Final Cut 6.0.2 and HV20 Workflow - Questions about Supposed 1080-24f Easy Setups.
 

Hello

My first post on the forum. I am a fairly new HV20 user, and have some workflow questions regarding Final Cut 6.0.2 wrt the HV20 - forgive me if this is answered elsewhere, I could not find a definitive answer concerning FCP 6.0.2s newer easy setups.

Short version - is it possible with FCP 6.0.2 to capture to a format (ProRes/HVD/AIC etc - I dont care) - but use log and capture AND get reverse telecine while capturing? I would expect this to be possible with interframe codecs like AIC/ProRes, not HDV with its long frame cruft.

I am aware of the reverse telecine process that must be applied to properly conform footage shot at cannons 1080p24 mode, and I am also aware that apparently the HV20 does not add cadence metadata to the video stream, so your reverse telecine software must have cadence detection and all that fun jazz (a la compressor).

With Final Cut 6.0.2 the release notes it mentions support for Cannons 24f mode with the easy setups as follows:

http://www.apple.com/support/release.../doc/uid/link2 (from here).

****
HDV Workflow Improvements
The following HDV Easy Setups should be used specifically for capturing footage in Canon HDV 1080F24, 1080F25, and 1080F30 formats:

HDV - 1080p24 FireWire Basic
HDV - 1080p25 FireWire Basic
HDV - 1080p30 FireWire Basic
****

Does this mean one no longer needs to use the HV20 workflow describe by Apple ?(http://docs.info.apple.com/article.html?artnum=306389)

Ive shot some material @ 1080p24 on my HV20, and tried three workflows as follows:

A) HDV 1080i60 FW Basic -> ProRes via Compressor Rev Telecine

B) HDV 1080p24 Firewire Basic (new in 6.0.2)

C) HDV - Apple ProRes 422 1080p24 (new in 6.0.2 - but mentions only the Sony HVR-V1)

My results are rather... interesting, so I feel I must be missing something basic about the new 6.0.2 improvements.

With the "traditional" workflow "A", I can log and capture and then transcode my material just fine. It seems to remove the pulldown as expected. I just dont want to spend the time capturing and transcoding :(
For reference, every 5th frame in the viewer in FCP before conversion is doubled up.

With workflow "B" I have captured a 29.97fps HDV movie, which has no pulldown removal as far as I can tell, every 5th frame shows the typical ghosted double frame ( like so http://www.flickr.com/photos/missilemike/494788313/ ) if I view in QT player of the viewer.

When I put this clip into a sequence created with the B easy setup, I am asked if I want to change my sequence from 23.98 to 29.97, which, uh, is not what I want to do. Selecting no gives me the nice little preview green line above my clip. *mutter*. However it seems to playback ok?

With workflow "C" its basically the same as workflow "B", but I have a ProRes encoded clip @ 29.97, with every 5th frame doubled (if i view in QT/or the viewer) , and I also am queried about changing my sequence setting. The other annoyance is that I was not able to log and capture, but presented with a window to name my clip, and then FCP just started doing 'capture now' (essentially).

However, with B and C it *seems* like it plays fine in the sequence if I tell it not to change anything? I'd love a bit of clarification about this. It seems like the sequence is properly skipping the redundant frames at 23.98.

What are others experiences with 6.0.2? What is the point of the 1080 24f support in 6.0.2 - does the HV20 shoot at this 24f mode ( does this refer to Cannons lack of metadata cadence info)? There is no way to capture at a true 23.98 framerate from an easy setup?

Thanks! Im a bit confused! (can you tell? :) )

I believe these workflow improvements are for the pro line of Canon camcorders, like the XL A1 and H1. Those cameras use the 1080F24/25/30 naming system, which are actually 1080p recordings.

Thanks Ian, that seems to clear up the issue for the easy setups, anyone know about the workflow questions?

So the HV20 is NOT using Cannons 24F, (my confusion stems from reading a posting on an alternate forum stating the HV20 was indeed 24F, and, within the manual its 24P mode is referred to as 24FP). Joy!

