AAF processing

Questions and answers on how to get the most out of FFAStrans
emcodem
Posts: 1811
Joined: Wed Sep 19, 2018 8:11 am

Re: AAF processing

Post by emcodem »

I see... thanks benjamin. No worries, this is not over... i just have a feeling that opentimelinio is the right way to go and they seem to be pretty close to the goal... It might already work with simple sequences, i'll give it a shot in the next days
emcodem, wrapping since 2009 you got the rhyme?
momocampo
Posts: 595
Joined: Thu Jun 08, 2017 12:36 pm
Location: France-Paris

Re: AAF processing

Post by momocampo »

Great!
You can also wait others feedbacks to see what are others opinions.
Cheers ;)
Ghtais
Posts: 164
Joined: Thu Jan 19, 2017 11:06 am

Re: AAF processing

Post by Ghtais »

Hi all

thank for all Emcodem. I understand that it is a huge project and it needs a lots of time and some money of course.
As we discuss before mixdown is not an very good option as it already works with quicktime reference (even if we loose some metadata).
I will continue to check this thread for news on this very interesting subject :)

bye
nknuelle
Posts: 19
Joined: Mon Jul 01, 2019 4:39 pm

Re: AAF processing

Post by nknuelle »

emcodem wrote: Fri Oct 11, 2019 1:45 pm Maybe @taner can help here, as the createaaf stuff worked for him on interplay viewtopic.php?f=5&t=892&start=50#p3952
@taner, maybe you could upload one of the aaf files that you created with the createaaf 0.0.0.1 tool? :-) Did you ever update to the recent version of the createaaf stuff and does it work for you?
Hi again,

I downloaded the updated AAF tool which still has the same problem with file input:

Code: Select all

C:\Users\Administrator\Desktop>createaaf-new.exe --debug "S:\Avid MediaFiles\MXF\KS-FFASTRANS\"
usage: createaaf-new.exe [-h] [--debug DEBUG] [--odir ODIR] [--oname ONAME]
                         [--testmode TESTMODE]
                         FILES OR FOLDERS [FILES OR FOLDERS ...]
createaaf-new.exe: error: the following arguments are required: FILES OR FOLDERS

C:\Users\Administrator\Desktop>createaaf-new.exe --debug "S:\Avid MediaFiles\MXF\KS-FFASTRANS\nina-schmidt_www-Stimme-Professor-K1-XQD-15von15_a1.mxf"
usage: createaaf-new.exe [-h] [--debug DEBUG] [--odir ODIR] [--oname ONAME]
                         [--testmode TESTMODE]
                         FILES OR FOLDERS [FILES OR FOLDERS ...]
createaaf-new.exe: error: the following arguments are required: FILES OR FOLDERS

C:\Users\Administrator\Desktop>createaaf-new.exe --debug "S:\Avid MediaFiles\MXF\KS-FFASTRANS\nina-schmidt_www-Stimme-Professor-K1-XQD-15von15_a1.mxf" "S:\Avid MediaFiles\MXF\KS-FFASTRANS\nina-schmidt_www-Stimme-Professor-K1-XQD-15von15_v1.mxf"
['S:\\Avid MediaFiles\\MXF\\KS-FFASTRANS\\nina-schmidt_www-Stimme-Professor-K1-XQD-15von15_v1.mxf']
Detected file from userinput: S:\Avid MediaFiles\MXF\KS-FFASTRANS\nina-schmidt_www-Stimme-Professor-K1-XQD-15von15_v1.mxf
Added S:\Avid MediaFiles\MXF\KS-FFASTRANS\nina-schmidt_www-Stimme-Professor-K1-XQD-15von15_v1.mxf
Created file: S:\Avid MediaFiles\MXF\KS-FFASTRANS\nina-schmidt_www-Stimme-Professor-K1-XQD-15von15_v1.aaf
Done
As you can see if I use 1 file path as input it throws an error. If I input 2 filepaths it recognizes only the second one..

This is the generated AAF:
nina-schmidt_www-Stimme-Professor-K1-XQD-15von15_v1.aaf

I can checkin this AAF into Interplay but all assets (mxf files) are recognized as AMA-linked..

So no real progress over here..
emcodem
Posts: 1811
Joined: Wed Sep 19, 2018 8:11 am

Re: AAF processing

Post by emcodem »

Hm i wonder how come that the files are now recognized on interplay, it still uses link external mxf from pyaaf2...
To make ot work with both files, instead of --debug just use --debug 1
emcodem, wrapping since 2009 you got the rhyme?
draldos
Posts: 3
Joined: Fri Jun 05, 2015 7:40 am

Re: AAF processing

Post by draldos »

Hi @emcodem
Was speaking to Steinar about this the other day.. Did you get anywhere with reading AAFs. We are looking at alternatives for QT Ref from Avid Media Composer to MP4 and wondered if a linked to AAF with effects rendered would work. I can see this thread is old but wondered if you had looked at it any further. Thanks S
emcodem
Posts: 1811
Joined: Wed Sep 19, 2018 8:11 am

Re: AAF processing

Post by emcodem »

Hey @draldos

Yeah well, i spent some time in looking at an alternative for quicktime player regarding qt ref but until now i didnt find any library that is able to resolve the full timeline. FFmpeg only supports "one single clip" in a qtref, so that didnt lead anywhere. And same as resolving the timeline of an aaf, it would be too much work for a sideproject to come up with a fully blown qicktime decoder.

Regarding aaf decoding, as said earlier, i was counting on pixar to get it done but they only support "some" timelines in OpenTimelineIO. I didnt want to interface their stuff until they support at least the samples that were uploaded in this thread.

https://github.com/PixarAnimationStudio ... issues/806

The best "alternative" would be to work with some timeline export format that describes it in an easy to parse format, like FCP7 xml or some other easy to parse edl format.
emcodem, wrapping since 2009 you got the rhyme?
draldos
Posts: 3
Joined: Fri Jun 05, 2015 7:40 am

Re: AAF processing

Post by draldos »

Hi @emcodem,

is there a way to go AAF to XML? Maybe with pyaaf
emcodem
Posts: 1811
Joined: Wed Sep 19, 2018 8:11 am

Re: AAF processing

Post by emcodem »

@draldos

the thing is that the whole issue is with pyaaf (even the one with opentimelineio), and it doesnt look like markreid (owner of pyaaf) is very active
on the project. And a plain conversion from aaf to xml does not take out the complexity, it just changes the data storage format but it does not make the structure more easy to parse.
2 years ago, i was under the impression that it should be possible to use pyaaf2 to parse the timeline out of the aaf just as i would expect opentimelineio to do it, so i started the work:
https://gist.github.com/emcodem/8ec146c ... ec29520e47

but after lots of work and testing, it occured to me that this work will go on for a veeery long time until one has covered all possible cases that can occur. Thats why i came to the conclusion that writing an aaf2edl parser would be too much for the ffastrans team, it needs to be a separate project and most likely it should be funded by someone, otherwise it would end up like the pyaaf2 project - work on it would just stop half the way.
emcodem, wrapping since 2009 you got the rhyme?
Post Reply