Exporting in FCPX

I've just tried exporting my video. The first time I exported, I exported as "Master File" which resulted in a 4 gig file (my choosing) using DVCHDPRO 1280x1080. It took about 20 minutes but worked and it looked very nice. I then exported as DVD. This was quite quick and it also worked flawlessly. Now . . . I'm trying to export ACC MPEG 4 for Apple Devices at the same resolution. FCP is estimating the file size to be 300MG but it's taking forever! Over an hour now (and I've tried it twice with the same result). Any thoughts as to what I may be doing wrong here?

Again . . . thanks in advance for an input!

Cheers!

P

You may be doing nothing wrong. It takes a while to compress that 4G to 300M, especially when, like me, you are running on a machine that is almost 5 years old. The amazement is that it does the job at all. :lol:

How long is your video? My machine can easily take 40 ~ 60 minutes to generate a 10 minute 5 Mbit/s. H.264 video using multiple passes.

Hello Marchyman,

Thanks for your post. It's always comforting to hear that you may not be doing something stupid. LOL!

I was wondering if it had to do with conversion or compression or something like that. In a perfect world, sounds like this is a task you do just before hitting the pillow for the night.

The video is 4 minutes in length.

My Mac is only a couple of years old (2010) and has 8 g's of RAM (love to bump that up!).

Based on your file size and the length of the video, I would say something else may be the issue. Since version 10.0.6 FCPX now uses the GPU to render and export. Not sure what computer you have, but possibly the GPU and amount of VRAM on the card is giving you less than optimal performance.

Well hello dgwvideo. We must stop meeting like this.

I'm just using an iMac-i7 quad processor with 8gig RAM and FCPX 10.0.6. I've since, rendered out more files of the same format and all are rendered quite quickly with the exception of the MPEG-4's. They take forever! I'm talking 2.5 - 3.0 hours for a 4 minute vid.

I must say . . . I've learned an awful lot in the past few days, much of which is thanks to your input and other posts here in the forum. Again . . . thank you all!

If you have Adobe Media Encoder, then consider outputting your file at the projects settings (if it was DVCHDPRO then leave it at that). Then add the file to Adobe Media encoder and use one of their H.264 presets.

However, if you have Compressor, you could always setup a virtual quick cluster to direct Compressor to utilize all your CPU cores and increase your rendering speed.

Although i've found that Adobe Media Encoder is faster and produces smaller file sizes.

I'm using a MacPro with 24gb ram, a nVidia GTX 570 video card and Dual 2.26 Xeon Quad Core Processors and I render H.264 files with 2pass VBR faster than realtime (ex: a 4 minutes 1080p Video shrunk to 720p H.264 would take around 3 minutes and 30 seconds). With Adobe Media Encoder I'd be surprised if it took you longer than double the length of the video to render on your iMac - however updating your ram will help.

Fordee,
FordeeTV.com

Okay, I'm salivating. Nice setup! Yes, updating my RAM is a priority for sure. Max is only 16 unfortunately but better than 8

I have both of those programs. I've never used the Adobe Media Encoder. Quite frankly, I didn't know what it was until now do, thanks for that! I did try running this project through Compressor but that'd be a whole new post (I got lost). By the sound if it though, I'd better revisit that approach.

Curious . . . Does it take longer to render a project via export when it has not been consolidated? When I ran into these delays I've posted about, my project wasn't consolidated to the timeline.

Thanks gor your input Fordee!!

What do you mean consolidated to the timeline?

Rendering your project via it's current settings shouldn't take long if the timeline is rendered (i.e.: background rendering is on and completed). And then Adobe Media Encoder is pretty quick... quicker than compressor in fact even when turning on Virtual Quick Clusters.

I have a video on enabling Virtual Quick Clusters... check it out:



Fordee,
FordeeTV.com

Question: you (and everyone else who has given similar setup tutorials) start talking about setting the number of instances of compressord in terms of CPU cores, then switch to number of gigabytes. Huh? Why? Who said?

Example: If I have a 2 core machine with 8 GB of ram should I not bother, use 1 (half the cores), 2 (all the cores), or 4 (half the memory size in gigabytes)?

You should still see a difference utilizing both of your cores. So either: A) Set it up to use 2 (both your cores) or, Set it up to use 4 (2 real and 2 virtual cores - because you have 8 gigs of ram). Do a test to see which is faster... I've heard differing results from different people.

Have a good night!

Fordee
FordeeTV.com

Hi,
Interesting tutorial. I have a Mac Pro 12 core dual 2.93 GHZ and have 64 GB ram. I used your method to setup 24 cores (I was not presented with an option for more than 24). Any ideas why 32 was not available?
Regards,
Dan.