Thanks again,

All the 24 I have seen from tho HV20 is interlaced.
what's this about?
How are people dealing
(Obviously you could deinterlace in your NLE - But what's the cost in Res w. the HV20?)
Thanks

Unless I am horribly ignorant and have mis-read just about everything out there on the HV20, while it shoots @ 1080i60 - in 24 Cine mode it inserts a pulldown pattern to the 60i material resulting in an interlacing pattern that is essentially progressive (if the two fields are from the SAME moment in time, its progressive...) , and this can be inverse telecined to achieve TRUE progressive 24fps footage.

However the caveat is that the HV20 does not put in the cadence metadata so your NLE/deinterlacing software has to either be smart enough to recognize the cadence on its own, or be explicetly told by you.

My questions above were about sidestepping the post processing after capture, and using FCP 6.0.2 to do on the fly pulldown removal during capture, eliminating a rather tedious second step.

Thanks Anton,
I am looking into the HV 20 primarily to work with a 35mm adapter as a smaller package
btw: I love your Max patches

Thanks! :) Im waiting to get a 35mm adaptor too. Im seriously pumped - the HV20 is the best bang for the buck ive gotten in a LONG time - I highly recommend it even without the lens adaptor!

Great
Can't wait to see the results !

So far, all you can do is download it as 108060i HDV and reverse telecine in Compressor, that is without resorting to third-party plug-ins.

great thanks to all - sorry for hijacking this thread

HV20's 24f don't have tags to the frames, so it does not work the same as G1/A1/H1 when you capture in FCP.

Thanks Kaku, I knew the tags werent there, but I guess that means there are no magic faeries to help me out and save some time. Back to compressor it is.

Thanks for the clarification.

No problem, on top of that, there seems to be no automatic indication that FCP captures 24f clip from HV20, so in order to cancel the pulldown frames, you have to figure out the certain frame and certain format pattern in the Cinematools.

Hrm, perhaps its because I got little sleep last night or am misunderstanding, but doesnt compressor find the cadence for you? At least in my tests it seems to be just fine if you select "de-interlace -> reverse telecine" option, things come out "ok" visually.

Thanks.

i know cineform is one of the only pieces of software thatll do a realtime inverse telecine during capture for the hv20 on pc and they have a mac version. havent used the mac version much, though.

Anton,

Didn't realize that in the compressor. I will take a look, thanks.

it's only in the latest version of compressor (FCS 2), it's hidden under the frame controls>(de)interlacing options, the last one is reverse telecine.

Its really not to bad with the new version of compressor... I just made a dropplet app that removes the pull down. So once I capture (1080i60), a couple key strokes with quicksilver and away it goes... Really painless.

These are the hoops you have to jump through when you try and make a sub 1k camera play with the big boys... :) worth it, IMHO

Scott, what do you mean by droplet app? Is it an automator thing? It would be awesome if you could explain how you have that setup. Thanks!

You can create dropplets in the new compressor that are little application.

I'm headed out for the moment, but when I get back tonight I will do a screen cast capturing my workflow... hopefully it will help clear a lot of things up, stay tuned.

Ok, as promised... here is a video detailing my PF24 work flow:

http://thatsawesome.tv/PF24.mov (11:37 - 214mb)

Disclaimers:

- I am not a professional, so some things may not be 100% accurate
- I did this on the fly, please don't mind my ramblings

Scott,
Thanks very much for that tutorial about Droplets in Compressor.
It gives me a 4x3 rather than 16x9 (maybe I need to go back and re watch)
so I set the size to be 1920x1080 and it looks great, but
I notice that diagonal lines are very jagged
and the file size is 5x more (this happens with pro Res)
I am almost tempted to use the deinterlace in Qt Pro (under Window>Movie Properties...Video)
It's immediate & you lose a bit of Res perhaps, but same file size

But you have me thinking,
There has to be a better way with Compressor
Thanks again for the info,
J

It only looks 4x3 in Quicktime, because Quicktime is not correcting for the pixel aspect ratio. If you bring the Pro Res footage into a 1080p23.98 sequence in FCP 6.0 you will see a 1920x1080 frame.

Remember HDVs native resolution in SQUARE pixels is 1440x1080, which has an aspect ratio of 1.333333, or, 4x3.

Final Cut will automatically do pixel aspect ratio conversion to the proper 16x9 aspect ratio. Similarly, within Quicktime -> Item Properties -> video Track -> View Options there is a 'view at' where you can set the "display" pixels. This will basically do what FCP does with the pixel aspect ratio conversion.

