5 Replies Latest reply: May 8, 2010 3:44 PM by Anig Browl RSS

    Multitrack session import/export

    Anig Browl

      So I've been using audition since it was CoolEdit back in the 90s, and nowadays I work in film. I was excited about the release of CS5 and downloaded a trial as soon as possible, but I'm quite disappointed. why? It's still impossible to move multitrack sessions around.


      Quick recap of multitrack audio capabilities in Adobe products...


      • Audition loads/saves .ses and XML files, with the latter recommended for future compatibility. It doesn't import other XML files.
      • Soundbooth only saves stereo files, but can load/export the ASND format (new in CS3/4?) to embed multitrack info (which can then be reopened and edited in SB).
      • Premiere exports OMF and its successor AAF (popular multitrack project exchange formats) and imports the ASND format. It imports and export final Cut XML files, but not Audition XML files.
      • Premiere also lets you select any clip and open it for editing in Soundbooth or Audition. That's clip, singular. Select two audio clips and the option is greyed out.
      • Neither Bridge nor Dynamic link allow multitrack audio sharing.


      Do you see the problem here? If you have a pile of multitrack audio clips in Premiere (like you would after editing a film) there is no way to get your audio timeline information, with many tracks and many clips, from Premiere into Audition or its little brother, Soundbooth

       

      This is, frankly, a disaster.

       

      Think about it: audition has by far the best forensic, analysis and editing tools of any DAW. But anyone using Premiere who needs to do fine-grained audio editing - because although Premiere is audio-friendly, it has its limits - is forced to export to something like OMF or AAF and then  bring into Pro Tools (spit). You can create audio in Premiere and export it as audio tracks (with blank space) to Soundbooth, from where you can throw it back and forth between SB and Premiere to your heart's content.

       

      But if you're editing a film (not a music video) then there is no way you would want to put the audio together in Audition or Soundbooth first, because 90% of the time it will be following the picture. You can get individual clips out of Premiere and into the others...but only one at a time, which defeats the entire purpose of post sound editing, where you are trying to blend together multiple clips from different takes. And even though Premiere offers multiple options for exporting all the clip information and even volume/pan and plugin data to standard platform independent multitrack formats, Adobe's audio tools can't read any of those files.

       

      When is Adobe going to fix this? I'm sorry to say that I'm becoming embarrassed to tell people I prefer Audition, because everyone in the business knows it's hopeless for multitrack portability.

       

      And I'm not saying this to be selfish - this problem is really, really hurting Adobe's sales of both Audition and Premiere. Avid owns Digidesign, maker of Pro Tools. I hate both, but from a logistic standpoint they're a great combination, it's pretty easy to get stuff out of Avid into PT, and of course both of them have 'tradition' on their side. Apple has Final Cut, and Soundtrack Pro (rather like Soundbooth) or Logic, for those who want more. I don't care for this platform either, but it's a viable solution. And here's Adobe, with a decent lightweight audio editor in Soundbooth and the best analytical/forensic audio tool - which is exactly what's needed for film post - in Audition.

       

      And yet there is no way to get your audio timeline out of Premiere and into either of these tools, nor do either of them load any industry-standard multitrack file formats, nor are they even compatible with each other. I was looking at an Adobe TV episode earlier today where they take a music project from Audition to Soundbooth...by exporting the tracks as individual audio files, which are then added to a brand new Soundbooth project. This is just ridiculous. I mean, if I'm consulting for someone about audio post for a film project or to outfit a studio, how can I possibly recommend Adobe's products? Individually they're great, but in workflow terms, forget it. Anyone who needs to get things done will probably go for Pro Tools, and since Avid and FCP are both more widely used than Premiere they'll probably go for one of those too.

       

      I really can't understand why the audio tools aren't integrated to the same degree as, say, After Effects. Premiere + AE is a winning combo, and yet audio is arguably even more important than color correction or VFX, as almost every film has a soundtrack and audiences are far more intolerant of poor sound than they are of visual shortcomings. Please, please fix this. I live in San Francisco, so I'm not far from Adobe's offices: I will happily donate a day or two of my time, and even supply a waiver of any claim in intellectual property, to answer any and all questions about audio workflow. I think I've learned a few things worth sharing over the course of 10 features and about 50 shorts. It's not a big deal for Adobe to fix.

       

      Message was edited by: Anig Browl to add the following: there is a LOT of stuff I do like about CS5 - Premiere has not gained a ton of new features, but it has had a massive performance overhaul, and the new features that do exist are very useful. The same is true of several other packages - I wouldn't want this to be seen as a 'nothing's ever any good' rant. The sad thing is that it won't reach its full audience because of the inexplicable barrier that remains in place between the NLE and the multitrack audio environment.

