How do you know that it is rendering audio files as opposed to video files?
Try a test and turn of all audio tracks (test export or render of a section (shortWork area bar).
Try a test an Delete Render Files. (ie start a fresh render)
Try exporting Audio Track only and re laying that into the Sequence.
Exports do nor rely on Render filesu nless you want them to. (ie a fresh render occurs when exporting)
Well if you took a look at the little screen grab I sent you would see it says rendering audio previews. If I delete the audio from the timeline I don't get the rendering audio previews message and it goes straight to exporting my video. I have tested this a bit and it's definitely the audio that is the bottleneck.
Just an observation. The 44.1KHz Audio should not be an issue, since PrPro Conforms that to 48KHz 32-bit floating point in the CFA files, upon Import. I think that you can rule the 44.1KHz material as the cause of the bottleneck.
However, I do not have any thoughts on what the issue could be, unfortunately.
I have never seen a separate Audio Render Window (and progress bar).
Its also interesting that you are saying Export ..yet I am seeing the Timeline (Seq) in your sreen grab. Not sure what to make of that because I would expect to see AME???
Any how...there is a Preference Setting that says something like "Render Audio with Preview".
Try it OFF or vice versa
It's all the same. I can render audio from the render sequence drop down and I get this window, or if I export I will get this render window before before actual exporting begins. And as I mentioned if I close the project it has to do it all over again, it will keep the renders as long as the project stays open. I'm having no issues with video renders not coming back. All render folders set to same as project which is on my RAID. IMO PPro is being asked to do WAY more than it actually needs to in regards to audio. Why in the world would it need to render audio previews from uncompressed files? It should just transfer what is needed from point A to point B. It should be way faster than real-time but is in fact slower than rendering 1080 video.
I have a similar workstation. My sequence is a little more complex, no multicam, but its says it will take 6 hours to render my 27 minute film. No effects are applied as I've not yet mixed. Rendering audio for export should not take this long.
Plus it fails when it reaches 99.6%. Infuriating.
Let me elaborate.
CC (fully updated)
Mercury wGPU Active
Timeline: 1920 x 1080 x 23.976
Master Track: Mutlichannel (2 ch)
No Submixes (yet)
(60) Standard Tracks containing Mono and Stereo
(3) 5.1 tracks of Surround FX
Removed all 5.1 tracks. Audio render for my 27 minute film now takes just a few minutes.
Well that's a clue. I don't have any 5.1 in my piece. In any case that still doesn't make any sense. Does Adobe monitor these threads?
Yes. I know they do. I get responses from "staff" people sometimes.
Okay, so removing 5.1 worked for a couple hours, now it's back to its same old tricks.
I've been getting crashes too while rendering. It finishes the audio but crashes halfway through the video. If I restart the mac it's fine but I'm having to restart a lot! I'm thinking it's something to do with the NVidia drivers because I start to get little drawing errors on the screen. All drivers and Cuda is current.
Sorry about your experience. Have you tried to flatten your multicamera sequence before rendering and exporting?
Yes. I know they do. I get responses from "staff" people sometimes.
Right, we have Adobe staff that do monitor the forums, myself included. However, it is a user-to-user forum primarily. Thanks for providing feedback to our staff, we like to interact with our users to improve Premiere Pro as much as possible.
Would flattening affect audio?
I do occasionally see noisy GPU screen printing for a split second when scrubbing non rendered portions of the timeline while waiting for the drive to respond to file reads. But they are very fleeting. Not a concern for me. I would describe them as flashes of digital "snow" in my viewer during scrub, never during playback or in rendered content.
I am having the same problem with a fairly high end rig... i7-3930K, 32Gb RAM, GTX 780 with MFE on, OCZ Revo system drive, SSD scratch drive, 3 7200RPM project disk and 1 7200 disk where I write my rendered files to. With my old projects and multiple videos and separate audio recorded at 96K never have a problem after importing from CS6 to CC. With a new project that had 3 video sources and one audio source I used PluralEyes 3 to sync the files. Worked great, but when I went to render the final video which is 47 minutes long, the video took maybe 10 minutes and the audio and hour…
I have had this problem too ever since upgrading to CC from CS6. I always use nested sequences for syncing up my system sound, so I have not tested if this bug or 'feature' is premiere feeling the need to make audio previews for nested sequences vs other ways of syncing sound with video clips. I did an experiment. One sequence had 192 sequences, but only 33 within the work area for rendering. When I went to render, premiere wanted to render 192 audio previews. When I copied just the work area to its own new sequence, premiere wanted to render 33 audio previews. So annoying that it wants to render a 'preview' for every sequence on the timeline, not just only what is in the work area. Why is that, Adobe?! Do you understand how much editing time this wastes when I can't be using my machine? I think from now on, I will be sure to keep a fresh sequence only for rendering out from.
