***UPDATED*** preliminary OMF Support

[kerryg]

OK, here are two OMF files generated by Ensoniq PARIS. The first is an export of one of the demo songs that originally came with the app. It's 245 meg and it fails with: "Can't create file (null) (object 66059)"

http://kerrygalloway.com/PARISRELATE...0no%20auto.omf

But the second is very interesting. It's a small test OMF I just made, and astonishingly it actually imports. There are a range of issues - files don't have proper names (this was actually always a problem with PARIS OMF export), neither do tracks (all are called "null), and the framerate and tempo are incorrect.

But I'm just a bit blown away right now at the fact that it opened at all [not, incidentally, because I doubt your work, but because this is an eight year old app whose OMF never functioned properly]:

http://kerrygalloway.com/PARISRELATED/OMF%20TEST.omf

The original audio file format used in the song is PAF, discussed in the PARIS thread here - http://forum.cockos.com/showthread.php?t=33102 Xenakios' libsndfile wrapper (found here - https://stash.reaper.fm/v/2421/libsndfilewrapper.zip ) may be needed for playback, I can't really tell what it's done with the audio (there's no file extension and no info as to file type).

[John Lundsten]

I'm getting a "reaper needs to close..." message
Reaper V 2.58 & using beta 5 omf.DLL

The OMF is an export from pro convert, which can open & understand file as can Neuendo 2

[airon]

V3 beta is required to use the OMF importer.

Find it in the Pre-Release Forum. I hope it works for you then.

[kerryg]

Originally Posted by airon

V3 beta is required to use the OMF importer.

Find it in the Pre-Release Forum. I hope it works for you then.

I actually didn't know that; my PARIS import results were achieved by using it with 2.58.

[404NotFound]

Originally Posted by airon

V3 beta is required to use the OMF importer.

Find it in the Pre-Release Forum. I hope it works for you then.

Actually, no. I don't use anything V3 specific.

[airon]

Sorry lads. I'd misunderstood something about new API stuff a while back. My mistake.

"Nothing to see, move on..."

[Bastiaan M]

I'm having problems with the OMF import. I created an OMF in AVID and i can't import it. It gives me a "Reaper stopped working" message. The OMF itself imports without problems back into AVID. The OMF itself is about 38 MB big, so thats way less the the 200MB's that will give problems. The OMF has some EQ's applied in AVID wich show up in AVID. I uninstalled Reaper and reinstalled it and i also installed the omf.dll again.

Any ideas?

edit: running Vista home premium here on a Toshiba A200 1P8 core 2 duo laptop
edit: you can download the omf HERE

[John Lundsten]

Originally Posted by John Lundsten

I'm getting a "reaper needs to close..." message
Reaper V 2.58 & using beta 5 omf.DLL

The OMF is an export from pro convert, which can open & understand file as can Neuendo 2

404 have you had any results running the very small OMF I included with this message?

[Bastiaan M]

Originally Posted by John Lundsten

404 have you had any results running the very small OMF I included with this message?

This sounds like the same problem i'm having.

Just thought of something: my OMF is at 25FPS. Maybe that explains something....

[antiClick]

Hello!
Thanks so much for your efforts on including OMF support in Reaper. This is a mandatory feature when doing audio for video.

I tested your extension by importing a FINAL CUT omf file, but failed after a minute of processing. I'd like to provide a log for the error, but didn't find any. The error seems related to some audio file names having characters with accents (?).

If you have the time for testing, I just uploaded the final cut OMF file here:
ftp://public@canreco.homeftp.net/PUBLIC/OMF/



Thanks a million for your efforts

[Andrew]

I'm trying to get some Sonar projects into Reaper.
I've played with this a bit today. here's my experience, exporting from Sonar 8 all audio no MIDI...

.omf opened in Reaper 2.58 under windows 2000 (I have old Aarks)
Reaper 2.58 crashed while attempting to open under Vista
Reaper beta 3 opened the .omf in Vista
Both set-ups have the following issues:
in Sonar many of the clips have the same name, they opened with the same audio, but on different tracks.
(I'm assuming this is a problem with the sonar export and naming convention?)
I gave each clip a unique name, exported again, and opened with 2.58 under win2000
the clips that are anchored the beginning of the project were fine and lined up.
Other clips were in relation to each other on their tracks, but did not line up with the rest of the project.