        • 1. Re: Multitrack session import/export
          SteveG(AudioMasters) Community Member

          Nice rant, but not everything's quite as bad as you think...

           

          You should check out AATranslator, which does begin to address some of your import/export issues, and is still very much an active project. It's just possible that Suite Spot might even investigate multitrack export possibilities from Premiere if you ask him nicely...

           

          Anig Browl wrote:

          I live in San Francisco, so I'm not far from Adobe's offices:

           

          Ah, but the Audition development team live in Seattle. Well, I suppose that in intergalactic terms it's not far! As for improved integration though, I think you'll just have to wait and see what, if anything, happens; the development team won't comment publically about any work they have in progress until a formal announcement is made in the form of a press release. It was ever thus, and will remain this way for the forseeable future.

          • 2. Re: Multitrack session import/export
            Stephen P Simpson Community Member

            Hello Anig.

             

            I upgraded to the full version of AATranslator recently because it allows an OMF exported from Premiere CS4 to drop all the multitrack audio information onto the Audition timeline. Oh what a godsend - no more boring, utterly soul-destroying, uncreative tens of hours of manual conforming!

             

            I've informed an editing colleague that, when a project is delivered on a portable hard drive, my preferred OMF export settings from Premiere CS4 are:
            Bits per Sample - 24;
            Files - Separate Audio (as opposed to 'Encapsulate');
            Render - Copy Complete Audio Files (as opposed to 'Trim Audio Files');
            Handle Frames - Not Applicable.

             

            Something to watch out for, which I had never seen before, are the 'disembodied volume envelopes'. On the Audition multitrack timeline, we're used to seeing clip volume envelopes and their volume nodes contained within the graphic box that represents the audio clip. When importing an OMF exported from Premiere CS4 via AATranslator, you might find that some clips display their volume envelopes way above the audio clip graphic box, in the above track, in fact. The volume envelope will follow the clip as you move it around the timeline, but will display in the track above. You'll find that the envelope and its nodes are 'ungrabbable' and therefore unmodifiable. Making a duplicate of the clip will result in another clip with the volume envelope also being out of bounds. Grabbing the clip from the Files Window and placing it on the timeline, however, will result in nearer to what one would expect, that is, the volume envelope will be within the boundaries of the audio clip graphical representation, although it'll be at maximum value and possibly will have lost volume node information.

             

            The reason for this peculiarity would appear to be quite straightforward: Premiere CS4 allows the volume envelope on an audio clip to have a maximum value of +6.02dB; AATranslator 3.0.00 faithfully translates this value; Audition 3.0.1 can only cope with correctly displaying a maximum volume envelope value of 0dB.

             

            I wrote to AATranslator's developer with this observation. He very quickly and cordially informed me that this behaviour had been noticed, and that Audition could sort out the issue with ease:

             

            Right mouse click on the clip -> Clip Envelopes -> Rescale Clip Volume Envelopes (6dB, or however much is necessary to access the envelope and all its nodes).

             

            Whilst rescaling the envelope down, this function also scales the clip's volume property up, resulting in no net volume gain change, which is what is required.

             

            When I've done what I need to on an audio timeline, I tend to supply my editing colleague with a finished stereo wav file. I've just had a quick look at AATranslator - I could be wrong, but I can't see a way of sending back to Premiere CS4 complex multitrack audio information.

             

            I have another editing colleague, but he uses Edius. Something I have yet to formally request from AATranslator's developer is that he add support for the Grass Valley Edius AAF export format. If that is ever implemented, then I'll be a very happy bunny...

             

            Best Regards,

             

            Steve.

            • 3. Re: Multitrack session import/export
              SteveG(AudioMasters) Community Member

              Stephen P Simpson wrote:

               

               

              I upgraded to the full version of AATranslator recently because it allows an OMF exported from Premiere CS4 to drop all the multitrack audio information onto the Audition timeline. Oh what a godsend - no more boring, utterly soul-destroying, uncreative tens of hours of manual conforming!

               

              Well there you go Anig - seems he's already done it!

               

              The reason that I wasn't directly aware of this is simply that I don't do video work any more - gave it all up ages ago.

              • 4. Re: Multitrack session import/export
                SuiteSpot Community Member

                I just got in from a tiring FOH mixing gig so I probably only read every 3rd line of all those posts but thanks SteveG & Stephen P

                Anig I feel your pain & frustration (and there is even more pain & frustration trying to write code to deal with all these formats).

                 

                Like Stephen P I was blown away with that little gem that AA deals with rescaling envelopes - brilliant!

                 

                I have quite a few users who are as happy as Stephen P but I also have some who are less happy mainly because we have yet to cater 100% for some (or even all) formats but we are enhancing AAT as fast as possible.

