Michael Gissing
Forum Replies Created
-
Capturing straight off the camera is a potential disaster. The editing codec is one question but the real problem will be the ability to recapture.
Straight off the camera will give you no device control and therefore no timecode if you recapture.
Uncompressed is huge so unless this is a very short project you will rapidly fill expensive RAID drives. I use uncompressed to recapture final edits but never would consider it for editing.
-
So far you have saved 90 gigs which is about $50 worth of hard drive. It probably cost you more in time to transcode than that so if you had asked before, I would have said stay uncompressed, especially if it has to be scaled up to match HD and buy more drive space if that is really a concern.
-
Big Endian is the audio codec (in this case PCM )and is not the video codec. Do you have quicktime player on your PC? If so when you open the file what happens?
If it says it need to download a codec and that fails then you should go back to the person who made the disk and ask them. I can’t possibly guess from the info supplied.
-
[Carsten Orlt] “I would think you cut in 24, do your sound post in 24, export picture from FCP as 24, change Fps in cinematools to 25, speed up your audio by 104% and all should be fine?
to naive? :-)”
Unfortunately yes.
Firstly 23.98 is not 24. Most audio systems can do real 24 but not 23.98. Also we had to cross convert the HD to SD to run with the audio system.
So we went with 29.97 NDF to maintain the real time of the sequence. Audio was imported from an OMFi (this was offlined on an AVID) and the picture came on digi beta.
So far so good as in audio we have a standard sample rate so the OMFi could be imported into a 29.97 sequence and still was in sync. Next tracklay & mix as usual. Then the 23.98 was sped up to 25 for the BBC version. The mix and stems had to be sped up by 4.01%. This figure is not entirely accurate as 23.98 is actually 23.9761ish and getting it right makes a difference in a 90 minute doco.
Not everyone has good software to do this speed up and maintain pitch without introducing audio artefacts. Luckily I do but it is far from typical software in any facility.
So with all that fiddle, the other complication is that I had to switch one of three studios onto a 29.97 video reference to sync audio & picture. All of this is a nuisance to the usual efficient running of a facility, as I had to use a second video reference signal and manually patch it to the machines, bypassing my otherwise efficient signal distribution system.
All this came about because the project required varispeed. Nobody spoke to me before production so I had to solve this issue on a tight turn around relatively low budget program.
The final issue was a reversion for The History Channel where I was presented with a 23.98 EDL to recut a 29.97 sequence. Here Cinema Tools came to the rescue and converted the EDL to 29.97. You may have seen an early Cow post from me about trouble getting Cinema Tools in FCP vers 6 to work. Luckily I have a second system drive with 5.1.4 and its Cinema Tools worked.
Very few facilities would have both the software and technical savvy to make this sort of complicated work around happen without blowing time and budgets.
There must be a way to do decent slo mo in HD without this sort of mess. If anyone out there knows of an alternate to Panasonic Varispeed then I would love to pass that onto the producers to make the next one a bit easier for all concerned.
-
Morning Carsten,
Good link and info. However the final paragraph say “Some additional work might be done for audio conforming”. This is a huge understatement and anyone who is thinking of shooting 23.98 for a 25 frame delivery should talk with their audio post people. It is far from trivial.
It was also interesting to watch the director and editor watching the 25 frame restripe and commenting on just how the 4.01% speed up changed the pace of their edit in a way that they felt was undesirable.
-
An easy option is to install a second boot drive with Leopard, CS3 and a newer quicktime. Choose which application to boot to and you can leave the old drive untouched with 4.5.
Having a second boot drive is easy to manage. Just hold the alt key during boot up and select the drive to actually boot.
-
The Panasonic codec is DVCPro100HD and is designed to be captured over firewire.
To monitor HD it is best to have a card like the Decklink or Kona. Search previous posts also about the Matrox MXO which is an HD monitoring solution.
My limited experience with varicam at 25 fps plus anecdotal stories leads me to be very wary of this format in 25 fps. Perhaps things have been improved recently but I have heard some bad experiences with varicam and 25 frame. It seems that in the 24 frame world, it works as many here will confirm.
I have just finished a varicam job that was shot 23.98 and then the final was sped up to 25 for the BBC. F**ing nightmare for all concerned. I was luckily only doing sound post on that one. I limit my picture post facility to HDV and HDCam.
-
Bit confused about your description. If you are trying to mix a six channel program on FCP then you don’t have surrround panning and buses so you can’t seriously be mixing for 5.1 . When you say Dolby Digital I assume you mean an AC3 encoded file which looks like a stereo interleaved file but if you put that in FCP – well I don’t even know if FCP can deal with it.
If however you mean discreet six channel as aif or wav files, then either Decklink or Kona cards embed multichannel audio in the SDI. You will need to decode this from the SDI and then to an appropriate amplifier.
If you mean something else then please – more info
-
I don’t believe it is uncompressed HD because that would be the biggest file you could make and it wouldn’t play on Macs or PCs without a RAID drive.
Walter is right – H.264 is the best codec. Quicktime 7 for windows is a free download.
-
[Ted Griffis] “The HDV to ProRes on the fly workflow is really cool, its limitation that I can see is the re-batch-capturing those clips if things fail.”
Sorry to harp on Ted but there is no reason why you can’t recapture again frame accurate with that workflow if you have device control via firewire. Why is it a limitation?
My 2 cents worth? Why transcode to ProRes and then after grading etc render again to ProRes. That is one more transcoding than capture as HDV native and render the final to ProRes. More space, less effort on your drives and one less transcode. If someone can give me a reason why my logic is wrong then please – fire away.