I've got a feature coming up and the producers want to post in Finland (for various financial reasons). We are shooting in the US. Finland says the material will have to be converted to 25 fps from 23.98. First they asked me if I would shoot at 25psf and I said no. The producers are being gracious at this point and saying that if I say no, they won't go.
The HD final will be produced in Finland. So then I guess it would have to be converted back to 23.98 for the US. A film out is planned. What are the issues with all this? If I can deliver a few good tech reasons not to do this then I think it's a done deal and we stay home. However, if it's easily doable and the final product will not suffer, lets take a vacation to Finland.
Ian Ellis DP
> Converted back to 23.98 for the US. A film out is planned.
No real issues, actually very straight forward and concise work flow - if you shoot @ 25 that is...
Have done this in HD recently, and the next feature I am preparing for (to shoot over here) is going to 25 for very good reasons, mainly that the funding is coming out of Europe, and there is no good reason other than fear of the unknown to not do it, the biggest problem i had was with LA based post house who were stone cold afraid of 25Fps....got all of them in a room and went over it step by step, took a full day, but at the end they all agreed that they did have the gear, and there was no reason why not to do it, and loads of good reasons to go to 25.....contact me off list with specific questions? will try my best to answer them.
Hd & Film Vfx guy
Vancouver, Canada & Shanghai, China
> Why not shoot 25fps?
Kenos, practical fluor’s and non square wave HMIs will flicker. Film out is in
Ian Ellis DP
> Kenos, practical flos and non square wave HMIs will flicker.
Still have non-square wave Hmi's around? thought they had gone away with my Cp16.....and there is work around for the Keno's, 50Hz bulbs.....
It can be done, and is done on a regular basis
HD & Film Vfx guy
Vancouver, Canada & Shanghai, China
Ian Ellis / Chris Ellis wrote :> Finland says the material will have to be converted to 25 fps from 23.98... Let me see if I have this straight...
>If the reason for conversion to 25p is for HD online
Yea - the caveat from me is that the stuff I do is often funded primarily out of the UK, so 25Fps is in the contracts, if as Lucas pointed out, the main distribution is USA, the 23.9 is your friend, if it is Europe then 25 would be the answer, neither is a issue really... both are possible, and I second his advice to run far and fast from anyone who cannot switch frame rates in HD post easily unless they have an old Sony switcher that cannot change frame rates.. in that case another reason to run.....
HD & Film Vfx guy
Vancouver, Canada & Shanghai, China
Why not shoot 25fps?
A temporal shift with pitch correction on the audio side should keep the film out a possibility. A minor compromise for the film-out (which may never happen). The long-term is a video release which would probably be in PAL, so why not?
Dailies could be an issue if you don't have a PAL downconverter available during production.
Personally I'd run a mile from anywhere that couldn't work at HD or 2Kat whatever speed was wanted.
As has been said earlier it may be the offline that's a problem but equally, that is dealable.
Geoff Boyle FBKS
Director of Photography
You are shooting a show in the US for delivery and distribution in the US, and the Finns are asking for everything to be converted to 25fps.
Sometimes your job is to save the ******* idiots from themselves...Speak the truth, and if they don't like it, walk away.
Take The Gig
Let them downconvert to PAL.
Master from PAL EDL list back to 23.98 in Online (Just like you would do if it was NTSC) Record Sound on the HD Tapes there is no need for any pitch conversion.
A Second of picture and sync sound in 23.98 is the same as a second of PAL and Sync Sound.
Just Like a Second of 30FPS of Sound is the same length of time as a second of 144FPS of Sound is the same length as 23.98.
A Second is a Second is a Second no matter how fast or slow one format is running. Test it you'll see,...uh Hear.
Do not attempt to or be persuaded to record downconversion on Set. Do not bother with any Sound recording device other than the HD CAM Camera. Do Not use Free Running Timecode.
Fight for everything I am recommending and you will be fine. Remember to ask yourself "If I was shooting Film It would be possible then and I know its possible here". If they cant handle it I would take Geoff’s advice.
Run to the next available competent house that Can.
B. Sean Fairburn
Director of Photography
Sean is mostly correct, but here are a few additional comments:
"Let them downconvert to PAL Master from PAL EDL list back to 23.98 in online (Just like you would do if it was NTSC)".
