AVCHD and FCPX – my two scents

[Joe Marler]

[Sebastian Howard] “I then re-converted in CLIPWRAP leaving the frame rate un-touched. This resulted in the same file sizes (since the number of frames didn’t change, only the number of frames per second.)

Although the file size was significantly reduced (30%), I decided to stay with the FCP-optimized media as I was concerned about quality and grade-ability of the footage. I know the original quality of the source file in AVCHD can’t be great to start with and therefore not very grade-able anyway but I figured I would keep my chances on my side with a higher bit-rate.”

You can just import the AVCHD package directly into FCP X. I usually copy the camera native folder tree (presented in Finder as a single file package) to the Mac then import from there — usually without optimized media. This is from a Canon XA25 with 1080p/60 material at 28 megabits/sec.

The FCP X import dialog is aware of the AVCHD package and you don’t have to drill down to locate the video files.

Clipwrap is great if you need easier Finder-level access to the files. However converting before importing adds a (mostly) unnecessary step. In general FCP X on a modern machine can edit most camera native files with adequate performance and quality. If for some reason you later want to selectively create optimized media on specific clips, this can be done within FCP X after import. Just CNTRL-click on the clip in the media browser, select Transcode…. and pick “optimized media”.

It is often assumed editing optimized media is superior to camera native media regarding color grading, etc. As this article shows, that’s not necessarily the case: https://www.macprovideo.com/hub/final-cut/fcp-x-the-truth-behind-performance-with-optimized-media

My advice is do a few simple tests with optimized vs native media, rather than automatically transcode everything (with all the associated costs) assuming it’s automatically better.

[Sebastian Howard]

Thank you all for your interesting and insightful responses.

First, upon reading many comments above from people that said they could import AVCHD natively directly into FCPX, I got a little flustered, couldn’t understand what I was doing wrong… So went back to FCPX and tried importing again. Well….I discovered (with great embarrassment) that I could simply uncheck the “Create Optimize Media” option and that the footage would import and play natively in the FCPX timeline.

Now I have been editing for a long time and I have to say I was red in the face with humiliation. How did I miss that, I don’t know…

However, I still couldn’t check the “Leave files in place” option. For some reason the files had to be “Copi[ed] to library” It’s probably just a question of re-wrapping the files with the proper extension… And I guess it’s useful if you are importing straight from the camera or card.

But all this doesn’t really solve the problem in the end.

Having worked mostly with XDCAM, Prores and some 5D footage, I wasn’t really knowledgeable of AVCHD and I was under the impression that it was a “semi-pro” format… But after researching a little further, I confirmed that AVCHD is indeed an Interframe format, meaning the video file in there is compressed for playback, not for editing. It’s a consumer format that was designed by Panasonic and Sony for enhanced playback experience, hence the folder structure that allows for menus and such. The root file is MTS but the codec inside is none other than H.264.

Now I know our NLE software systems can work well with these compressed files now days but these files still remain Not-Frame-Accurate. And that’s an issue for me. Plus they can get buggy for a number of other reasons when you start to apply filters and other effects like colour grading and speed changes.

So bringing AVCHD into FCPX natively is no different than bringing in footage from the Canon 5D. It’s just another consumer format that is not favourable for editing. Yes it can be “optimized” through FCPX or transcoded through software like CLIPWRAP but the source footage remains 4:2:0 and thus a lower quality, especially with regards to colour space.

The article posted above by Joe (Thank you Joe for that) has more information on this topic.

And this video here offers a well summarized explanation of video formats and codecs: https://www.youtube.com/watch?v=WpBjGUlBTHU

This all makes me wonder why Technicolor bothered creating the CineStyle profile for Canon DSLR cameras when the video recorded by these cameras is still just H.264. Anyway…

So what is the FINAL conclusion? This is it: AVCHD is consumer. If you want good quality footage to work with (with wide colour space possibilities and frame accuracy), then try to use cameras that can record straight to a 4:2:2 “un-compressed” format, like the Canon C300 or XDCAM HD – OR – use an external recorder to record the raw footage straight to the desired Prores format, like the Atamos or AJA Ki-Pro.

Sorry for the long post. I felt the need to try to clear up the subject (somewhat). Hopefully it can help others.

Cheers

Sebastian W. Howard
Batchfilms
Sculpting Life Into Moving Pictures

Some contents or functionalities here are not available due to your cookie preferences!

