C100: Need Advice with these Post Settings

nedcam

nedcam

Well-known member
I've been shooting for various Russian clients, mainly TV shows and they pass my name around. Usually I send raw footage. However, it appears this new client for Tuesday wants me to ingest the footage (I have FCP7) encode and export with the following settings and I don't know why they put 1080p under the audio section, I assume it was lost in translation:

2. Тech requirements:

А) General requirements:
- File Format: MP4
- Video Format: H.264 (compression type)
- Data Rate: 3000-5000, Optimized for: Streaming - Image Size: 960x540 HD
- Frame Rate: 24p
- Keyframe: Every 24 frames

B) H.264 video options:
- Restrict profile(s) to: Main
- Encoding mode: Best Quality (multi-pass)

C) Audio settings:
- Format: AAC-LC (music)
- Data Rate: 320kbps
- Channels: Stereo
- Output sample rate: 44.000 kHz - Encoding quality: Best
- Resolution: 1080p

So dear friends, knowing this, and the fact that I am filming 5 hours of speakers at a symposium, what do you think are the best settings for the Canon C100? I have 3-16GB cards. I'll be working alone so I'd prefer not to eject and download on the fly unless I have a lunch break. If I learn that the speakers will be moving around a large stage I may use my Panasonic HPX370 so the servo zoom will track them better than on my 24-105.

This video is for a MOOC course (online university) that's probably why their window size is strange. I own MPEG Streamclip and Clip Wrap so maybe I will convert there instead of the time for FCP7.
I do not have an on board recorder nor do I want to rent one for this rate. In fact, I didn't even know about this time for post when I accepted the gig!

Thanks for your advice.
 
Last edited:
And yes, they surely made an error by putting 1080p in the audio settings.
The settings on the C100 doesn't really matter. Though I'm not sure you'll have enough with 48GB to run for 5 hours. With a bit rate of 24, you'll have about 4 hours of footage with your 3 cards. At a bit rate of 17, you'll get a little over 6 hours.
What that really matters, is the way you'll compress the footage in post. But the again, they seem to know what they want by giving you the settings.
 
SD cards are so cheap that I don't understand why you don't just buy some more higher capacity cards. Get some 32GB or 64GB cards for longer recording. If you don't trust them, get two and run them in parallel.

You don't have to spend a fortune on high value cards either. We have 28 Tanscend cards and have never had a problem, yet their SD cards are unbelievably low cost compared to Lexar and Sandisk.

I never go on a job without two or three times the card capacity than I think I'll need and you're cutting your own nose off by limiting your bit rates just to fit on such small cards.
 
Yes I figured I buy some more cards. I've sent them an email trying to get them to accept the raw footage to get out of the encoding hassle. But here's a 10 hour difference and I'm sure they're busy with the Olympics (and Ukraine).

I also figured there's no lesser setting I can get away with if they've requested 1080. Don't know why they want 24 if it's for streaming. Thanks for your feedback.
 
Those specs look like fairly standard FCP7 "Export using Quick Time Conversion" settings, with a typo on the audio spec.

The following comments assume this job is relatively time-critical.

It sounds like you don't need to edit the clips, so I'd say FCP7 is definitely the wrong tool for the job, as it's not optimised for an MPEG4 workflow. If you take it as a given that you will shoot AVCHD at full quality 24Mbps (a lesser codec will not really be an advantage - bear with me), FCP7 will need to transcode this to something like ProRes 422. You've then got to export and re-transcode, using Quick Time conversion. This is where FCP7 shows it's age, not really being designed for MPEG4 workflows - although it's still my preferred editing tool.

If you do need to edit (add captions, top and tail etc), I would strongly suggest running a test. Use the gruntiest Mac you've got (iMac 27 Quad Core or top spec MBP.) Shoot exactly one hour of AVCHD, make a note of how long it takes to ingest into FCP7, add your manipulation/editing time and then export with QT Conversion. Once you've got a total time to do that, scale all of that up to 5 hours of original material.

You may otherwise want to consider Adobe Premiere/FCP X - both much more optimised for MPEG4/H264 workflows. They will allow you to run an H264 native workflow right through from ingest to delivery.

If you're not editing - and just delivering the stream from the camera without manipulating it - Clipwrap may well be the best bet. You've just got to be able to unwrap the native Canon wrapper and then transcode/re-wrap to the lower res H264. If that doesn't work, FCP X will definitely do this workflow nice and quickly for you.

Whatever you do, preparation and testing will be your friend!

If it's not time critical and you can run the export workflow overnight, just get some larger capacity SDHC cards (32 Gig+) and don't worry about it.

Sorry - have edited this a couple of times, as digital workflows have so many variables and I wanted to give you some useful advice, based on painful experience!

Ben.
 
Last edited:
Thank you so much Bengiles. I do have FCPX and also Clip Wrap. Which do you think would be best? I am not doing anything to the footage such as captioning, fades, titles, etc. Just prepping it for the FTP.
 
Hi Ned.

I think FCP X will require less steps. If you get an online connection, you should be able to set up the exact output spec and FTP server as one of the "Share" destinations. Or you can just fire it at a hard drive.

Again, test, test, test!

