On Sun, September 2, 2012 4:43 am, Thomas Vecchione wrote: > On Sat, Sep 1, 2012 at 4:47 AM, Robin Gareus <robin@xxxxxxxxxx> wrote: > >> On 2012-09-01 07:54, Ralf Mardorf wrote: >> >>> On Sat, 2012-09-01 at 15:50 +1000, Patrick Shirkey wrote: >>> >>>> The movie industry >>>> >>> >>> Linux isn't use by the movie industry for everything >>> >> >> You are soooo wrong! >> >> They all love Linux. "They" in this case are Hollywood based WB, >> Technicolor and LA Center Studios. Probably others too. Their >> Datacenters >> runs almost exclusively on it. While it's true that actual editing work >> happens predominantly on Win/AVID, there is some customized HW/SW around >> (e.g. for color-correction) that is Linux-based and the amount of it is >> rising quickly due to fact that it can be customized, adapted and >> maintained much more easily than Win or OSX. None of it is >> free-software, >> though. >> >> best, >> robin >> >> > > I want to point out and repeat what Robin actually said here, as this is > on > two lists and it is hard for me to follow what post is on what list as I > read them all as one thread... > > The Movie Industry DOES use Linux, HOWEVER it is primarily on render > farms, > and backend support. The majority of the editing, mixing, etc. does not > happen on Linux yet, sorry. > > A short time ago I went looking for, and demo'd the available options for, > a decent video editor on Linux that could do a simple task. Allow me to > easily, and quickly, green screen out a source clip, and overlay it on top > of a background and layer in additional generated content (Titles, images, > etc.) in between them and on top of them(While obviously maintaining a > high > level of quality). The end result? I purchased Final Cut Pro X (Already > had Final Cut Studio) and do the work on my mac laptop. I hope to see > something like Lightworks change this, but honestly I wasn't able to work > nearly as quickly in the tools that could do this basic workflow(And I am > on a timeline where I would have an afternoon or less to get the work done > before I have to do other projects, and that includes editing the audio), > and many tools I expected to be able to do this (The Blender NLE for > instance) couldn't in a way that was close to acceptable for the timeline > I > needed to work in(And I specifically went asking several people that are > fairly knowledgeable in using the NLE and all aspects of Blender as well > when I found I couldn't do it myself). > > Even with this, I have now hired on a couple of people to pick up the > video > work so that I can concentrate on audio. Sadly right now there are severe > limitations on this even. My primary workstation is Linux for the record, > I accept audio stems and work on them and render out new audio for them to > remux to the video. If they haven't got something cut I can use, I either > have to manually pull in the original audio, find the section they used, > and recut it, and manually pull it into sync, or I have to get them to > redit and rexport the audio. Why? Because I am dealing with people on at > least two different platforms and currently there isn't a good solution > for > AAF or OMF that works with ArdourXChange out of FInal Cut Pro (Either > Studio, which I tested with John without Automatic Duck that no longer > exists, or X as Automatic Duck went no longer exists and I can't even > their > previously released plugin for X anymore, and am now hearing the > laternatives dont' work well either and there is no option to write to it > from FCPX). > > Or I can go into Windows and possibly use AATranslator, which by all > accounts I have heard is great software and would do what I want, but I > don't have a Windows machine and we are discussing native solutions on > Linux. > > So what I have right now is a substandard solution, that thankfully I am > typically either just doing basic edits on long pieces or working on short > (approx 5 min) clips so that I can deal with it. But doing a full length > feature having to work with dozens if not hundreds of others on such a > workflow? Talk about painful. This doesn't include mixing in a so called > 'standard' destination format like 5.1 etc. that is also still fairly > painful on Linux, even with the revamped VBAP panner in A3. > > Now don't get me wrong, this doesn't mean I don't enjoy working in Linux. > Most of my stuff is primarily stereo destined, I thoroughly enjoy working > in Ardour and Mixbus in Linux with Jadeo for doing sound for video, or > editing down music, etc. But it works for me because I know how to work > in > it, and don't need to relearn the 'quirks' of the workflow etc. I can > work > faster in Linux than I can on OS X using the same tools. But many tools > just don't exist that would enable a true workflow as is needed for it to > become popular for a majority of uses that many people are using the other > OSes for. I didn't submit bug reports to each of those pieces of software > I tried above.. I probably should have, but was under a very tight > timeline, especially then as I was doing at least two people's jobs. And > this is true of most 'professionals' in the industry. If it can't be > fixed > quick then it is broken and you have to move on to get other work done and > bills payed. > > In particular Blender with 3D is doing amazing work, but it is only one > piece of the puzzle. I am looking forward to reading up on their success > with their new open movie, where they started incorporating blender into a > live workflow, and seeing how they did, I hope it addresses some of the > reasons I couldn't use it above, honestly. I enjoy using Blender for the > little 3D work I do, and would have enjoyed using it for the video work as > well quite likely. > > Remember, most of the 'professional' users of software in the industry > aren't the really big houses(Unless talking about very specialized > software). They are the smaller commercial studios, project studios, > independant companies, etc that outnumber the larger studios tremendously > and typically do a lot of the commercial work(Commercials, but also > industrials like training videos for companies, corporations, release > videos, etc.) > This problem comes down to the strangle hold that a few companies have on the audio tools market and the impact that has on the Linux user base and development environment. There is categorically no reason why this problem could not have been solved by now if the people using the software in professional context were actively feeding that information back into the development pipeline. Clearly the problem has been solved in programs like smoke and maya which have been running on Unix platforms for longer than they have been on Linux. There is a large group of companies that build solutions that run on Linux that we never hear about. For example did you know that there is a company that specialises in porting AMD's instruction set to a paraellised architecture for a closed source version of ffmpeg? They are actually paid by AMD to do this and provide services almost exclusively to Government, Military and HPC researchers. The main problem here is not that the software or skills are not available but that there is not a big enough pool of professionals participating in the development cycle to drive the open platform forward. To solve this problem we need more publicity singing the praises of the platform and demonstrating the real world value that professionals gain from providing feedback on the tools that we know they are using. It requires a very conscious decision on the part of the user to participate in the feedback cycle and to encourage it there needs to be some kind of incentive which makes it not only fun but a critical business tactic. We are working on that problem here but it is still a murky area at the moment. Rest assured that several heads are involved :-) -- Patrick Shirkey Boost Hardware Ltd _______________________________________________ Linux-audio-user mailing list Linux-audio-user@xxxxxxxxxxxxxxxxxxxx http://lists.linuxaudio.org/listinfo/linux-audio-user
