PDA

View Full Version : Save VBR analyzing results


davexnet
12th July 2003, 00:35
Hello,
there is this option in environment/cpu,
save analyzing results to cache.

It sounds as if the VBR analyzing info is somehow "saved"
so that if subsequently a project needs to be reencoded,
the analyzing stage need not be processed again.

How is this actually used?

I've not got any result with it, it always re-does the
analyzing stage even with this checked?

Dave

Kika
12th July 2003, 00:41
Newer Versions of TMPGenc do save the Results in a native format (RAW Data, a little like a Cache). It is used to speed up the second pass and is removed from the HD after Encoding or stopping the encoding.

Guess, there's no chance to use this Analysing-Data an other way than the one, it is created for (to do the speed up of the second pass).

davexnet
12th July 2003, 00:52
Thanks for the info - if it's as you say, then it seems to have
no use at all.

2-pass VBR by it's nature *must* save the results of the first pass,
so that it can refer to it when it's actually creating the mpg (2nd pass) and surely this will happen whether this box is checked or not.


If the above is true, then what extra does checking off that
selection do for me Vs. not using it?


Dave

Kika
12th July 2003, 01:09
Didn't you read what i wrote? The results of the first pass are saved. This is a must be in 2pass VBR. But the save Date can't be use by other Programs than TMPGEnc and you can't manipulate it like in CCE.

davexnet
12th July 2003, 01:24
Yes of course I read it.
But your explanation didn't *fit* with the way I thought it
worked - hence my ongoing confusion.

Are you saying that if that box is not checked, any given
2-pass encode will take longer

Kika
12th July 2003, 01:31
Yes, if you don't check the box "Save analyzing results... ", the second pass takes longer - a lot longer. But be sure, 2 pass is aktivated, not 2 pass (old Mode).

Try it by yourself, you will see, that the second pass nearly is as fast as in CCE if this option is checked, and it is much slower if you not check it.

davexnet
12th July 2003, 06:43
Hello, I did a test (repeated on two different clips), about 15
minutes each, 480*480, ntscfilm, vbr.

I encoded each 15 minute clip twice, once with the "Save analyzing results... " selected and once without. The encode time were
virtually identical.

There is a sub-folder called cache, and I noticed something does
get saved in here, it seems to be the ivtc data, and it works.
As long as you don't change the "project", the old data is used
on a re-encode,
saving *lots* of time. (it also stays in there forever and has to be manually deleted)

BUT, based, on my test, it seems to me as if "Save analyzing results... ", whether checked or not - does nothing.
Certainly no file that I can see gets "saved".

Dave

Kika
12th July 2003, 12:01
Then there's something wrong with your Machine, your settings or you are doing heavy Filtering/Resizing.

If i'm encoding 10 Minutes 704x576 MJPEG to 704x576 MPEG2, first pass takes ca. 25 Minutes, second only 10 - 12 Minutes. But if i'm doing filtering or something else than only encoding, second pass isn't faster. That's normal.

davexnet
12th July 2003, 22:25
Well I'm disappointed I didn't see anything.
I'll look at my setup and see if I can find any problems.
In the meantime, I have emailed tmpgenc directly to see if they
can provide details.

If I get back anything interesting I'll post it in here.

Dave