Questions about editing with canon 5dmark ii footage on FCP7

Hello Everyone,

I have been trying to read up on this but still have some questions to you. I am feeling a bit lost here and am grateful to anyone with some advice.

Computer: MacBook Pro 15inch, 2.7GHz Intel Core i7, with OS10.8.
Editing on FCP 7.0.3.
Camera: Canon EOS5d markii.

1.>I already started editing by just dragging and dropping my 5D .mov H.264 files into FCP. I dropped them into a sequence and then I answered yes to having the sequence set itself to the clip settings. It actually works. But I know that here everyone is strongly advising against this. So I do really need to convert to ProRes422, correct?

2.> I do have this problem that when I mix freeze frames (captured from the footage) with footage on the same sequence timeline, the playback keeps stopping. Is this because of H.264?

3.> Which Apple ProRes 422 format do I convert to? The 422 HQ?

4.> WHich software is best for converting to ProRes 422?? I seem some people just use media manager from within the browser which seems easiest. But will this give the best quality? Or others use MPEG stream clip.

I am interested in getting the highest quality possible for my footage.

Thank you very much to all.

Cheers,
Robert

> I already started editing by just dragging and dropping my 5D .mov H.264 files into FCP.

Big no-no. FCP7 can't edit H.264 files properly, and you'll run into a lot of trouble doing it this way. Convert all footage to ProRes QuickTime movies first. The Canon E1 plugin will allow FCP7's Log and Transfer function to batch-convert the footage, and add timecode.

> I seem some people just use media manager from within the browser which seems easiest.

Also incorrect. Media Manager is used to consolidate and trim footage. What you want is Log and Transfer.

> I am interested in getting the highest quality possible for my footage.

It's unfortunate that these days all people think about is "quality". Video quality is only one aspect in the equation. Safe-guarding your editing decisions is another concern. There's also long-term stability, flexibility, compatibility with your system, amount and speed of your storage units...and so on, and so on. If your editing workflow is solid, then you can capture at offline quality to save storage space and then online only part of the raw footage upon picture lock. But if you're ingesting a 10-hour documentary miniseries with a 200:1 shooting ratio and you're capturing at the "highest video quality", say uncompressed 10-bit HD, you'll be dooming yourself to needing to buy 100TB of storage for no reason. Yeah, it's "high quality", but you'll be wasting tens of thousands of dollars on unnecessary equipment.


www.derekmok.com

Hi Derek,

Thank you very much for your sage advice. This is a great help.

>>Now I am trying to start from scratch and want to convert my footage to pro res422. I have copied all files already from the CF cards to my hard drive (and have already reformatted/erased the CF cards).

I have installed the Canon E1 plug in version 1.1. I was unable to find the v1.2 anywhere on the internet. So when I open log and transfer in a new project in fcp7.0.3, I try to import the volume, ie the folder with the .mov and .thm files. But it then tells. me

"100EOS5D_Mark02_Street" contains unsupported media or has an invalid directory structure. Please choose a folder whose directory structure matches supported media."

Maybe I need the v1.2 but I cannot find it anywhere.

So what can I do? Or use mpeg stream clip?



>> Thanks for your advie on the quality issues. I understand what you are saying. Usually I make docs but this is a relatively short music video so I will try the pro res 422 HQ.

THank you and take care,
Robert

> So when I open log and transfer in a new project in fcp7.0.3, I try to import the volume, ie the folder with the .mov and .thm files. But it then tells. me

Another extremely common amateur mistake. You probably copied only the .MOV files. First rule of shooting tapeless: You don't mess with the original folders, even when they appear empty or useless. When you back up the cards, you have to grab every single file, folder and item from the root of the memory card and keep everything where it was.

There are ways to "fool" FCP by trying to simulate the original file structures, but I've never had success with any of them. At this point you're probably better off using Compressor or MPEG Streamclip to convert the files.

