Editing D90 footage

mattsand said:
the plugin checks that the output is 1280x720 and does nothing if it isn't. in preview mode it is, since everything is rendered as square pixels. you can try modifying the script. it's easier than you think. just replace all 1280's with 960's and 640's with 480's. or use a square pixel codec, although i guess there are none that are realtime in 4.5...

btw why rt safe? unlimited is better if you're fine with a frame drop or two.

/matt
Click to expand...

So, what I gather is that DVCPRO HD 720p is not a square pixel codec? There are no other options? Time to see about upgrading, although that's bound to open other cans of worms.

I'm fine with changing the script code if I need to, but are you saying that it won't work when it comes time to render for output/ make a master?

I was using RT safe as an example. I usually use unlimited.
 
Try this streamclip - fcp workflow. Got this info from chris f. I'm on Pal so I ignored the stuff in brackets...

"Batch Process all of the video files with Streamclip using this setting: http://michaelmistretta.com/2008/edi...-in-final-cut/

(Then open those new quicktime files with Cinema Tools and convert them to 23.98fps)

And then import them into a Final Cut project with timeline set at (23.98 and) the same Apple Intermediate Codec setting from the above link."
 
you're right, i forgot about aic, but i'm not sure if it's a realtime codec in 4.5 or if it's even included or has to be installed via imovie? i also don't quite see why you'd want to go to 23.98 since fcp supports 24? better to convert on export if you need ntsc compatibility?

anyway, another solution i just remembered is to convert to photo jpeg then hack the rt enabler to enable rt for resolutions over 320x240, google it. that should give you great rt performance with a high quality codec. keep the quality over 75% and it will look as good as the original.

/matt
 
gdrifter said:
Try this streamclip - fcp workflow. Got this info from chris f. I'm on Pal so I ignored the stuff in brackets...

"Batch Process all of the video files with Streamclip using this setting: http://michaelmistretta.com/2008/edi...-in-final-cut/

(Then open those new quicktime files with Cinema Tools and convert them to 23.98fps)

And then import them into a Final Cut project with timeline set at (23.98 and) the same Apple Intermediate Codec setting from the above link."
Click to expand...


I tried this and it seems to work well using the Apple Intermediate Codec. I put the quality up to 100% and converted several files, put on CC and the d90 rescaler all in real time, and made a dvd. It looks great, and I don't see hardly any difference in the results using a ProRes workflow on the same clip. When I import the clips into FCP it brings up a message that "Apple intermediate Codec is not fully supported in FCP and they recommend using something else..." But hey, it works better than anything else, so far.

My D90 is shipping today, so that will be the true test when I get a big project going.

Also, your right Matt, I waited to convert anything to 23.976 till it got to compressor for the dvd ver, although I'm curious how it will sync when I start shooting double system with sound recorded on the laptop. Logic has the ability to run at 24.000, so probably i'll choose that, or 30.000 might be better and still sync.
 
I just got my d90 and have started to play around with the footage in Final Cut Pro 6. FCP 6 seems to handle editing fairly well without transcoding. I take it that the problems reported by others with Final Cut refer to earlier FCP versions.

Are others using FCP 6 natively for editing? My understanding is that the d90 codec has a 4:2:2 color space, which should make it good for color correcting and FX.

What are other's more experienced experience??

(I hate to reinvent the wheel)
 
The color subsampling is not the same as the color space in the true sense of the word. A 4:2:0 codec has just as many colors, just at a slightly lower res. It does matter for some effects like keying but not so much for primary color correction. Besides prores is a 4:2:2 codec too and it has 4:4:4 resampling built in so in some ways it could be considered better. I like to stay in the original format because it's, well, the original format. i can create an edl and go to another system losslessly, maybe a windows system that might have a better mjpg codec, maybe with deblocking? Not that i often will but i like maybes. :-) /matt
 
I chose the wrong word--color sampling is 4:2:2 as I understand it with the d90 codec. This is better than HDV (and miniDV, and I think DVCPRO) which does 4:2:0. In other words it is a reasonable codec for color correction and FX.

So it sounds like you do stay in the d90 codec for all your editing/CC/FX needs.

I will give that a try--certainly the easiest route!
 
Robert Altman said:
I chose the wrong word--color sampling is 4:2:2 as I understand it with the d90 codec. This is better than HDV (and miniDV, and I think DVCPRO) which does 4:2:0. In other words it is a reasonable codec for color correction and FX.
Click to expand...
the choice of word is ok, i knew what you meant, but what i'm saying is that this matters very little if at all when it comes to color correction. you got just as many colors in all of the above codecs with the same 8-bit latitude. the subsampling only determines the resolution of the color channel. while low resolution is a bad thing in itself it doesn't make the codec any better or worse suited for color correction.

/matt
 
mattsand said:
the choice of word is ok, i knew what you meant, but what i'm saying is that this matters very little if at all when it comes to color correction. you got just as many colors in all of the above codecs with the same 8-bit latitude. the subsampling only determines the resolution of the color channel. while low resolution is a bad thing in itself it doesn't make the codec any better or worse suited for color correction.

/matt
Click to expand...

No, i think Robert is right, a 4:2:2 is better for working color in post than a 4:2:0. A 4:2:0 is higher compressed color information; and thus more likely to color artifacts when working on color. A 4:2:2 is much better for chroma keying or other color work because you have more color information to work with.

HDV is in 4:2:0. DigiBeta is in 4:2:2; this is why broadcasters still use Digibeta.
 
artforme said:
4:2:0 is higher compressed color information
Click to expand...
no, it's just lower res, which again is bad in itself but not more so when it comes to grading. if you downsample an image to 4:2:0 then grade it you get the exact same result as if you grade it first, then downsample. you're perfectly right that 4:2:0 is more prone to color artifacts, but again that's true whether you grade it or not. bottom line: 4:2:0 is not as good as 4:2:2 but it's equally worse no matter how much you grade it.

/matt
 
You must log in or register to reply here.
Back
Top