesize).   This is the upper bound on filesize.  Use the
>       value of -b, add the audio rate, and then add about 2.5% for the
>       other information written to the stream.   Use that total and
>       multiply by the playing time - that is the maximum size of the
>       file.   Then vary -q according to the desired quality (lower -q
>       values will cause the file to more closely approach the max
>       filesize).
>       
>       As an alternative use constant bitrate encoding ;)

As I understand this:

The more I go down with -q the more close I go to constant bitrate
meeting my -b setting, right?

>       What I often do is encode a representative sample, between 5 and
>       10 minutes worth of data, and extrapolate the final size of the
>       file.    

My first goal was to automate the process. I found dvd2iso.pl
(http://james.nontrivial.org/projdvd.htm):

It uses transcode ...-w 3800 ... -N 0x50, -F 5 but I want 5.1 sound
and thought the bitrate is rather low. My test orginal (pulp fiction
PAL) was at 8800/1100 and the files I extracted had about 4.9 GB.

So there was no need for a drastical reduction to 3800.

I used -b 8000 -q 6 (got there by try&error fullencoding). In spite
I did not ran dvd2iso on the this disc I suppose that the difference
between 3800 and 8000 would have been visable.

Moreover I thought it is a better idea to start at -F 8. I konw it is
probably possible to extend the xsvcd profile to suit dvd needs, but
for my end there are enough traps left even, when starting at
profile made for my needs.

So I skipped dvd2iso.pl and tried on my own, but ...

>       The effect of -q is data sensitive and there is not a fixed/firm
>       relationship between increasing -q by 1.  I have seen in some
>       cases going from 5 to 6 a change in the average bitrate around 15%
>       but that varies from run to run depending on the data.

..now, I understand that the prozess is hardly automatable :(. I hope
my guessing of good values for -q and -b will increase in line with
experiance.

Thanks for your explanations.

-- 
bye maik


-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01
_______________________________________________
Mjpeg-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/mjpeg-users

Reply via email to