Hope this info helps!
Yours,
A-

[unmateria]

for me worked great after some tries... Im using it for converting my last cubase projects into reaper... I managed to get it working even with big projects, with this setup

http://www.losgen.com/x.jpg

(im sorry, I dont know how to insert images here)

thanks for your work man! solved me the last thing that was hailing me in cubase If you need some sort of beta testing you can add me

[mim]

Yeah it works with Nuendo 3/4, too ! quite long but it's ok ...

The only thing I miss is Volume and Pan Automations, do you think it is possible to get this working ?

Good work 404

mim

[404NotFound]

OK, since I haven't got the time for further OMF development I've decided to release the plugin's source code. Feel free to improve.

Some release notes:
- You may have to modify some project settings (e.g. paths) in order to compile it on your machine.
- The code is not quite elegant, includes a few hacks for different OMF dialects and lacks comments. Sorry for that.
- Theoretically, it should be possible to compile it under OSX without any modification.
- The package includes two third-party components:
pstdint, a portable stdint.h replacement by Paul Hsieh (BSD license)
SQLite, a small and portable database system (public domain)

A big Thank You goes out to SWS who helped me figuring out some licensing issues.

Grab it here: http://404notfound.bplaced.net/reaper/

[boatclub]

hope to see somebody grab hold of this soon and take it further! OMF support is exactly what i've been looking for

many thanks 404.

louis.

[Runaway]

Originally Posted by boatclub

hope to see somebody grab hold of this soon and take it further! OMF support is exactly what i've been looking for

many thanks 404.

louis.

We've been looking at OMF2 support in AATranslator (off & on) for some time now (in between working on PT, CEP, Vegas & more AA functionality).

As we are about to do another release which incorporates more AA, CEP & Presonus support we plan to start putting some more resources into OMF2.

While this won't be a Reaper plugin hopefully (if we can pull it off) it will be of some use.

[Runaway]

Started looking at OMF2 again - who makes up this stuff?
Getting there slowly - all I can say is hats off to 404 for his perserverance with this stuff!

[sergiovillard]

Ask kindly that do not stop the development of this plugin, because it is very important to the community, back to your time and your way but please continue.

[Runaway]

Thanks mate, encouragement like this always helps :-)

OMF2 support not far off now ;-)

[Runaway]

Originally Posted by Runaway

OMF2 support not far off now ;-)

Famous last words.

For an 'industry standard' fomat there seems to be a lot of 'standards' out there when it comes to omf.

So far our success rate on converting omfs from Sonar, FCP, ProConvert, PT, Cubase/Nuendo and a few others is very high.

Still some challenges with FCP pan envelopes, track volume & track pan.

Does anyone know for sure that Track Volume & Track Pan values are stored in the OMF?

[Runaway]

I guess smarter people than me have been bogged down with this crud called omf but....
Track & Pan values don't appear to be stored in the omf.

Avid (bless their little sox) either have their own 'secret' spec which differs from everyone else (who would have thought) or just want to be different.

We appear to have a big chunk of the major players' omfs working and have now cracked most of Avid's omfs. However I need a few (small) Avid (Media Composer?) created OMFs which contain volume &/or pan envelopes.
I've noticed that even between Avid omfs there seems to be a difference when it comes to clip gain but i need to explore this a bit further.

So if someone would like to donate a couple of small Avid created omfs so i can sort out these envelopes I would appreciate it.

[Bastiaan M]

Hi Runaway.

If you look higher up you'll see a few omf's i created in an avid-system. I think they contain pan and level information.

EDIT: http://forum.cockos.com/showpost.php...6&postcount=14
(looked my post up for you)

If i look at your efforts from a video-view i wouldnt invest too much effort into translating panning and levels into Reaper. Usually you make a temp mix in the Avid so that you can hear what has to go there. The whole exercise of moving that audio into Reaper is to be able to work on the audio in Reaper. The audio-guys at work usually remove any levelling and panning i apply because they want to do it by themselves with the tools at their disposal.

So if you can wing it, great, very great in fact, but if you cannot, your tool is still very very usefull.

[John Lundsten]

Originally Posted by Bastiaan M

Hi Runaway.

If you look higher up you'll see a few omf's i created in an avid-system. I think they contain pan and level information.

