Vorbis dev - Jul 2002 - [fwd] CVS: ogg123 rocks! vcut no so much so... (from: wayfarer42@postmaster.co.uk)

Some detailed feedback that looks useful.

----- Forwarded message from Michael Semon <wayfarer42@postmaster.co.uk>
-----

Delivery-Date: Tue Jul 16 05:29:37 2002
X-Mailer: MIME-tools 5.41 (Entity 5.404)
From: "Michael Semon" <wayfarer42@postmaster.co.uk>
Subject: CVS:  ogg123 rocks!  vcut no so much so...
To: feedback@vorbis.com

Hi!  Being impatient for RC4, I downloaded enough of cygwin so I could CVS
ao/ogg/vorbis/vorbis-tools and transfer it to my Internet-free Linux boxen. 
[Can't there be an easier way?]  Reviews so far:

ogg123 rocks!  Before, I had to make "playlists" using symbolic links.
Now, I just make a playlist, like any sane human being would do.

I wish that ogg123 -@ *.m3u would concatenate playlists, instead of considering
the first playlist as a playlist and the rest as failed ogg files.  Also, I wish
there was an option for ogg123 to sort directories.  At least with ReiserFS on
Linux, file order in a directory is basically random.

Oh, for oggenc on my files, RC3 used 62% CPU on my Pentium 75.  The latest CVS
uses 44% CPU.  Good work!  Also, I'd be glad to share my recording
processes, though they'd not be nearly as sexy as the vorbis.com articles.

Full bitrate management works but is slow.  No big deal, at least it works now.

I couldn't get --advanced-encoder-option to work.  However, it seemed to
work as --advanced [option=whatever].

vcut is still rough.  Sometimes, I'll get the "file too short?"
errors as I'm trimming four-hour-long files, and this failed:

vcut 0100rhc.ogg a.ogg b.ogg $((9600*60*202))

...but this will work...

vcut 0100rhc.ogg a.ogg b.ogg $((9600*60*202-1200))

Note that vcut can make files in which playing will succeed but seeking will
fail.  In the past, I'd test a file by doing something like `ogg123 foo.ogg
-k 15`.  If the file works in ogg123, then it will work in Winamp and Xinf.  RC3
ogg123 would be graceful, but CVS ogg123 will hang to the point where Ctrl-C
doesn't work, and I have to `killall ogg123` from another virtual console. 
Something to work out?

I'm not sure what to make of ogginfo.  With RC3 ogginfo, I could do
something like this...

find -type f -name *.ogg -exec ogginfo {} \; >> files.txt
grep "^total_playtime=" files.txt > lengths.txt

...or grep whatever.  I'm looking for a new strategy to replace my old
"grep header_integrity=" test.  That's not a request for help,
only an observation that I like the geeky RC3 ogginfo program better so far.

Anyway, that's a status report and nothing more.  I was merely being
impatient.  I usually record 11 hours of radio per day (6 hours automated + 1
hour automated + 4 hours started by hand), and RC3 did not fail once in many
months of use.  The latest CVS has passed the 4-hour test and the 1-hour test,
and I expect it to pass the 6-hour test as well.  It doesn't sound very
different at 9600-Hz mono, -q1.25, but I like the progress anyway.

Good luck!

Michael L. Semon

<p>----- End forwarded message -----

<p>--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

>I couldn't get --advanced-encoder-option to work.  However, it seemed to
work as --advanced [option=whatever].
It's --advanced-encode-option (there was a mistake at one point in the 
manpage in 1.0 which indicated --advanced-encoder-option. Not sure why
--advanced works at all - maybe some quirk of getopt.
>
>vcut is still rough.  Sometimes, I'll get the "file too
short?" errors as I'm trimming four-hour-long files, and this failed:
>
>vcut 0100rhc.ogg a.ogg b.ogg $((9600*60*202))
>
>...but this will work...
>
>vcut 0100rhc.ogg a.ogg b.ogg $((9600*60*202-1200))
>
>Note that vcut can make files in which playing will succeed but seeking will
fail.  In the past, I'd test a file by doing something like `ogg123 foo.ogg
-k 15`.  If the file works in ogg123, then it will work in Winamp and Xinf.  RC3
ogg123 would be graceful, but CVS ogg123 will hang to the point where Ctrl-C
doesn't work, and I have to `killall ogg123` from another virtual console. 
Something to work out?
Not sure about the ogg123 bug, but vcut is known to be very buggy. I wrote
it more as a proof-of-concept than as something for people to really use
much. It works for simple cases, but doesn't handle some things at all.
>
>I'm not sure what to make of ogginfo.  With RC3 ogginfo, I could do
something like this...
>
>find -type f -name *.ogg -exec ogginfo {} \; >> files.txt
>grep "^total_playtime=" files.txt > lengths.txt
>
>...or grep whatever.  I'm looking for a new strategy to replace my old
"grep header_integrity=" test.  That's not a request for help,
only an observation that I like the geeky RC3 ogginfo program better so far.
ogginfo in rc3 didn't actually work correctly in many, many cases. So I
rewrote it, and did the output in a nicer way. ogginfo was never intended
to be a tool for automated stuff like that (though you can doubtless still
do something), part of the reason for the new output format was a concious
desire to discourage it further (too many broken programs were using it
in particularly bad ways).

<p><p><p>--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

[I can't seem to get through from a different unsubsribed address
(scben@t2.technion.ac.il).  I don't get any TMDA message requesting
confirmation either.  Anyway here it is...]

Quoting Monty forwarding Michael Semon
<wayfarer42@postmaster.co.uk>:
>
> Hi!  Being impatient for RC4, I downloaded enough of cygwin so I could CVS
> ao/ogg/vorbis/vorbis-tools and transfer it to my Internet-free Linux boxen.
> [Can't there be an easier way?]  Reviews so far:
>
Nigthly cvs tar.gz snapshots?  Look on xiph.org for "download" (iirc).
> ogg123 rocks!  Before, I had to make "playlists" using symbolic
links.  Now,
> I just make a playlist, like any sane human being would do.
>
As Segher once pointed out, unix tools need no playlist support:
ogg123 `cat playlist` does the job (unless you spaces in file names, I wonder
why bash has no split-by-lines command substitution variant...).  Using
ls, find, vorbiscomment instead of cat you can have dynamic playlists windows
users don't even dream of...
> I wish that ogg123 -@ *.m3u would concatenate playlists, instead of
> considering the first playlist as a playlist and the rest as failed ogg
> files.
That would be broken, it would mean the -@ treats all rememaining arguments
as playlists.  ogg123 `for f in *.m3u; do echo -@ $f; done` will produce the
needed command line.  But in the spirit of Segher's idea, ogg123 `cat *.m3u`
should work (modulo spaces, again) and ogg123 -@ <(cat *.m3u) is good too
(bash has this nifty temporal named pipe feature, don't know for other
shells).
> Also, I wish there was an option for ogg123 to sort directories.  At
> least with ReiserFS on Linux, file order in a directory is basically
> random.
>
Again, ogg123 `ls *.ogg` or ogg123 `find -name '*.ogg' | sort` should
do.
"A program should do one thing and do it well."


-- 
Beni Cherniavsky <cben@tx.technion.ac.il>

<p><p><p><p>--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.