Although this will work, there are better ways to do it. Offline in a film project on a Film Composer, and the lists coming from offline will always be at 24fps. If this is done, there will be no need to ever deal with a 25fps list conforming back to 24fps in the online bay.
Also, most systems have better conform methods than EDL. Smoke and Fire offer OMF conform. DS|HD offers AAF and AFE conform. An AAF conform on a DS is infinitely better than an EDL, as it brings over a ton of additional information about the clips, bins, effects, titles, and project.
"A Second of picture and sync sound in 23.98 is the same as a second of PAL
and Sync Sound. ... A Second is a Second is a Second".
In camera, a Second is a Second. In HD post, a Second is rarely a Second, and could really burn you in this particular situation.
Scenario #1 :
1) Shoot 23.98
2) Downconvert and offline to Slow PAL.
3) Audio mix to slow PAL masters.
The final mix will now be wildly out of sync with the original masters in a 23.98 online.
Scenario #2 :
1) Shoot 23.98, with audio recorded on F900
2) Downconvert and offline on Film Composer with PAL method1
3) OMF to ProTools for final audio.
Now you are screwed, because audio has been slowed down in offline because of your choice to digitise with method1.
Scenario #3 :
1) Shoot 24
2) Downconvert and offline Film Composer PAL method2
3) OMF to ProTools
4) ProTools mix assuming pulldown at 47.952 (instead of 48)
5) Sync mix in online.
The mix will now be 1/1000 off, and will drift.
Bottom line -- either hire a post coordinator who is *very* familiar with HD workflows and can navigate this stuff. Or, get online and audio post involved in the process at a very early stage, so formats and data transfer methods can be ironed out and finalized before you get screwed.
Hates Dealing With Sample Rate Mismatches
Either way Its not Ian’s problem how post Sound gets the Job done nor is it his responsibility to make life effortless for them. (I like to be considerate but not unnecessarily)
Your Off-line and DS solutions are very good ones and should be looked at strongly.
I Find it Odd that NTSC at 29.97Psf 3-2 pull downconverted from 23.98 works like a charm but PAL at 25 frame downconverted from 23.98 can't lock sound and make it work.
Something doesn't seem right about that.
Ian Have them look at the whole picture and all its parts looking at real proven answers not speculation then make decisions that serve the overall good of your Picture.
But First Take The Gig
>Either way Its not Ian’s problem how post Sound gets the Job done
[Lucas] When the Pro Tools guys are saying "this was a dumb way to do it" and when sync sound in DaVinci doesn't work and everybody is asking "why was it done this way?" then it definitely becomes his problem. I've been in more than one HD Online where the Producer rips the production team apart because they didn't think through the entire process. Besides, he wouldn't be asking about post solutions if it weren't at least partially his problem.
"I Find it odd that NTSC at 29.97Psf 3-2 pull downconverted from 23.98 works like a charm but PAL at 25 frame downconverted from 23.98 can't lock sound and make it work. Something doesn't seem right about that."
[Lucas] Even in NTSC, you have to worry about whether audio is recorded with or without pulldown. With a DA98, Is it 48K, or 47.952K? In a dualsystem, is Nagra locked to 60Hz, or 59.94? It can all be worked around in post, but it doesn't always work like a charm.
PAL downconverted can work -- you just have to be careful about it. There are two different ways to downconvert PAL from 23.98 and two different methods to offline it on a Film Composer. Out of 4 possible combinations, only 2 work. So, there is a straight 50/50 chance that audio will be SNAFU unless these considerations are taken into account.
Your comments are great and very helpful.
But as you know it can go both ways "Its a Dumb Way to do it" also comes from lazy Post people or those that like to say We have always done it this way. Most are knowledgeable and informative like yourself and have real answers that are considerate of the big picture. Unlike the Comment Shoot in 25P HD (because it will be easier for sound to do its job)
You know as well as I do that Ian will be Praised and Cursed and Praised for whatever decision he makes and undoubtedly he will be blamed for decisions that a Producer made over his objections. Happens more than it should.
Lucas. You Are Right…But 2 things are still true here. Ian Can't do anything about it other than shoot 23.98 (Really he has no other viable choice) and whatever Post must do is out of his hands even if he wanted to its not his Job.
So I thank you for educating Me on great Post solutions and I know that Ian will make great looking pictures with his Camera regardless of how
sound deals with it.
B. Sean Fairburn
Lucas Wilson wrote:
> When the ProTools guys are saying "this was a dumb way to do it"
I think this happened on a recent 24p feature I did. what may cause it to unsync on DaVinci ?