EDIT: http://forum.cockos.com/showpost.php...6&postcount=14
(looked my post up for you)

Thanks Bastiaan.
We have been using your "omf2 test.omf" already, MC ver3.1.1, very useful ta.
AFAICT it has no clip gain or Pan/level info. It's got fades tho. Are there any others of yours I have missed (OMF 2 only btw)
John L

[Peevy]

Works great so far for my Cubase sx3 projects.
The process of transferring tracks from Cubase to Reaper was so tedious before. Now I can Have my Cubase tracks transferred to Reaper in a few minutes. GREAT!!

[Runaway]

I guess because I like pain I decided to 'have a go' at the Paris omf cited earlier in this thread. Well I think I have this converting - all the output looks good but will confirm when i can get to a daw to test.

Just downloading the other one (245mb) and will let you know how that one goes.

So it looks like Paris omfs may be supported in the next AATranslator release due in a few days time.

BTW It looks like Paris omfs do not contain frame rates or if they do they are hidden under some obscure entry and not the 'standard' "TCCP" either way we have a work around for it by a user selectable rate.

Just converted that 'crop circles' omf only issue with both is associating each clip with the correct audio - I will try and get that sorted tonight.

[dawhead]

a quick note to say that we are now using 404NotFound's code in Ardour for OMF import. I plan to talk with him about various issues along the way, since I believe we can make it much faster than it currently is. Not that important, though.

[kerryg]

Originally Posted by Runaway

I guess because I like pain I decided to 'have a go' at the Paris omf cited earlier in this thread. Well I think I have this converting - all the output looks good but will confirm when i can get to a daw to test.

Just downloading the other one (245mb) and will let you know how that one goes.

So it looks like Paris omfs may be supported in the next AATranslator release due in a few days time.

BTW It looks like Paris omfs do not contain frame rates or if they do they are hidden under some obscure entry and not the 'standard' "TCCP" either way we have a work around for it by a user selectable rate.

Just converted that 'crop circles' omf only issue with both is associating each clip with the correct audio - I will try and get that sorted tonight.

The only excuse I can give for not acknowledging this very important piece of info was that I was on the road when you posted it and it must have got lost in the email shuffle. This is *extremely* cool news for us, thanks VERY much! We'll look forward with great interest to hearing anything else you have to tell us. Is there any way we can be of some sort of help in this process?

I can't help but feel that ID, who were no slouches as coders, must have figured it was pretty much working when they announced it and all that was needed was final (and possibly trivial) tweaking to nail it - and then events overtook the project and it ended on the 3.0 release instead. The hypothesis I want to test is that rather than PARIS OMFs just accidentally omitting all that info, that it's actually in those OMFs *somewhere* but not properly parsed, and it might just be a matter of locating it and allowing for it.

Would it be useful for me to open and export the same PARIS project as an OMF several times, with trivial changes between them - say just changing track names before re-export - to let us look at what changes occur between those resulting OMFs?

[Runaway]

Originally Posted by dawhead

a quick note to say that we are now using 404NotFound's code in Ardour for OMF import. I plan to talk with him about various issues along the way, since I believe we can make it much faster than it currently is. Not that important, though.

That is excellent news for Ardour users. I have had Ardour on my list of daws for sometime but have been bogged down with other more pressing issues.
404NotFound has done a lot of good work there and as you will/have found there are many ways to do similar things.

Keep up the good work.

Avid knew that one day there would be those that would finally crack the inner workings of the omf thats why the first moved to their own 'secret mens business' version of omf and then finally to AAF :-)

We have made big improvements in our omf reading along with adding more formats and improvements across all daw formats for the next release.

As far as future omf support goes we have enlisted the support of RonN from over in the Audition camp and with his help have finally got some preliminary writing of omfs. Still early days but with luck that functionality will make the following release.

[Runaway]

Originally Posted by kerryg

The hypothesis I want to test is that rather than PARIS OMFs just accidentally omitting all that info, that it's actually in those OMFs *somewhere* but not properly parsed, and it might just be a matter of locating it and allowing for it.

Would it be useful for me to open and export the same PARIS project as an OMF several times, with trivial changes between them - say just changing track names before re-export - to let us look at what changes occur between those resulting OMFs?

What you are saying may well be the case. I would be more than happy take what you can give me and see what we can sort out.

