GH1 Pulldown Removal using FCP / JES Deinterlacer / Compressor

Isaac_Brody

New member

NOTE: I have found that to properly reduce the pulldown I've only been successful from importing the original "Private" file into final cut with prores, and then removing pulldown with either Compressor or JES Deinterlacer. I've actually found that JES looks a little better than compressor. Removing pulldown from just the MTS clip with no metadata from the original folder has been damn near impossible, but with the original folders intact it works as it should. You wouldn't import P2 by stripping the folders apart, and you need to take the same approach to AVC-HD. You remove that original file from the folder structure and you lose all the metadata for it.


My method is what I used when I had an HF10, it worked perfectly for removing pulldown and avoiding any interlaced ghosting or artifacts. This method picks up after you've ingested into final cut with log and transfer.

You can download JES deinterlacer here for free:

http://www.xs4all.nl/~jeschot/home.html#DEI

Take your footage that you log and transferred in Final Cut and open it in JES.

1. Under Input make sure Topfield is checked.
2. Under Project select inverse telecine, select detect cadence breaks (I sometimes will select suppress interlaced scene changes, but am not sure if it makes a difference) and set output framerate to 23.976.
3. Don't mess with color, just leave it on default.
4. Under output put it in a different folder than original material.
5. Under video output select export and choose quicktime movie. Choose prores.
Under your Prores settings make sure framerate is at 23.976, set to progressive, dimensions 1920X1080
Sound setting is Integer(Big Endian) Sample rate 48.000kHz, 16bit, Stereo
and hit OK.
6. And hit OK to have it run. It should strip the pulldown and get you back to native 23.976.

Also, you can setup JES to run as a droplet, meaning that I'll select all my footage in the finder, drop it on JES and it will automatically do the pulldown and set it to a folder. You can select the droplet option under Preferences ==> File. You can also setup a folder to process to under Preferences ==> File as well.

In my experience JES runs MUCH faster than Compressor.
 
Compressor Pulldown Removal:

24p FCP AVCHD workflow:

1. In FCP, Log and Transfer
(AVCHD-> ProRes 422)

2. In the Clip window in FCP:
Highlight all clips and go to Export-> Using Compressor

3. In Compressor:

Pick Format as Prores 422 HQ

Go to Video-> Frame rate-> Custom, Enter 23.976

Turn on frame controls-> Deinterlace settings-> Reverse Telecine

Set your output folder and click submit (this renders pretty fast and Compressor automatically finds the pulldown cadence and removes it)

4.Now import the new folder with new clips into FCP and you have 1080/24p
 
There is probably an obvious answer to this question but, do you have to conform to 24p prior to editing or can you conform your timeline after you have locked your picture?
 
Does JES give you 4:2:2?

For the Compressor method... Why Prores 422 HQ and not normal Prores? You still get 4:2:2 and the AVCHD footage is not 10 bit right? So going to HQ would just use up more drive space and tax the computer more. Or am I missing something here?

Also, is there any way at all to go straight from the AVCHD to Compressor to do the pulldown? Rather than having an extra generation of recompression from going AVCHD -> FCP -> Compressor?
 
4:2:2 is a chroma subsampling used in interlaced video. Look at the illustration in this article http://en.wikipedia.org/wiki/4:2:2#Sampling_systems_and_ratios

Going to 10 bit can give benefits when applying course colour correcting and/or in secondary post treatments like blurring, graining, other 'styles' and colour correcting the result of that.

I suspect/hope that the workflow you ask for will come after the release of Snow Leopard and QuickTime X in September.
 
4:2:2 is a chroma subsampling used in interlaced video. Look at the illustration in this article http://en.wikipedia.org/wiki/4:2:2#Sampling_systems_and_ratios

I understand that, but I read in the other threads that other methods (Neoscene for example) are upsampling chroma to 4:2:2 . I would be color correcting in Color and the 4:2:2 would help.


Going to 10 bit can give benefits when applying course colour correcting and/or in secondary post treatments like blurring, graining, other 'styles' and colour correcting the result of that.

Okay, here my understanding has been that you can't take advantage of the 10 bit if it wasn't shot in 10 bit. That's pretty much what I read and hear all over the place. Has the consensus thought about this changed?
 
So take your non 10 bit footage. Add heavy grading - you've made a NEW image. Not that you've added any original information to the original, but you're adding new stuff on top of it. Vignettes maybe provide the best example. Say you add a vignette in your color grading. 8 bit has a tough time rendering some vignettes without banding. Same goes for any other heavy grading. Whatever your software is adding to the image or significantly altering may be better rendered in 10 bit.
 
That makes sense. So would it potentially be better to have Compressor just do the frame controls/reverse telecine and output as 8-bit, then have Color output as 10-bit when color grading? Would Compressor actually be re-compressing the media again, or just doing the reverse telecine process and repackaging the frames in the MOV file? Is there a way to do this easily in Cinema Tools which won't recompress and then go to 10-bit in Color? I'm just trying to think of a way to save a generation of compression.
 
I noticed JES has a setting to change gamma (default on mine after download is 2.22)
Any comments on adjusting away from quicktime gamma?
 
Don't do it. It's annoying that FCP and QT try to guess what you want, but all of the monitoring in FCP is done assuming that all you monitors and everything else is in Mac Gamma. It compensates to 2.2 so if you also change your material to 2.2 you exacerbate the problem.

Tx!
 
If I use Neoscene or VoltaicHD will it preserve the original time-stamp of the original clip? This is important due to the fact that most people in professional situations will be shooting double system sound. We need the original time-stamp of the clip so we can easily sync it up to our audio files.
 
Aren't all version of ProRes, HQ, LT, even proxy, 10-bit? Also don't they all convert to 4:2:2 unless you're using ProRes 4444?
 
"The Remove Advanced Pulldown and Duplicate Frames option in the Log and Transfer window preferences allows you to remove redundant (duplicate) frames recorded in variable frame rate DVCPRO HD 720p footage. You can also select this option to remove advanced (2:3:3:2) pull-down from 29.97 fps footage such as 1080i and 480i, resulting in 23.98 fps (24p) footage on disk after ingest.

The Panasonic AG-HVX200 camcorder can record 24p footage on P2 cards using three methods:

Native 23.98 fps (no pull-down or extra frames)

23.98 fps with standard 3:2 (2:3:2:3) pull-down

23.98 fps with advanced (2:3:3:2) pull-down

Important: The Log and Transfer window cannot remove standard 3:2 pull-down. To convert footage with 3:2 pull-down, you can ingest the media at 29.97 fps and then use Cinema Tools to remove the pull-down."

Isn't GH1 some flavor of 3:2 pulldown? So this wouldn't apply. I believe this feature is for removing pulldown on stuff in a 29.97 fps stream? Hmm wish apple would support us. Right now I am using the compressor method for when I am shooting 1080p but haven't done much 1080p work...will probably be using JES if I do more but I've had a few stability problems with it (probably solved by making a bottle)
 
Last edited:
Back
Top