Ours was simple 23.98p HDCam shoot/post and downconversions to DVCam for Avid. I thought it was backup/temp sound on HDCam-originals (mix not yet complete), but in retrospect must've been temp sound laid in from Avid that was out-of-sync on DaVinci and even more so in a re-assembly online of all 5 filmout reels.
LA based DP
Mark Doering-Powell wrote
>I think this happened on a recent 24p feature I did. what may cause it to unsync on >DaVinci ?
The easiest place for this to go wrong is in the digital audio transfer from audio post to Online/DaVicni. The reason is that Media Composer, up until the latest versions of software, did not have a 23.976p project setting and only had a 24p project setting.
Anybody not interested in deep geek online stuff can turn the channel...now.
Film Composer was always built for Film. Film shoots at 24fps. Adding pulldown to 24fps results in 30fps. Lovely, except that NTSC is not 30fps, but rather 29.97fps.
So in NTSC-land, film is slowed down in telecine to 23.976fps, which equals 29.97fps when pulldown is added. Audio is also slowed down accordingly, so that what is recorded at 48K really becomes 47.952K.
In a 24p project in Film Composer, video is digitized at NTSC, and pulldown is removed, leaving 23.976fps video. Then, the video is sped up by .001% to undo the telecine slowdown and play back true 24fps video. Audio is also sped up accordingly.
Audio typically goes to ProTools from offline via a digital audio file called an OMF. This is a fine process for film, and works well.
The problem comes with 23.976fps video. In 23.976fps video, the ‘origination’ format is 23.976fps. There is no telecine slowdown. So, when a Film Composer digitizes the video and audio and speeds it up by .001%, it is inaccurately speeding up the original audio. So when a mixed file gets to a 23.976fps HD online, audio has been sped up by .001%, and will drift.
It is exclusively for this reason that the newest versions of Media Composer have both a 24p and a 23.976p project setting.
Lucas brings up an important point in that the entire project including the post production sound workflow is critical to streamlining the process. As a former on line editor I can't begin to tell you the number of horror stories that ended up in the edit suite and were resolved by the producer or director uttering "I will never hire that crew again." Simple issues that could have been easily resolved in the field were never addressed until it was too late.
It's not a matter of post production personnel being lazy rather it is about working and interacting with each production department down the chain after the field production is done to insure your work will comply with an established post production workflow. The bottom line is protect your future employment and reputation by delivering materials that minimizes problems instead of creating issues which "you" will be the ultimately held responsible for creating the problem in the first place. I am not saying that we have to bend over backwards or "forward for some" to satisfy all the post production departments instead "work within reason" and when in doubt ask the right questions to the right people.
It's amazing that we work in the "communication business" and the least amount of communications often occurs between ourselves and others in the
>Audio is also slowed down accordingly, so that what is recorded at 48K really >becomes 47.952K.
Maybe I'm being a bit dense here, or maybe my memory is failing me. First of all, it seems to me that everything you're saying only applies if sound is digitized separately from video, i.e., sound is not brought in from the daily video masters. Often done in features, but it should be pointed out nevertheless. Second, I seem to recall that the Avid audio interface had a pulldown switch, which you set to "on" if the audio has already been pulled down (as it is if you're digitizing from a synched daily video master) and "off" if it hasn't (as it is if you're digitizing directly from a Nagra or a DAT that's set to sync to itself). In the case of the former, both the video and audio are "sped up" for 24fps editing and stay in sync.
In the case of the latter, only the video is "sped up" and the audio remains the same, thus "pulling up" the speed of the picture to match the sound. This is all a "layman's" description, as I'm aware that all this really has to do with is the sampling rate of the sound. Given that, couldn't you simply digitize the 29.97 sound with the pulldown switch "on" and cheat the system into not doing a separate pullup? As for the OMF files, wouldn't the show be dubbed to a 29.97 timebase (in most cases in the US) anyway, and, if so, wouldn't any sync problems be obvious during the dub? Or, more directly, during sound editorial?
BTW, I'm not disputing the usefulness of a specific 23.98p project setting, but I don't see where anything other than sound pulldown is affected - so it doesn't really "do" anything you couldn't do before, provided my previous paragraph is correct. Of course, I could be completely misunderstanding the entire pulldown switch concept (many others have misunderstood it long before me.
