Vorbis - Dec 2002 - Technical MP3->OGG transcoding question

> CAN it be done in a lossless way?
No.
> And are the two file formats THAT incompatible between each other?
Yes, they are.


-- 
Graham Mitchell - computer science teacher, Leander High School
Cats are Spiderman,
Cats are Spiderman,                                                   ^ ^
Cats are Spiderman,                                                   o o
NO-body KNOWS who they ARE!                                          ==Y=
--- >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-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.

Hi guys.

Let me put this first: I've read every past thread about mp3 -> ogg
transcoding and I agree to the fact that it is BAD--for several reasons
indeed.

This post is just a technical question, which I hope someone who knows
stuff better than I do will answer me.

Another disclaimer: although I'm a programmer, I've never worked with
anything audio-related, nor do I know anything technical about Mp3 and
Ogg Vorbis. So everything you will read from here on is PURE GUESSWORK.
Feel free to flame me, bash me, etc. as long as you correct me! ^__^

But let's skip to the main question:

<p>QUESTION:
---------
Let's assume one wants to transcode an Mp3 file to Ogg Vorbis, for his
own reasons, which I'm NOT going to discuss here.
CAN it be done in a better way than Poor Man's Transcoding? 
(i.e.: mpg321 -w - file.mp3 |oggenc -o file.ogg - )
CAN it be done in a lossless way? 

<p>As I said I don't know anything technical about Mp3 and Ogg Vorbis.
I
suppose that both formats contain the outcome of a cosine transform
performed on the audio data, or something along those lines. 
Better said, they must contain a particular _clipping_ of the cosine
transform, specifically computed in order to discard most information
while retaining some "hearing quality".

As such, the differences between mp3 and ogg vorbis can be of the
followin types--and ONLY of them:
- Fundamental differences in the "cosine transform" or such
mathematical
  transformation of data.
- Differences in the "clipping" (ie. "discarding") part of
the encoding
  process (which of course must be the hardest part of the whole encoding
  process, because it involves psycho-acoustic and such stuff)
- Differences in the way the resulting encoded data is stored in the
  file--that is, differences in the file format.

Am I totally wrong here?  ?__?

I suppose the latter two aspects are very different between mp3 and ogg.
In particular the fundamental differences in the second issue are
exactly why Poor Man's Transcoding almost always produces ogg files
which are sensibly worse than the "original" mp3.
But is the first issue ANY different between Mp3 and Ogg Vorbis?
And are the two file formats THAT incompatible between each other?

That brings me to a better phrasing of the original question:

<p>QUESTION IN OTHER TERMS:
------------------------
Is it possible to pick up the "frames" stored in an Mp3 file, one by
one, and write them off as Ogg Vorbis frames?

<p>Please don't answer me: "No, because it's
wrong/evil/<put your own ethic
adjective here> to do so!". That's not a proper answer to a
technical
question! :-)

<p>TobiX
--- >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-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.

TobiX wrote:
> Let's assume one wants to transcode an Mp3 file to Ogg Vorbis, for his
> own reasons, which I'm NOT going to discuss here.
> CAN it be done in a better way than Poor Man's Transcoding?
> (i.e.: mpg321 -w - file.mp3 |oggenc -o file.ogg - )
> CAN it be done in a lossless way?
This has been discussed before as well ... I forgot the exact technical
reason why, but MP3 and Vorbis are too different - it's not possible to
transcode without completely decoding MP3 and re-encoding with Vorbis.
This is also the main reason why it's "evil", there is just no
other way
than the way of the Poor Man.

<p>Moritz
--- >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-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.

('binary' encoding is not supported, stored
as-is)
> QUESTION IN OTHER TERMS:
> ------------------------
> Is it possible to pick up the "frames" stored in an Mp3 file, one
by
> one, and write them off as Ogg Vorbis frames?
I also thought about this, and I came to the conclusion, that, even if it would
work this way, you'd never get a REAL ogg file. For example the progressive
Format of the data, the peelers use, couldn't be achieved this way. (I know
peelers are not yet good, but as far as I got it, it already works, just not as
good as it could)
So I decided not to spend any further thoughts about it, cause it would be
really bad if the ogg files in a collection are of a "different"
format and
can't all be treated the same way.

Just my 2 cents.

-- Daniel

<p>You know it's going to be a bad day when you want to put on the
clothes
you wore home from the party and there aren't any.
--- >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-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.