So rather than making up 480 pixels that arent there (1920 - 1440) within compressor and bloating your Pro Res output by that amount, keep the same number of pixels and change the pixel aspect ratio in Compressor to HDV, o redit the final output movie to be 1920x1080 VIEWABLE within the item properties for the video track.

HTH.

John, this could also be why you are getting jagged edges...?

I see no jagged edges in my footage, but I am keeping the pixel aspect as HDV.

its been awhile, bit I had done some testing with outputing back to 1440x1080 vs outputing to 1920x1080, and the file sizes were more or less identical, sharpness didnt seem to be different either

Yes this is it perfectly !
I looks great - beautiful workflow
Thanks,
J

were you changing the pixel aspect ratio or actually changing the pixel count? Im not sure what would happen, as I haven't tried it :) :)

Thank you! That screencast was great!

Thank you, Scott, the tutorial is very helpful. I have one newbie question:

Do you apply this workflow to individual clips before editing or can you build your entire movie and then go through the reverse telecine process?

It seems like it would be burdensome to do this with individual clips and then edit.

Bob

Hmmm... interesting question. I don't think I have a good answer.

I always just convert all my captured footage first and then do my edits in a 24P timeline. My guess (I'm just thinking out loud) is that if you were working in a 1080i60 timeline with the captured footage, did all your cuts, etc... and then converted your footage, I think your edits would be messed up because your removing (or combining?) those interlaced frames. I really don't know... I would think you would want to work in a timeline the same as your output.

Were you talking about exporting a ref movie and then running that through compressor to remove the pulldown?

I'll have to try this and see, or maybe someone with more experience can clear this up.

I set the size to 1920 and the pixel aspect to square.

As far as editing before removing pulldown, you can do it either way, but it depends on your final format. A DVD will look fine, but a web movie will not.

You can remove pulldown with compressor as your outputing to m2v, this works pretty well, but again on a regular DVD/tv you can't tell

Check out:
http://file.meyersproduction.com/testdvd.img

That image has 3 versions of the same 24p footage, it was all edited in a 60i timeline and the track labeled "24p" had inverse telecine run on output to mpeg2.

Personally I remove pulldown prior to any editing since most work is primarily web based.

Potentially a better workflow is to do a very rough cut, and then use the media manager to to move only used clips and then remove pulldown on the resulting files and edit that (media manager will break each edit into a self contained clip)

That clip actually looked the best on my Mac with DVD Player app.

I tired briefly on Compressor 2's reverse telecine and it seems to work well, I'm posting a comparison between Cinematool's and Compressor 2's reverse telecine out from the same footage.

Here's the link.
http://www.onebikeguy.com/KakugyoBlo...t/Podcast.html

I think you must remove it before editing. Say you have 2 clips which are being faded from one to the other, with different cadences, how is the output going to know which cadence is correct? The cadence of one will be lost and when pulldown is done on the final output it will be incorrect.

Scott- Thanks!
 

Scott,
Thanks for posting the video.
It's made my life much easier.
Now I can focus on shooting, capturing, waiting... waiting... cutting.

As for the edit before/after... It's a toss up based on your conversion speed. If you have 5 hours of raw tapes captured, does it make sense to convert 5 hours when the roughcut would be an hour for a final edit of 20 minutes or less?

I would follow the media manage workflow mentioned above.
The caveat is that you offline HDV 60i, online with ProRes 24P. You'd capture your footage; do a rough cut with edits only (no transitions, effects, or alterations to footage). Once you're cut down to the "selects," then media manage to a new folder, and run that through compressor. Reimport the 24P footage back into the media managed timeline, and THEN start your online.

To the best of my knowledge, Media Manager does NOT Recompress frames. It duplicates the section of the file itself. The upside to the HDV rough cut is that you can back this up to HDV tape (print to tape) and connect your camera to an HDMI Widescreen TV and see a fairly good rendition of what's to come. (The print to tape adds a level of compression to the HDV, but its a rough cut.) Saves time on building a DVD for a rough cut... Most HDTVs are 1080i anyhow, so you're not losing.

FINAL THING: 24Frame is not 24P. The HV20 is 24P. Other, more expensive Canons are Frame. Use what you got, but I'm for true progressive, not frame. HV20 is a must have due to this.