                 

                Anig as far as your workflow AAtranslator will read a Premiere OMF and AAT also creates a Final Cut Pro xml which Premiere will read.

                Is it perfect?  Probably not but it might just do the job for you.  So you can either download the demo (which has pretty strict track/clip limitations) or contact me via the AAT website and we can organise to convert one of your sessions for you and that way you can see if you need to visit Adobe or not

                 

                BTW We are currently working on a new release which has (among other things) a very much faster and improved omf conversion (well it will when its finished)

                • 5. Re: Multitrack session import/export
                  Anig Browl Community Member

                  I'm glad of these reports, and have been checking out AAT, which has potential. I do have some reservations about it, but this is not the place to critique helpful solutions; I'll limit myself to saying I'm a bit troubled by the obvious dependence on an old UI library and the conversion speeds, which seem slow for a task that is basically text-based (ie metadata manipulation). Look out for a follow up email, Michael, and thanks for your previous reply.


                  All the same, at bottom this is an Adobe problem - third party conversion utilities shouldn't be needed to move projects between Adobe's own software. It's a shortcoming of the CS4 & 5 suites, in that file-conversion approaches like that of AAT are necessarily destructive and thus preclude dynamic file exchange. I mean, would you buy a new car if the salesman said the fenders fall off above walking speed, but that you could fix the problem with $100 of duct tape?


                  Suppose, for example, one is polishing audio for a film, but audience tests or the director's opinion mandates a scene be cut, or (worse) lengthened. Making a change in the NLE would require export from Audition to a Premiere-compatible format (probably via AAT to AAF or FCP-XML, but see below), import to Premiere, consistency and compatibility checks, adjustment of the edit timeline as required, and subsequent re-export for conversion by AAT, import to Audition, followed by further compatibility and consistency checking - plus a separate render and media export of the video, since Audition can only handle a single video track.


                  While both hardware and sanity dictate that it's probably best to manage the project as a batch of single scenes or reels rather than moving an entire feature or TV episode at once, this process is likely to take at least an hour before audio editing can resume. In theory it should only require a few minutes for export, conversion, import, adjustment, re-export, conversion and re-import, but that process offers 6 stages for things to go wrong. Even if everything is usually fine, time-consuming sanity checks are still required, unless you like all-night emergency rescue sessions. In practice, a single small timeline change in the NLE may well be easier to accommodate by manually moving the audio in Audtion, and cross-checking the timecode in both applications.


                  When I started doing audio for film - back when it was film, and the audio was recorded on DAT or even analog tape - this was the norm and those skills still work. But it undermines the whole point of using metadata and products like Adobe Bridge for version management. This is not the fault of AAT in any way; it's just a fact that when your workflow involves export/import via third-party tools, you have to treat every such iteration as a separate project file and adhere to strict versioning and administration/backup practices even though these would be better performed by the computer. In a perfect world, we'd stick to a scheme whereby the editor locks the movie once and forever, the audio engineer and colorist work their magic on the soundtrack and the picture respectively, and then their efforts are united for the final master. But we don't live in such a world, and clients don't appreciate being told that changes which were easy and swift at an early stage are now either impossible or hugely time-consuming.


                  This is what I mean when I refer to being embarrassed to say Audition is my favorite editing tool. Most people want Pro Tools, because although it's expensive it's standard and the workflow is predictable. So that's what I end up using on a great many jobs; it's capable for most tasks, I'm good at it, and to be frank I can get paid a bit more for my ability to use it. The sad thing is that I can work better and faster in Audition, where things like the spectral editing, noise fingerprinting, and advanced filtering offer really amazing possibilities...but getting projects in and out of it are a terrible struggle. Soundbooth makes things marginally easier, but only marginally. I can see a possibility of rebuilding an audio timeline from Premiere using Adobe's Javascript Extensions, and then generating an audition-compatible XML file from there...but I can't say it looks like fun


                  As for exporting to AAF, Premiere CS5 has other problems. Many things can not be exported, like per-clip gain levels (because the Advanced Authoring Format doesn't properly support them, according to Premiere, although the protocol specification appears like it should). In any case, Premiere CS5 has trouble importing AAF; after exporting a short scene using all possible settings (normal, legacy, or normal with audio embedded), attempting to reload the exported file into a blank project causes an instant crash due to a kernel error (the import filter appears to launch a 32-bit process).


                  In any case, although AAF is still being developed it seems to be moving towards a BWF-style XML format, and most of AMWA's development seems concentrated in that direction. Final Cut XML appears to be the best solution, but this is (obviously) rather sad, and furthermore Premiere does not strip invalid data (eg >6.02dB gain, which Premiere clamps at +6 on import) from FCP-originated projects, but re-exports it - not an altogether bad thing for working with FCP partners, but not exactly optimal either.