Might I suggest also (if the option allows) both embedded and reference omfs.
If you could also give me a summary of what I should be seeing with any values in samples that would be excellent.

If we can get this organised soon hopefully any Paris specific changes can make our next release.

[Runaway]

Thanks for that - its good to hear from someone who knows something about this stuff :-)

Understood about the TCCP being optional as it appears that almost everything is optional in some daw's version of omf :-)

As you know the framerate is required to workout the session offset (omfvalue / samplerate * framerate) and depending upon if it comes from FCP or others there may be other numbers that come into play eg divide by 1000 etc.

The frame rate is no deal breaker for Paris omfs - what is a problem is how they store the reference to the source media for each clip - it seems to be very different to anything else i have come across - or it could just be another 'ID Ten T error' :-)

As far as omfi:effect:MonoAudioGain, omfi:effect:MonoAudioPan & omfi:effect:StereoAudioGain I am aware of those and (without refering to the code) related to clip values rather Track values.

I would eagerly read your 100 page dissertation :-)

[Runaway]

Originally Posted by Omfoo

In the absence of the TCCP property you can get a frame rate by walking through each object in the file and testing for the presence of a property (any property) using an “omfi:ExactEditRate” data type (rev 1) or an “omfi:Rational” data type (rev 2). The structure is the same for both types:

typedef struct _RATIONAL
{
UINT32 Numerator;
UINT32 Denominator;
} RATIONAL, *PRATIONAL;

Divide the numerator by the denominator. The quotient will either look like an audio sample rate (48000, 44100, etc.) or a video frame rate (30, 29.97, 25, 24, etc.). If it’s a video frame rate then there you have it (there can be only one video frame rate in an OMF).

I guess the theory is fine but no - Paris omfs do not have a frame rate either in the optional TCCP field nor in any “omfi:Rational” data type nor do they have any property which is “omfi:ExactEditRate”.

Not all omfs are created equal - least of all Paris ones ;-)

Well at least it confirms that I'm not going crazy - well at least not as crazy as I thought :-)

[Runaway]

Trust me I take no insult :-)
Consider me a sponge happy to soak up any insight you have to offer.

Unfortunately I have had to work most of this stuff out without any doco and a lot of hack work & trial and error. My limitation is based on the omfs I have encountered. As you can see this is essentially working http://forum.cockos.com/showthread.php?t=51875

If there is a specification and no doubt there is one then almost every daw has a different view of that spec - Avid's omfs being the most different to the rest.

But no I'm not hung up on the fps - I extract what I find and use it as I can.

My hang up with Paris OMFs is that I cannot for the life of me associate the clip with the source media. The SourceID for each clip seems to point to the last listed media file. Now maybe its something simple much like myself or they are doing something a little different to the rest but any suggestions would be gratefully received.

I've wasted too much time already on Paris omfs and while they are a very small section of the omf community they are active on this forum and I would like to help them if I could.

Thanks in advance

[Runaway]

Ahh yes I remember picking that up and the Hebrew-Portuguese dictionary and wondering which would be the more interesting to read.

I think I picked the wrong one ;-)

Thanks

[Runaway]

Thanks for that but I have that Kick drum and these

Source Table Entries: 5
1 "H:\VB\AATranslate\JL\KICK.wav"
2 "H:\VB\AATranslate\JL\TEST-001.wav"
3 "H:\VB\AATranslate\JL\TEST-002.wav"
4 "H:\VB\AATranslate\JL\TEST-003.wav"
5 "H:\VB\AATranslate\JL\TEST-004.wav"

The last 4 being 6mb each of silencech.

I also get 16 tracks of kick drum (it seems).

So by good luck (or not) this file translates fine.

For Paris OMFs its not isolating the audio itself its matching the SEQUs & SCLPs etc up to the correct media.

If I recall correctly I would match up the MobID with the SourceID and all is well but not in other Paris OMFs where multiple audio is used.

I'm happy to be smacked over the snout for stupidity but what works for me with other omfs don't work for me with these ones.

[Runaway]

Got it!

Paris uses the SourceTrackID - now I have to have somebody killed :-)

Just when you thought it was safe to open another omf......

[Runaway]

Ok I can now load Paris omfs into Reaper (and obviously all the other daws Audition, Samplitude, Vegas, Cubendo etc).