Another thing: The Canon E1 plugin/FCP7 combination doesn't work on Canon 5DmkIII footage, only 5DmkII. With 5DmkIII footage, you have to use Compressor/MPEG Streamclip as well, or another kind of media conversion software. Digital Rebellion Pro Media Tools may also do that job, but I haven't used it. (Jon Chappell may jump in with some more recommendations here)


www.derekmok.com

Hi Derek,

thank you again for your reply and for your advice. I guess one learns from one's mistakes.

>I finally just found canon e1 v1.5 and it seems to work!!. It is transferring them and I can open them in fcp. So do you think this will be okay? (I am probably making more mistakes....)

>I created a new project for the log and transfer. I noticed it stores the files in a separate folder with the name of this project, within the capture scratch folder. How do I get these files into my edit. By import-> file??

>As I said I had started editing with the h.264 clips. Is there some way to replace these easily with the pro res clips? Or do I need to start from scratch?

Thank you again.

Best,
Robert

> I finally just found canon e1 v1.5 and it seems to work!!. It is transferring them and I can open them in fcp. So do you think this will be okay?

Try it. Wrong editing methods don't reveal their issues right away. A stable system takes weeks, even months to field-test. Edit with the media, test the system with multiple projects and learn what you can about how the system behaves. Filmmakers are usually inclined to use one months-long project to test their systems and their knowledge. That's not how you learn editing. That's why people who claim to be editors, but are actually from a directing/producing background, often fail when put into a pure-editing context. They don't have the necessary experience and have not been put into the correct context for learning hard-core editing. They know how to work one project on their own system, that's it. They know how things behave and where everything is, but their methods aren't workable in a pure-editing context.

You learn editing by starting multiple projects, over and over, so that you are forced to go through the entire workflow repeatedly. That's how you learn what you need to.


www.derekmok.com

Thanks so much Derek for all of your help. Believe me I have great respect for people like you who dedicate themselves to the great art of editing.

Sorry to bother you with 2 more questions but it would be a great help, especially Q1:


Q1>As I said I had started editing with the h.264 clips. Is there some way to replace these easily with the pro res clips? Or do I need to start from scratch?


Q2>I created a new project for the log and transfer, Canon E1 process. I noticed it stores the files in a separate folder with the name of this project, within the capture scratch folder. Do I get these files into my edit simply with import-> file??

THanks!
Cheers,
Robert

> As I said I had started editing with the h.264 clips. Is there some way to replace these easily with the pro res clips? Or do I need to start from scratch?

Easily, no. H.264 clips are unstable and have no timecode, so an automatic reconnect to the ProRes versions may not retain your actual editing decisions. You may need to conform these by hand. H.264 clips should all have 00:00:00:00 as the starting timecode, so Match Frame each edit, spot the In point timecode, then locate the identical frame in the ProRes version.

This is why editors don't start projects until they are sure of the accuracy of the workflow. Fixing these problems after edits have started takes 10 times longer. I spent two weeks fixing up a short film in September before I could start the recut, because they edited with improper media and FCP workflows. The minute we switched editing stations, all the sync sound reconnected wrong and had to be renamed and then resynced and reconformed by hand. If the work had been done properly in the first place, I could have started editing right away instead of dealing with technical snafus.

> I created a new project for the log and transfer, Canon E1 process.

Why?


www.derekmok.com

Quote
Robert
WHich software is best for converting to ProRes 422?? I seem some people just use media manager from within the browser which seems easiest. But will this give the best quality? Or others use MPEG stream clip.

derekmok gave Robert a nice lecture about quality and methodology, but there is a simple question in the quote, which deserves a simple answer. I think it makes no quality difference which software you use to convert H.264 to ProRes (in one or another flavor). This is because H.264 decoding is univocal (while there are great differences in H.264 coding) and ProRes coding is univocal. Really horrible software can screw up anything, but this particular transcode should be straightforward.

* * * * *

"Codec" abbreviates "code + decode". A codec gives a description of the decoding: the algorithms by which the data in the video file produce the colors at pixels on output. A codec may or may not give a description of the coding: the algorithms by which the colors at the pixels of the input produce the data in the video file. In some cases the coding is obvious from the decoding, so there's just one way to do it. Conceivably the coding may be unobvious from the decoding, but still with one best way to do it described in the codec. There's a third possibility that the best coding is unobvious from the decoding, and there is no best way to do it, but due to proprietariness, the codec must describe the way it will be done. I don't know which of these is the case for ProRes.