And I too can confirm CC renders audio slower than video. Just got my new core i7. The video encodes soo much faster than my old rig, but the audio seems to be stuck in molasses. Would really love to hear an answer from the high software paid engineers at Adobe who get health insurance why they have taken audio rendering back 20 years. What do we have to do to not get this annoying glitch? Not use nested sequences? That makes keeping sync soo much harder if that's the case.
Message was edited by: Kevin Monahan
Abusive language. Please do not use profanity.
That did the trick! And make sure you do the same for all multi-cam clips - not just the ones in the work area. It tries to render audio from the entire sequence. Weird bug.
What did the trick? This thread is long, can you be more specific please? was it flattening the layers?
Yeah what did the trick?
What did the trick? I'm having the same issues.
I believe it was the post by Kevin Monahan regarding "Have you tried to flatten your multicamera sequence before rendering and exporting?" I haven't tried this yet and feel like this is a work around when the problem never manifested itself with CS6...
Thanks jt20194. I tried that, and although it reduced the full number of audio clips that needed to be referenced (not sure why that makes a difference), it was still over 100 audio clips that "needed" to be rendered. I had chosen a 10 second clip with In and Out points, and the MAXIMUM possible audio clips that are in any way connected to the video clip is 4.
It's a bug, or a bad design choice. Either way, it needs to be changed.
Just submitted a bug report. If anyone finds any real solution to this, please let us all know.
Any headway in this problem? I'm just running into it now, rendering out 10 second .wav audio files takes 20 mins the first time, only a second the 2nd or third time though. What gives?
For me, when I use Sequence->Render Audio in Premiere Pro CC, there is a huge difference in rendering time depending on whether the audio tracks are locked or not. This is obviously a bug, because clicking the lock icon on/off shouldn't affect anything about the audio itself.
Which is it? Do you get faster renders with locks activated or locks deactivated?
Just to give a few more details, the project is 100 minute film with 22 audio tracks, and stereo master output. The audio is fairly complex, but no nesting. With all tracks unlocked it takes less than 5 minutes. With all the tracks locked it went for almost 2 hours before I cancelled it; I've never had the patience to see it all the way through.
After rendering all audio with the tracks unlocked, if I then switch even one track to locked, it will re-render, even if that newly locked track has no audio on it. Again, no sound is being moved or modified in anything I am doing.
We have exactly the same issue, the audio render files never save and when we want to export the sequence it takes 20 minutes to 'render required audio' before the export starts. Its a fairly simple sequence audio-wise, 6 audio tracks, with some reverse speeding on some audio tracks, video footage is RED, no difference if I nest video it still needs to render audio on export. I've not been using Premiere Pro for long, migrated from FCP7 so no great depth of knowledge of it as yet. I thought to 'nest' the audio but seems that isn't an option. Super annoying issue, only have it in this 1 project, all other projects we have no issues like this - surely it can't be the reverse-speed effects?
I looked into this a little closer. Are those of you with this problem using multicamera editing? If so, when you set up the Multicamera Source Sequence, is Audio set to "Switch Audio" in Sequence Settings? See image below.
When Switch Audio is used, all source audio tracks are available in the nested sequence and since multicam is enabled for the audio portion of the nested sequence, audio belonging to each video track needs to be rendered separately. So if you have 4 cameras, it will take 4 times as long to render than if you had multicam disabled on the audio portion of the nested sequence.
I think that is why when I suggested flattening audio, the rendering time was as expected.
For locking tracks causing longer audio rendering times, or if you would like to see the "Switch Audio" feature designed differently, please file bug reports: http://adobe.ly/ReportBug
Sorry for any trouble this has caused you.
My instance, noted above, does not involve multi-cam or any nesting of any kind. I have filed a bug report. Thanks.
Let me start by saying I am only a hobbyist and do this more for my son and niece since they are musicians and play live in a number of venues. I am not an electrical or audio engineer (I am an aeronautical engineer) so what I am about to say is based only on what seems logical to me… First rendering audio should simply be faster than rendering video given the amount of digital data that needs to be processed. That stated, I know that the Mercury engine using CUDA to process video vastly speeds the rendering of video. Regardless, using today’s CPUs (i7 3930K in my case) should and does more than make up for the additional processing power for rendering video using CUDA.
It seems to me that there is a bigger problem than whether or not we are flattening, nesting or locking our tracks. It shouldn’t take as long as it is taking for some of us to render audio!!!
I only experienced this problem one time. I am not even sure what I did to change the outcome on the same project that I have since edited and rendered many times. But now the audio renders very quickly (less than a minute). I will say that in almost all cases my music videos have different audio sources (44.1KHz, 48KHz 16bit, 48KHz 24bit, and 96KHz 24bit) and that I set all my sequences to 96KHz since my master audio sources are recorded at that. I use an external firewire RME Fireface 800 using RME ASIO divers to handle all audio. I do find that this setup can be stubborn in that there are times PPro CC doesn’t know what to do regarding my audio so simply won’t allow me to play any track. When this happens I simply remind PPro CC of my settings by going to Edit/Preferences/Audio Hardware then check that the Adobe Desktop Audio is set to ASIO FireFace and click on ASIO Settings to check that the buffer size and sample rate are correct. If they are not I read the flash from the RME and all is good.