This happens because the functionality/content marked as “Google Youtube” uses cookies that you choosed to keep disabled. In order to view this content or use this functionality, please enable cookies: click here to open your cookie preferences.

[Jeremy Garchow]

[Sebastian Howard] “However, I still couldn’t check the “Leave files in place” option. For some reason the files had to be “Copi[ed] to library” It’s probably just a question of re-wrapping the files with the proper extension… And I guess it’s useful if you are importing straight from the camera or card.

But all this doesn’t really solve the problem in the end.”

Again, you don’t have to rewrap as FCPX can import AVCHD transport streams natively (as of version 10.1.0)

You have to dig in to the package (right click on the AVCHD file and choose to “Show Package Contents”, then right click on a file called “BDMV” and chose to Show Package Contents, then open the folder called “STREAM” move the streams out of the AVCHD structure and import.

This cuts down on unnecessary copies of the media on the same drive, and still allows transcoding if you need it. It also breaks the native AVCHD structure and may cause you to lose any spanned clip information, and it might (depending not the camera) create a lot of files that are named the exact same thing, like 00000.MTS

AVCHD is a pain of a format to work with. FCPX (and OSX) have purposefully tried to make this easier, and simply by using FCPX to import and rewrap the footage to .mov saves some trouble at the expense of disk space.

[John Davidson]

I don’t think of AVCHD as complete garbage. My take on it is that to use it professionally you just have to know exactly how you want your footage to look when you shoot with it, set the camera accordingly before you hit record, and don’t have a ‘fix it in post’ mentality.

It’s a lot like working with HDV. You just have to get it right, right from the start.

John Davidson | President / Creative Director | Magic Feather Inc.

[Joe Marler]

[Sebastian Howard] “However, I still couldn’t check the “Leave files in place” option…AVCHD is indeed an Interframe format…the codec inside is none other than H.264…these files still remain Not-Frame-Accurate. And that’s an issue for me. Plus they can get buggy for a number of other reasons when you start to apply filters and other effects like colour grading and speed changes…So bringing AVCHD into FCPX natively is no different than bringing in footage from the Canon 5D….the source footage remains 4:2:0 and thus a lower quality, especially with regards to colour space.”

You should be able to select “leave files in place” if you copy the AVCHD bundle to your local hard drive. FCP X disallows that option (for obvious reasons) on the camera card.

You’re right AVCHD is an interframe codec, but so is MPEG-2, MPEG-4, XAVC and XDCAM. We rely on the codec to reconstruct the P-frames just like we rely on the codec to reconstruct the missing macroblocks within each frame.

I personally have never found a problem with using AVCHD, MPEG-4 or any other interframe codec in FCP X. I have tested the all-intraframe codec (All-I) on my 5D Mark III. It produces 3x the file size at 3x the bit rate, but I can’t see any significant difference in image quality or editing.

However I am biased toward documentary work, where our shooting ratio is typically 50:1. If I transcoded all that before or during import it would be a huge task.

If I was shooting mainly scripted narrative, I might record straight to ProRes using an Atmos Ninja or similar. That would still effectively be 4:2:0 from the 5D3, regardless of the storage codec because that’s all the camera sends.

Vincent Laforet shot “Reverie” at 4:2:0 on a 5D2 and nobody complained. However he transcoded to ProRes for editing since that was on Final Cut Studio. But he did that because he had to — Final Cut back then wouldn’t edit H.264.

[Andrew Moore]

Hi mate

I know this post is old but I saw it when looking something up and thought I would mention a couple of bits that might help. You said on your original post that the files had to go in the library if you are converting to optimised. Just in case you are not aware you can change that and put them wherever you like, external drive as would be normal. Just go to library settings before you import and change the location.

With the AVCHD files you have to put them on your hard drive or better save as archive first, you can then leave in place as you mentioned. This is a safe way of doing it and retaining file data as well.

To be honest the way I did things is based on the fact that I shoot both photo and film here and there so…. I right click a few times as stated above in previous post and copy the MTS files into my client folder template. I then import like any other file. I then immediately build proxy files. No matter how you get the AVCHD or MTS files in they are a pig to edit with. I assemble with the proxy and can normally do some good basic colour work with them before switching back to original for final work.

This is now a mute point anyway as I shoot everything with XACV anyway now and they are all MP4 files so no problem, FCPX rolls with fine. The quality is way better as well. I guess you know most of this now but just in case… Cheers