Dan,

You can only setup 24 cores because you have a 12 core machine. It's not based on your ram, it's based on the cores. So, 12 core can run 1-12 real cores, anything above 12 is a virtual core... you are able to assign a virtual core for every core you have. Although you may see better performance setting it to 12... so try both 12 and 24. The rule of Thumb for Ram is that you should have 2gb of ram for every core you assign. So, if you didn't have 48gb of ram, then you shouldn't assign 24 cores (12 real, 12 virtual). Since you have 64gb, you can assign up to the max amount of cores for your machine. Hope this helps.. have a good night!

Fordee,
FordeeTV.com

@pwillis one thing to look at is the setting for apple devices do you have it set to 1 pass or 2 pass encoding. If it is set to higher quality and 2-pass it will take significantly longer (not just twice as long) this is because the way Final Cut Pro x works when outputting to h.264.

Let me explain:
On a single pass h.264 (mpeg 4) encode Final Cut Pro will take advantage of your GPU to do a lot of the rendering and encoding.
On a two pass encode it can not use the GPU (this is also true with hardware acceleration add ons like the Matrox Max)
So on a two pas encode it doesn't take twice as long but significantly longer on both the first and second pass since you are not using the GPU

2-Pass can provide better quality on high action footage and footage that has lots of cuts and transitions, in most cases a high bit rate with a single pass will get the job done.

If you must do a two pass encode I always recommend exporting a master file then encoding from that since half of the heavy lifting is already done (its been rendered)

Hope this info helps

Sorry for being MIA the last couple of days.

This is all fascinating to me as I'm such a nube to the FCP world. That said, this discussion is making sense to me. I'm curious though about re-rendering a "master file". I don't know if that's the correct term or not but let me go off for a moment with some of my "nube logic" here and get this straight:

If I were to say, export a Master File that was of lesser quality than the original (project) would that not limit me in terms of the quality I could have achieved when exporting h.264 from the project initially? I hope I've worded this cohesively. Obviously we're in the digital world here where degenerative copies are not as obvious as they are in the analog world but still, wouldn't it be best when making a copy (exporting) to make one from the source? If I could do this and lessen the waiting/rendering period, well . . . great! But, respectively, it sounds too good to be true.

For example, if I were to make an MP3 copy of a 16bit .WAV file, I would end up with a far lesser quality sample. In the end, it's all about what you can hear/see and then of course, the age old argument ensues: What is the end user watching/listening with?

For YouTube, one could easily argue that it's all crap but there are notable differences between work which was dome well and that which was not.

As per the "core" issues, I'm afraid I'm barking up a tree which I can't climb at the moment given my 4 core limitations and I'm assuming the price I'll have to pay for that is . . . waiting for rendering to complete.

When you render your project at your project settings... by default it's ProRes 422, so you shouldn't be seeing a quality hit unless you started with something better (Red footage for example). When you render to a smaller file format it equals compression which equals some loss of quality. However, same as audio... start with shit, end with shit. By starting at a higher quality file you'll end with a higher quality file.

Am I making sense? I haven't had any coffee yet and I'm pretty tired today for some reason.

Fordee
FordeeTV

fordeetv wrote: The rule of Thumb for Ram is that you should have 2gb of ram for every core you assign.

That was the missing information. It was implied, I guess, but never stated. At least I don't remember ever seeing it stated. I guess I was looking for a statement such as you can allocate up to 2 times the number of physical cores to a virtual cluster so long as you have at least 2 GB of ram for each allocated core. That would have opened my eyes much sooner.

Geeze, I replied to this but it looks as though it was never posted. Pilot error I guess.

Well Fordee, I'm gonna give this a shot. Sounds great, and sure would be saving me some time. What is Red footage exactly? Not that it's gonna happen anytime soon, if at all but, if I were submitting footage to a network or similar, what would I expect them to request of me in terms of format/quality?

My vid's are all really amateur but I'm hoping to improve substantially and so far, thanks to hearing what you guys have had to say, is helping a lot. Sorry for all the audio analogies, I'm just trying to find a constant somewhere to grab onto.

marchyman wrote: That was the missing information. It was implied, I guess, but never stated. At least I don't remember ever seeing it stated. I guess I was looking for a statement such as you can allocate up to 2 times the number of physical cores to a virtual cluster so long as you have at least 2 GB of ram for each allocated core. That would have opened my eyes much sooner.

Much more eloquently stated...

Fordee
FordeeTV.com

Pwillis,

Red footage comes from a Red Camera... you are probably far off from that workflow at this point. Tackle one step at a time and you'll see your knowledge grow by leaps and bounds.

Good luck with the continued learning!

Fordee
FordeeTV.com