|
|||||||||
|
Thread Tools | Search this Thread |
October 29th, 2010, 03:34 PM | #1 |
Trustee
Join Date: Oct 2009
Location: Central Coast Australia
Posts: 1,046
|
Color Correction?
Hey to all,
Can anyone explain a simple workflow for color correcting in Vegas? I've read the Glenn Chan articles, and to be honest, I'm left a little confused. ATM I'm converting everything to neoscene, Project settings- 32bit floating point (video levels), which gives you gamma of 2.222 by default. Editing, color correcting with the RGB parade, then saving as a .veg file and nesting that in a new project of 32bit floating point (full range) with a gamma of 1(linear). It gives me really clean looking video, but the blacks seem a bit more crushed than they should (it is uniform though, so not too hard on the eye). Should I have an srgb to crgb (or the other way round?) on the original footage while col correcting? Does anyone have a step by step guide?
__________________
http://vimeo.com/livewebvideo |
October 29th, 2010, 04:16 PM | #2 |
Regular Crew
Join Date: Jul 2010
Location: London, ON
Posts: 175
|
If you haven't changed the default for the Cineform codec, it decodes to Studio RGB, regardless of the full or video range 32bit setting.
So, yes. While colour correcting/grading, you will need to have a preview device using sRGB, or convert it. The easiest/safest is to use a proper DV preview device, but a second windows monitor works OK was well, if you can overlook the interlacing issues. BUT you have to set the secondary monitor into full-screen mode, and you have to make sure you've checked the 'Use color management' and then the 'Use Studio RGB' options. Then, when in full screen you'll see it correctly on the second monitor. The sRGB->cRGB filter is the last choice, as it is a output FX, not just a preview FX; all output will be converted, even if you're rendering. So if you forget it on and your rendering codec expects sRGB (as Cineform does by default) you will get crushed shadows/highlights. Somewhere near the bottom, of one of the pages, Glenn gives a pretty good workflow description for the various modes, but that's the gist of it. With Cineform, it gets a bit easier; some of the Sony codecs switch between sRGB and cRGB depending on the full or video range setting. Hope this helps,
__________________
CraigL |
October 29th, 2010, 04:39 PM | #3 |
Inner Circle
Join Date: Nov 2005
Location: Elk Grove CA
Posts: 6,838
|
Add the studio RGB to Computer RGB filter at the preview window, for a very close approximation to what you will see with the finished product.
When render time comes, uncheck that same filter.
__________________
Chris J. Barcellos |
October 29th, 2010, 06:36 PM | #4 |
Trustee
Join Date: Oct 2009
Location: Central Coast Australia
Posts: 1,046
|
thanks heaps for that, after all the the long winded reading ive done, now i've got something i can use.
On the Cineform box Ive ticked the upgrade to 422 color space, does that matter? So I can see what Im doing now in the preview monitor, now the big question is do i go back and re do the col correction on the 2/3 of the wedding ive already done, I can feel a stress headache coming on. On that note, when col correcting, what do you guys do, just each camera in the project bin, or each cut in the timeline? I did each cut in the timeline, and it prob took a good 3-4 hours to do a 10 min sequence, is this normal?
__________________
http://vimeo.com/livewebvideo |
October 29th, 2010, 08:35 PM | #5 |
Regular Crew
Join Date: Jul 2010
Location: London, ON
Posts: 175
|
The 422 is a colour subsampling parameter, not a colour space. In 422, two pixels in a row are averaged to produce one chroma value per two luma pixels. So your colour resolution is much better than 411 for eg, which uses four pixels to average one chroma value per four luma pixels, or 420 which is also 4 pixels averaged, except a two by two square. However, when converted from YCbCr to RGB, the studio or computer RGB question still exists.
Doing the colour correcting is never a quick job. If you can't shoot a card/clapper on each take, you are likely in for some work. I would think doing one camera take per location/room/lighting condition at a time is probably the simplest. At least that's what I'd try to do. As the room changes, or the lighting drastically, tuning might be required. But if you've shot and/or set your camera with a card, it should (mostly) be fairly close, or fairly easily set-able with the card. I should imagine though, that it is far, far too easy to get caught up and fine-tune the correction way beyond what it actually required. We can tend to be way too over critical on ourselves sometimes. Not much help up there I guess, but I hope there's something. Good luck,
__________________
CraigL |
October 29th, 2010, 09:07 PM | #6 |
Trustee
Join Date: Oct 2009
Location: Central Coast Australia
Posts: 1,046
|
thanks Craig,
Its looking like this is a case of "On the next project........." A few people have seen the already wrongly col corrected MP4's and the general consensus is that it looks beautiful . You're right about being to critical, I can see how the blacks are a bit crushed in the dinner suits, but no one else cares. This has been great to learn though, Ive learnt more about col correction in the last 2 days than I have in the last year. Thanks again.