Clips are associated with the right media on the right tracks at the right offset, indent & size.

I haven't worked out what crazy scheme they are using (and I'm sure its a doozy) for clip volume or gain & pan envelopes yet but I'm busting to try and get another release of AAT out, well, it was going to be tonight, but now hopefully within the next 24 hours.

So this is as much Paris support as I can muster right now but if there is enough demand I will revisit.

And I used to think Paris was a small town in the back of New South Wales and a few months ago I couldn't even spell OMF!

Thanks all for your patience, input & support.

[kerryg]

Originally Posted by Runaway

Ok I can now load Paris omfs into Reaper (and obviously all the other daws Audition, Samplitude, Vegas, Cubendo etc).

Clips are associated with the right media on the right tracks at the right offset, indent & size.

I haven't worked out what crazy scheme they are using (and I'm sure its a doozy) for clip volume or gain & pan envelopes yet but I'm busting to try and get another release of AAT out, well, it was going to be tonight, but now hopefully within the next 24 hours.

So this is as much Paris support as I can muster right now but if there is enough demand I will revisit.

And I used to think Paris was a small town in the back of New South Wales and a few months ago I couldn't even spell OMF!

Thanks all for your patience, input & support.

Runaway, you and OMFoo freakin' ROCK! This is going to be huge news on the PARIS forums - we'll test an expanded range of OMFs to be sure, but I'll be launching an AATranslator license drive on the Forums as a way of saying "thanks". Major, major kudos!

[kerryg]

Originally Posted by Omfoo

Hi. I was just googling around and discovered this thread.

Fyi the TCCP property is optional. That makes sense to me because there is no reason to demand that an audio-only project make any commitment to any timecode format if the project doesn’t need or use timecode. After all, why force the project to decide?

In the absence of the TCCP property you can get a frame rate by walking through each object in the file and testing for the presence of a property (any property) using an “omfi:ExactEditRate” data type (rev 1) or an “omfi:Rational” data type (rev 2). The structure is the same for both types:

typedef struct _RATIONAL
{
UINT32 Numerator;
UINT32 Denominator;
} RATIONAL, *PRATIONAL;

Divide the numerator by the denominator. The quotient will either look like an audio sample rate (48000, 44100, etc.) or a video frame rate (30, 29.97, 25, 24, etc.). If it’s a video frame rate then there you have it (there can be only one video frame rate in an OMF).

Don’t forget to byteswap the UINT32s when necessary, and beware of Denominators set to zero. If the Denominator is zero then don’t divide anything; just use the Numerator.

Yes – in rev2. There is an Effect Definition Class ‘EDEF’ and an Effect Invocation Class ‘EFFE’. The EDEF will have an ‘EffectId’ property with a data type called ‘omfi:UniqueName’.

The unique names you are looking for are:

omfi:effect:MonoAudioGain
omfi:effect:MonoAudioPan
omfi:effect:StereoAudioGain

Afaik there is no stereo pan.

I could write a 100-page dissertation on this stuff but I’m stopping here.

Good luck.

Hope you do consider writing that 100-page dissertation some day, I would read it with *great* interest - and given the difficulty people have historically experienced with OMF formats, it would be an enormous public service (especially for those of us that are "married" to older systems where OMF is the only possibility for session export). A heartfelt thanks on behalf of the PARIS community for the insights you just gave - and a belated Happy Valentine's Day, that was a heck of a bouquet!

[kerryg]

Originally Posted by Runaway

Ok I can now load Paris omfs into Reaper (and obviously all the other daws Audition, Samplitude, Vegas, Cubendo etc).

Clips are associated with the right media on the right tracks at the right offset, indent & size.

I haven't worked out what crazy scheme they are using (and I'm sure its a doozy) for clip volume or gain & pan envelopes yet but I'm busting to try and get another release of AAT out...

I'll save you the effort on one of those - PARIS doesn't use clip volume, it uses track-based automation only - volume, pan and mute for each track.

[Runaway]

Originally Posted by kerryg

I'll save you the effort on one of those - PARIS doesn't use clip volume, it uses track-based automation only - volume, pan and mute for each track.

Cool - Any chance of getting hold of some of the values I'm expecting to find eg sample values of some of the envelope points and maybe a few pan & gain values at those points?

I'm trying to see if I can make some headway with this before this next release.