IATSE Local 600
Michael D. Most
>First of all, it seems to me that everything you're saying only applies if sound is >digitized separately from video
Could be I'm not explaining it well... the problem actually applies specifically to audio digitized *with* video.
>Second, I seem to recall that the Avid audio interface had a pulldown switch
This is all correct.
>Given that, couldn't you simply digitize the 29.97 sound with the pulldown switch 'on'
If the pulldown switch is set at .99 (on,) then audio is being sped up. "Pulldown On" tells Composer that audio was slowed down in telecine, and should be sped up inside the software. That is also why you cannot digitise digital audio when the pulldown switch is on -- Composer is resampling incoming analog audio at 47.952 and playing it back at 48 to achieve the speedup. Digitize with pulldown set at 1.0, and the system does not do a separate pullup, but then there are two digitize passes -- one for video, one for audio. But there is no way, in a 24p Avid project, to properly digitize audio and video concurrently that come from a 23.976p-originated material.
>As for the OMF files, wouldn't the show be dubbed to a 29.97 timebase (in most >cases in the US)
No, because audio and video are slowed back down on playback. This only occurs when an OMF file is transferred, because the OMF is not slowed back down. This entire problem targets a very specific situation -- transferring OMF files from a 24p Media Composer project to a 23.976p online with 23.976p HD - originated material.
I've explained this problem before, and it's really hard for me to do it without a whiteboard and an ability to draw it out. Sigh!
>BTW, I'm not disputing the usefulness of a specific 23.98p project setting
You're correct in that the entire purpose of the 23.98p project setting is to address audio pulldown issues. But it *does* so something you couldn't do before. It allows digital audio to be treated properly when working with NTSC downconverts from a 23.976p HD-originated program.
But think about this... you are an experienced post professional and VFX supervisor. You're used to dealing with problems like these, and it's confusing for *you*. Think about how screwy it gets with the average offline editor, who is a good storyteller, but usually not so strong with deep geek details!
Hates the Pulldown Switch
For more information...
One of the best resources for anything relating to Avid and Film is Alan Stewart's www.zerocut.com
for more details. He explains it better than I do.
Fantastic responses. Lots of info to work with when I talk with this post house in Helsinki on Tues. When do you guys sleep?
Ian Ellis DP
Ian Ellis / Chris Ellis wrote ;>When do you guys sleep?
Another great resource is Michael Phillips site 24p.com
Vincent G. Hogan
Cineworks Digital Studios, Inc
Lucas Wilson wrote :
>This entire problem targets a very specific situation transferring OMF files from a 24p >Media Composer project to a 23.976p online with 23.976p HD originated material.
Thank you for the explanation, it's very informative. One question though: why would you send OMF files to an HD online if the OMF's were created from downconverted videotape input (i.e., offline cassettes)?
In my experience, sound directly from the Avid (offline) is only really useful if it's digitized from a very high quality, low loss source - like the original DAT, a DAT clone, a Deva, a DA-88 (if you're multicam), or a time code Nagra. Otherwise, it's an "O" track only, and can either be sent to the online via digital cut tape output, or assembled from original tape sources - and the sync problem doesn't exist.
I must be missing something here, unless DVCam (a common source format
for Avid editorial these days) is considered an acceptable audio source - or unless more people than I thought are digitizing directly from HDCam machines (via live downconversion).
IATSE Local 600
I think I'm getting a headache.
Dan "just shoot film!" Drasin
Marin County, CA
> and there is work arounds for the Keno's, 50Hz bulbs.....
I'm sorry, I didn't realize that you needed different bulbs for different power. I thought it was the ballast that took house power, and those might be sensitive to line frequency. I thought that the bulbs all flickered at something like 250,000 hz. in which case I'm not sure that there would be a difference between shooting 24 (23.976), and 25.
Am I mistaken?
Cinematographer - Gladstone Films
Cinematography Mailing List - East Coast List Administrator
Better off Broadcast (B.O.B.)
New York, U.S.A.
No you are not.
Most Kino Ballasts flicker at about 4 Mhz. The lamps are the same everywhere.
It is my understanding that Kino ballasts are 25Khz - suitable for any frame rate. Any magnetic ballast will flicker at the exact frequency of the incoming line. Occasionally, there are electronic ballasts in upscale environments like office buildings they MAY have a higher frequency that will allow various frame rates.
The bulbs have absolutely nothing to do with the frequency.