For a codec like H.264 there are competing coding algorithms. Some are fast; some are slow. Some provide much better picture quality than others. Etc. This by the way explains why it is treacherous editing H.264 footage directly, in any NLE. Effects editing requires re-encoding. How well does the NLE do it?

Dennis Couzin
Berlin, Germany

> This by the way explains why it is treacherous
>editing H.264 footage directly, in any NLE.

You should try cutting in Premiere for native editing. There are a lot of under the hood processes that facilitates native editing.

Yes. As Derek said, if you are cutting in FCP7, use log and transfer. It let's you batch recapture the footage later on as a fallback.



www.strypesinpost.com

strypes: I always convert H.264 footage to ProRes HQ before editing. At worst, this is unnecessary work and expense. A test can determine if Premiere's direct editing of H.264 is as good as editing by the indirect route. The test should include complex time effects, layered. We already know that HQ is solid for this, so differences between the two outcomes will point to weaknesses in the direct H.264 editing.

Dennis Couzin
Berlin, Germany

Thanks very much to Derek, dcouzin and strypes for your advice! I greatly appreciate this.

Now that you guys have gotten me this far I have a further question so that hopefully I will do things the correct way:
>Derek and dcouzin, I used the Canon E1 v1.5 within log and transfer. I do not know what the best procedure is for doing this. I had seen in other links that people make a new project. And when I did this the footage is automatically stored in a separate folder with the name of this project, within the capture scratch folder. So it seems Derek that you are implying that there is some more efficient and correct way to do this?

Thank you all again!

Best regards,
Robert

>A test can determine if Premiere's direct editing of
>H.264 is as good as editing by the indirect route.
>The test should include complex time effects,
>layered. We already know that HQ is solid for this,
>so differences between the two outcomes will
>point to weaknesses in the direct H.264 editing

Test it out then. Most of us are quite happy with the performance of H.264 sources from DSLRs in Premiere that we see no benefits in transcoding.

> I do not know what the best procedure is for doing this.

How to manage media in FCP7 depends on the type of project you are working on. For short form work, renaming the clip to include the reel name (which should include the date of shoot) is sufficient. There are renaming options available in the log and transfer window.



www.strypesinpost.com

H.264 is stable in Premiere. You can cut and manipulate it as if it were any 'proper' codec.

Quote
strypes
Test it out then. Most of us are quite happy with the performance of H.264 sources from DSLRs in Premiere that we see no benefits in transcoding.

Quote
Jude Cotter
H.264 is stable in Premiere. You can cut and manipulate it as if it were any 'proper' codec.

I don't have Premiere. It might be enough to find out what bit rate Premiere uses for its "native" H.264 when it renders an effect. At high enough bit rates H.264 quality no longer depends on the encoder and might be as "visually lossless" as ProRes HQ is. Someone with Premiere can render a long H.264 double exposure and calculate the bit rate.

Dennis Couzin
Berlin, Germany

I don't think anyone with Premiere cares enough about why it works well in Premiere and not in FCP7 to do tests. It works, ergo, we work.

If it matters a lot to you, you could download a free 30 day trial and do the tests with that.

>It might be enough to find out what bit rate
>Premiere uses for its "native" H.264 when it
>renders an effect.

It doesn't decompress h.264 and force you to render to h.264. You basically play and edit h.264. You only render in timeline if your machine is not fast enough, which may be the case if you have very complex effects that your machine cannot process in real time. And for this render format, or what Premiere calls "preview format", this can be one of many formats including QuickTime codecs such as Prores, Uncompressed, or MXF formats such as DNxHD, XDCAM HD422, XAVC or even Avc-Ultra, depending on what your workflow is. This would be a little similar to having h.264 in a Prores timeline in FCP7 except that Premiere doesn't choke or crash with these formats and the playback is smooth. Also the real time performance of unrendered effects is much better than FCP's RT Extreme engine. So you generally render less during the editing process.