Suggestion:

- Consider employing an assistant/DIT to keep a watch over this process and to organise cards while you shoot. It might be a false economy not to.
- Get some more SDHC Cards, format them all beforehand and run a dual card workflow on the C100 (recording in parallel to both slots.) 16 Gig cards (approx 90 mins per card) should comfortably capture each speaker, with plenty of headroom - but do the math beforehand.
- Use a larger capacity 64 Gig card in slot B as your rolling backup. That should capture the entire 5 hours of material.
- At the end of each speaker, ingest the latest Card A, leaving the "optimise media" unchecked. That will pull in the AVCHD and simply re-wrap it as a .mov. You don't need FCP X to also transcode to Pro Res (what it dumbs down to "optimise media.")
- Once it has ingested (test the time it takes to ingest/re-wrap beforehand - it's pretty quick), add any relevant metadata (clip name etc), File/Share the clip/s to the destination you set up before.
- Keep the B card/s until client has paid you/signed off on successful delivery.
- This workflow also allows you to view/check clips through the day, while it ingests and uploads in the background (as before, use a grunty Mac.)

Cheers,

Ben.
 
+1 for Ben's thoughts. Probably too late for this now, but:

https://mattdavis.clarify-it.com/d/turbo264templates

I use Turbo264 for fast turnaround stuff at high bit-rate - it's quicker than real-time, and would be very easy to cook up a template from their specifications (but do double-check that 960x540 size - I can understand it if it's just a simple talking heads thing with slides...). I am asked to post rushes like this quite a bit, and high bitrate (5,000-10,000 kbps) H.264 is usable in the right hands, though being restricted to the Main level is a bit sad. LOL

With high bitrate (and with the High profile) on talking head, the value and 'time suck' of multi-pass is somewhat devalued. When you're doing your final encode and honing your compression in limited bandwidth situations, multipass is essential. With huge gobs of bitrate, with the intention of transcoding to an edit friendly codec at the other end, multipass is spending unnecessary time IMHO. Far better to use the High profile and get even more efficiency out of the encoder for a given bitrate.
 
Turbo264 is available as a software-only version - though the one that came with my dongle worked just fine too.

There was a sort of '2001 bone banging' moment when I, erm, forgot to plug it in when using my then shiny new MacBook Pro, and it went FASTER than real time. Gosh, that was a moment to savour.

You can just drag .MTS files straight to the window, but that's not how I use it. I've always used FCPX, ClipWrap or 5DtoRGB as middle-men, but taking MTS files and turning them quickly into something you can view is what Turbo.264HD all about.

It's been my dirty little secret for some time now. LOL
 
Matt, what's the H264 output quality from the turbo.264? Better/worse than (say) Compressor or Media Encoder, or even the X264 plugin to Compressor (which I know is really slow)?

I've read both good and bad reports but none of them from people I think had a clue what they were talking about!

I've found my MacBookPro (2.5Ghz i7) is 5 times faster than my 8 core MacPro 3,1 (2.8Ghz Xeon) for h264 encoding, but if I could get it even faster that wold be awesome.
 
X264 is still the best H.264 encoder bar none. Really very, very good - and give it time and multi-pass so it can make its magic. But save it for the final outputs.

Turbo.264HD is on a par with Compressor, if not (hmmm, maybe) better than that (Compressor's only been really good for DVDs). We're talking high bitrate and High profile, and quite frankly one shouldn't be seeing too much difference until you start comparing noise-differences. Adobe Media Encoder? Pfft. Bundleware. :beer:

I don't want to bang on too much about Turbo.264HD as being a high quality thing - it's as exciting as a colander. It's just a process one has to go through in the endless chain of edit-encode-web-client-changes-encode... But then again, when doing straight rushes, it's fine. It gets the job done on time within budget to specification.

Now, if we're wanting to talk quality, final renders, I use Episode Pro with X264... Slightly different dynamic!
 
Thanks Matt. I'm finding my MBP exports H264 at about double real time, so the 13 minute sequence I just exported to H264 from FCPX took around 6m22 seconds. Is H2.64 going to be much faster than this?

Comparing X264, is this any better in Episode than the free Compressor X264 plugin?
 
Thanks All. I dodged a bullet. Got the Russian client to accept the raw files via FTP so I don't have to spend (non-paid) time messing around with conversions.
 
Hi Friends,

I could use some more advice. Turns out I did NOT dodge a bullet and the client now wants me to do the conversion after all. Guess it is the language barrier and 10 hour time difference.

First I bought and downloaded Compressor 4 to use with my FCPX and all it does is freeze and I have to force quit. So now I have an email into TURBO H264 to make certain their software will keep the audio tracks separate and allow me to resize to 960x540. They haven't answered yet so...

I figured I should get started with FCP7. I am importing the original AVCHD Private file into Clip Wrap and exporting as Apple Pro Res 422 so I can bring it into 7 and export using the client's specs. I have Compressor for FCP7 but I figure I can also export from FCP itself. I realize this will take awhile but it's OK, I'm not busy and I can do it overnight too. It is 3 hours of talking heads but only 33GB.

I am mainly a DP with rudimentary editing skills. Am I on the correct workflow path? I will FTP them a short example first.

Thank you
 
If you're doing the initial conversion with Clipwrap, I'd just use Compressor to do the transcode to H264. Create a recipe/destination in Compressor, do a quick test with a short clip, play back in QT7, using the Movie Inspector window to check spec of the exported clip (careful that the size it displays is not just a scaled size.) If all OK, you can then just batch export with that recipe and bypass FCP7. FCP X would be quicker, as you're not bouncing up to baseband, only to then transcode to ProRes, back to baseband again and then on to H264. But this workflow should do the job.

Ben.
 
Thank you for that idea. Meanwhile I did hear back from Turbo H264, they sent me a demo and I did a 3 minute test. I was able to keep audio stereo plus meet the other specs of the client. I sent them the test and will hear back if it is OK in which case I'll just buy the program.

Many thanks!
 
You must log in or register to reply here.
Back
Top