__________________
http://vimeo.com/livewebvideo |
October 29th, 2010, 10:36 PM | #7 |
Regular Crew
Join Date: Jul 2010
Location: London, ON
Posts: 175
|
One last thing, although you working in Cineform is a special case (ie. always in sRGB, unless you choose differently) there are differences between Vegas 32bit full and video range (not 32bit vs 64bit =) for some codecs, as well as static images, generated media, etc.
Always refer to Glenn's chart: http://www.glennchan.info/articles/v...s-9-levels.htm To know what range the specific element might be in depending on your mode. In particular, even when you think your video is all in sRGB and adjust appropriately, the images and generated media (like text, etc) are almost certainly in cRGB, and should be adjusted appropriately. For eg, change the pure black on the text from 0:0:0 to 16:16:16 and pure white from 255:255:255 to 235:235:235, or use a cRGB to sRGB filter to bring it back to the same as all the video. The pixel smoothing that goes on will likely look better if the white starts at sRGB white and fades to black that cRGB white, the only reason I suggest changing the white. That chart is a big help. Of course, I'm assuming that Vegas v10 does the same as v9 did, I don't think Glenn has a v10 page up yet.
__________________
CraigL |
October 30th, 2010, 09:43 PM | #8 |
Trustee
Join Date: Oct 2009
Location: Central Coast Australia
Posts: 1,046
|
So one more question,
im doing another project, Cineform 1080p, 32bit floating point (video levels), with the SRGB to CRGB filter on the preview, (tried the external monitor with color management, looked great, but of course it is by default on Best Full, so playback is pretty bad), So I color correct to this and remove the filter on the preview when I render it out for DVD, yeah? So, If I want to render out for the web, Vimeo etc, do I just leave the preview filter on? or Do I save as a veg (with the filter on/off, ? ) and drop it into a new project with full levels and a compositing gamma of 1. You see what I'm getting at, a project that is color corrected one way, and just flicking a switch or two depending on where the final render will be, DVD or web. It would seriously suck if you have to color correct twice.
__________________
http://vimeo.com/livewebvideo |
October 31st, 2010, 11:55 AM | #9 |
Regular Crew
Join Date: Nov 2008
Location: Maidstone, Kent, UK
Posts: 190
|
As far as I am aware, YouTube, Vimeo and various other flash media players (not sure if it's actually the websites or Flash itself that does this) seem to expand sRGB to cRGB automatically, therefore video should be uploaded in sRGB and it'll get played back expanded. If you render out with the filter on and then upload, it'll still get expanded so everything black or white will get crushed.
(with thanks to whoever it was on DVInfo who brought this to my attention, sorry I forget your name) Dave
__________________
www.tubeshooter.co.uk www.youtube.com/ukairscape and www.youtube.com/tubeshootermag |
November 1st, 2010, 01:18 AM | #10 |
Trustee
Join Date: Oct 2009
Location: Central Coast Australia
Posts: 1,046
|
hmmm, dont know about this, Ive been uploading lots of clips of my sons band recorded from TV, they always looked very "soft" and washed out on youtube.
But since learning abit more recently about colour correction (the span of this thread mainly, lol), the last one I put up I applied a SRGB to CRGB filter when rendering the MXF capture to h264 and it looks remarkably better than any other TV upload Ive done, even at a really low bit rate (1-1.5mbps) compare this YouTube - SHORT STACK- Caralyn. with SRGB to CRGB filter.... with this (same bit rate and codec) YouTube - SHORT STACK -LIVE at Sydney Ent Centre recorded the same way, only difference is the filter.
__________________
http://vimeo.com/livewebvideo |
November 1st, 2010, 08:03 AM | #11 |
Regular Crew
Join Date: Jul 2010
Location: London, ON
Posts: 175
|
Careful with some of the filters in linear light mode, sRGB-cRGB in particular doesn't work correctly. Take a look at the SMPTE colour bars (which is cRGB) and convert it to sRGB. Normally, this will make the blacks on the in the plunge bars look correct, the 0 and 16 are both black, and only the dark grey stripe will be visible. But in linear mode, the whole plunge will look black. I think this will be the same for both levels and colour corrector. The secondary monitor using sRGB colour space is not affected by this. Glenn also has info on this:
http://www.glennchan.info/articles/v...t/linlight.htm I had always been under the impression that YouTube et al. did do an sRGB->cRGB conversion. Part of the reason I've never figured out a way to determine if it is true, was because I could never figure out how to record something definitively in cRGB space. I just realized that now with the v10 demo I can, using MPEG-2 in the two 32 bit modes, so I recorded some SMPTE bars in 32 bit full and video range. And guess what? I think the answer to what YouTube does is; it depends. The one I saved in full range, was definitely not converted. The super-black is still black, and the video black is the washed out grey. The video range one was definitely converted, the video black and super black are both real black now, only leaving the grey bar. Very weird... Almost as if they scan it and if they see sufficient values >235 or < 16, they don't convert it, but if they don't see it then they do. Unless the MPEG-2 has some way of indicating colour space? But I didn't think colour space information was included. Need some time to digest this...
__________________
CraigL |
| ||||||
|
|