But I diverge, there is a problem Adobe and it needs to be fixed. This never happened to me using CS6…
Hi Kevin. You may want to re-read some of this thread. Numerous people here are saying that flattening the audio didn't do it and some are not even using multi-cam. And as I mentioned in one of my posts, I per-rendered ALL of the audio into a stereo track in all of my multi-cam sequences in the correct sample rate to make it as easy as possible on Premiere but still had the same issue. So can I ask again? Why would Premiere need to "render" an entire 60-90 minute multi-cam timeline when I'm only using 30 seconds of material from it?
And the OMF export on this project was epic failure. It took 90 mins to attempt it but then crashed. I XML'ed it to FCP and it did it in 90 seconds and had no issue with 2GB limitation.
I would hope alarm bells would be going off in the hallways at Adobe over this but from your posts it doesn't sound like it. The thing is this stuff is bread and butter to all editing these days and it's something that people do day in and day out and it needs to be bullet proof. I can't imagine the reaction of having a paying client in the room under deadline and moving to the simple task of OMF export and having to wait 90 mins and then have it crash. Would you want to be in that room trying to explain that?
I think Adobe needs to realize that Premiere is still getting it's but kicked by a program that's been dead for years now and and hasn't been updated since. And also I think Adobe needs to realize that pretty much nobody expects Premiere audio to be "better" than Avid/FCP but they do expect it to be as good and as fast. So when I hear all audio gets converted to 32 bit or whatever I think people need to realize that if this impacts performance then it needs to be an option or just kill it because nobody cares.
I really hope Adobe doesn't just brush this off because I for one have been in a lot of pretty high end facilities and the conversation comes up of Premiere and the thing I hear over and over is it's just not there yet.
The Reason for me was:
- Mixdown all the audio tracks to one WAVE or MP3 file
- Re-Import it to the multicam sequence
- Delete all the audio tracks of the sequences, that are used in the multicam sequence (muting is not enough...)
- Export the multicam sequence.. and voi-là.. my sequence was rendered in 3 minutes instead of 2h30...
And at last, report this bug to adobe.. It can not be, that "professional editing software" causes such issues.. It's a shame, that adobe knows this issue since over half a year, but does not fix it until this day..
Hope this soultion works for you too guys...
I am using multi-cam, only 4 cameras.The audio is not set to switch. In fact the camera audio is muted and I am using a seperate stereo mixed track of the band.
Every time I export a single song (from a set work area) Premiere renders 'required audio file'. It even says it is rendering six Audio Previews. I am sure it is re-rendeing the whle of my imported stereo audio as the single track has been cut into segments for each song.
I can't charge the client for this rendering as it is a technical issue. A song that should take about a minute to export is taking over half an hour.
This has not happened with any previous version of Premiere, even with very complex projects.
By the way, I don't seem to be able to dynamically link to Audition to refine my audio tracks. The link is greyed out in the menu.
BTW I am using a high end i7 Windows 8 machine with RAID 0 disks and an Nvidia Quadro 4000 video card.
Arrrrg, what is up Adobe? Serious, what is taking so long? This issue is over 6 months old and still not resolved? I can't tell you how many Premiere updates I've downloaded, and none fix this problem?! That last update was to more smoothly integrate Lightroom and RAW into Premiere? Who cares about Lightroom when using Premiere? Why isn't this fixed, we pay good money for this every month!
Hi Viva Minolta,
Sorry this issue is frustrating you (and others on this thread) so much. I talked with engineering and they said that the workaround given by Lou Colt in post 24 was probably the best way to deal with the issue for now.
Basically, they said that if you have a multicam sequence that is used often with many cuts, then it may be best to pre-render the underlying multicam sequence.
If this is an unacceptable workflow, again, feel free to file a bug report: http://adobe.ly/ReportBug
Thank you for your efforts, but the situation is still unacceptable. If this was some sort of super advanced real-time 3d raytracing feature that was brand new, I might cut you people a little slack. But this is SOUND for film. That technology has been around since the 1920's! The idea of synchronizing sound and picture has been implemented for over 90 years, yet Premiere chokes when it tries to render a multi-cam sequence with cuts?
Its even more frustrating because your real-time 3d raytracing implementation actually works admirably well. So what is wrong with rendering AUDIO?!
Edited by moderator to remove personal attacks.
I read the original version of your post, and I understand you're upset. Not much I can do but ask again that you file a bug report: http://adobe.ly/ReportBug
Please also check that you are not creating a multicamera sequence using the "switch audio" function as I reported in post 29. That function drastically increases render times.
Concise bug reports are the best way for us to deal with bugs. I hope others on the forum also write a bug report. Thanks for understanding.