On final output, you can choose to render out to whatever your delivery format is. You can also choose to use the preview files that you created during your edit for the export.



www.strypesinpost.com

strypes: Thanks for this information. So Premiere has a different idea of sequence settings than FCP7 has. In Premiere, as I understand you, when a clip with or without effects is playable, it plays. When it is unplayable, then it must be rendered to one or another chosen "preview format". In FCP7, a H.264 clip with some or no effects in a H.264 sequence plays. But change the sequence settings to ProRes, or to any other codec, and a red render-required bar appears (at least with my render settings).

I suppose FCPX shares Premiere's philosophy on this. They regard the video as the primary thing being edited and the codec as an incidental cipher for the video. So they're happy to play back a sequence consisting of clips in an assortment of codecs provided they're each playable. FCP7 regards the codec as essential to a video's appearance, so the mixed sequence must be homogenized into a single codec for valid viewing.

Dennis Couzin
Berlin, Germany

To some degree... FCPX seems to prefer consolidating to the QuickTime container as in the case of XDCAM EX and Avchd. I suppose that helps media management. Premiere will work with the file directly.

FCP doesn't really need to have everything in one codec. Quite often I have different codecs in the same timeline - I usually set my edit sequence to prores proxy for faster renders then toggle it back to Prores HQ when I'm done editing. The issue with H.264 is that they did not optimise the program to work with H.264 so the software becomes more unpredictable with it and many users encounter frequent crashes. Also the RT Extreme engine is very outdated.

Premiere doesn't force you to render or transcode. If it cannot playback a clip in real time, the user can also choose to reduce playback resolution in the timeline.



www.strypesinpost.com

You can throw a bit of everything into the FCP7 sequence, and FCP7 will then put either a red (needs render) or dark green (doesn't need render) bar onto anything whose codec differs from the one in the sequence settings. (I'm assuming resolution and framerate already agree with the sequence settings.) When H.264 is in the sequence settings all the non-H.264 clips get red (at least with my render settings). (Likewise when codec None is in the sequence settings.) But when a ProRes codec is in the sequence settings many kinds of clips don't need render. (Likewise when an uncompressed 422 codec is in the sequence settings.)

Suppose codec X is in the sequence settings and a clip with different codec Y is in the FCP7 sequence. How does FCP7 decide between a red bar or a dark green bar for this clip? Does FCP7 measure the computational difficulty of decoding Y then re-encoding as X then decoding X? Is all that required just to play the clip? I doubt this since None is easier to encode and decode than ProRes is. How FCP7 decides red bar or green bar is a mystery, but the decision surely depends on both the X and the Y. All matters of bugginess and optimization aside, this double dependence distinguishes FCP7's treatment of codecs from Premiere's (as I understand it). Why is FCP7 this way? The charitable answer: FCP7 regards the codec as essential to a video's appearance. A damning answer: FCP7's programmers were lazy.

Dennis Couzin
Berlin, Germany

Fcp decides what gets a light green or dark green bar by doing system profiling. This is done on the initial start up of FCP7 and whenever preferences get trashed. The red render bar also turns on for formats that are not optimised by the RT Extreme engine.



www.strypesinpost.com

According to Apple, RT Extreme concerns effects playback. With "Safe RT" render settings, any H.264 clip, containing no effects at all, gets a red bar if the sequence settings have a codec other than H.264. But there's no render bar for that H.264 clip if the sequence settings have codec H.264.

PCP7 is sometimes thoughtful about its red bars. Even changing the parameters in a filter can make the bar come and go. What is FCP7 thinking in the first case?

Dennis Couzin
Berlin, Gemany

Fcp assumes it doesn't have to render because it is the same codec as sequence. But it does not mean that the RT Extreme engine supports it. Drop any effect including a cross dissolve and the red render bar shows up.

Btw, switch to unlimited RT. "Safe" was a legacy setting. With unlimited RT set for a sequence with supported codec settings, the worst you will get is an orange render bar which means FCP may beachball or drop frames to keep up with the playback. The render bars with unlimited RT effects is decided by system profiling.



www.strypesinpost.com