seek tables for MP3s?

Archived from groups: rec.audio.tech (More info?)

Is it possible to create MP3s with seek tables? I've seen just very brief
references to them, usually in connection with Shorten or FLAC. I use FLAC
format extensively, but for some recipients MP3 is really warranted. The
main use I would have for seek tables with MP3 is that such a thing would
allow taking a whole concert or other multipart program, and making it
available with divisions to which the listener could skip, but without the
audible gaps that would occur in listening to a succession of MP3 tracks.

Is there a program out there that could take a group of WAV files making up
a concert, and join them together, encode a large MP3 for the ensemble, and
include a seek table with pointers to the original file transitions?

A related question would be: How many popular MP3 players (software mainly,
but hardware, too) work with seek tables?

Archived from groups: rec.audio.tech (More info?)

> Those gaps are a matter of your player, not the MP3 files. Adjust its
> configuration or switch to one that allows a zero-length inter-track gap.

I don't think it's that simple. You may be able to eliminate the actual dead
space--there are plugins that supposedly do that, and they probably do. Even
so, when you've encoded a bunch of separate MP3 files, the nature of the
encoding process means that there will not be a seamless transition from one
track to the other, since the adjoining tracks were processed independently
of each other. So there will be an audible discontinuity, even if there's no
dead space.

So I'm back to wanting to combine a set of WAV files (that were split
seamlessly from a larger original), create a single large MP3, but retain
track length information and put it in a seek table. Is that doable? Is it
being done?

Archived from groups: rec.audio.tech (More info?)

RWG wrote:

> Is it possible to create MP3s with seek tables?

I think what you are asking about is more commonly known as cue lists.

> I've seen just very
> brief references to them, usually in connection with Shorten or
FLAC.
> I use FLAC format extensively, but for some recipients MP3 is really
> warranted. The main use I would have for seek tables with MP3 is
that
> such a thing would allow taking a whole concert or other multipart
> program, and making it available with divisions to which the
listener
> could skip, but without the audible gaps that would occur in
> listening to a succession of MP3 tracks.

That's known as a cue list when applied to .wav files.

> Is there a program out there that could take a group of WAV files
> making up a concert, and join them together, encode a large MP3 for
> the ensemble, and include a seek table with pointers to the original
> file transitions?

The way this is commonly done is to create a single .wav file that has
an imbedded cue list. The boundaries of the component .wav files are
in the cue list. Some CD burning software builds the proper
structures on a CD, when the CD is burned from the .wav file that has
the imbedded cue list.

Programs like Audition provide the means for both creating the cue
list and burning a CD from .wav files with imbedded cue lists. CD
burned this way are burned using a standard feature called "Track at
once". The entire disc is composed of one file, and cue points are
used by the CD player to access the boundaries of the individual .wav
file that were concatenated to create the actual .wav file that was
used to burn the disc.

A freebie called Cuelisttool can be downloaded from
http://www.stefanbion.de/cueltool/index_e.htm , and helps imbed and
manage imbedded cue lists. The web site also contains some helpful
technical information about using cue lists.

> A related question would be: How many popular MP3 players (software
> mainly, but hardware, too) work with seek tables?

A few posts like this one

http://forums.etree.org/viewtopic.php?p=7551

and web sites like this one

http://www.true-audio.com/codec.format

mention seek tables and programs that support them.

Archived from groups: rec.audio.tech (More info?)

> If the player is intelligent enough to preload the new data stream, and
> if the break occurs in such a way that the after-decoding last byte and
> first byte are reasonably continuous -- which shouldn't be hard to
> achieve -- it _ought_ to be possible to play straight through. Worst
> comes to worst, a bit of interpolation could be applied to the waveforms
> and any error would be sufficiently momentary not to be noticed, much
> like error recovery on CDs.
>
> I really think this is a player issue, not a data issue.

All I can tell you for sure is that when I've used Gapless, there were still
audible discontinuities in concert material, though they were less evident
than without it. Things like Gapless are easy to do, so probably worthwhile.
But I still think there will be some audible artifacts. IAC the kind of
thing (on the encoding end) to which I referred does not seem to exist, so
at this time there's no need to concern myself with it further.