Kevin Cannon
Forum Replies Created
-
Agree with Paul – even if you find GPU and CPU performance is acceptable on the old mac pro, you’ll have to use PCI slots for e-sata and USB3, and really having no option for thunderbolt drives – we’re starting to see a majority of USB3/thunderbolt only drives come in. An iMac (plus the cheap TB>Firewire800 and USB3>e-sata adapters) should give you all the fast connections you need. Same should be true of the new mac pro.
Cheers,
KC
-
Not having identical timecode between the editorial media and camera originals is a tough one. Did they rename the ProRes in a way that you can tell which editorial file corresponds to which camera original file?
We’ve solved a similar issue in the past by doing this: make a list of the editorial media and the starting timecode of each clip. Add the camera media to the media pool and sort it in the same order as the list. Then go down and go into “clip attributes” and change the “current frame timecode (while on the first frame)” to match the starting timecode of the editorial media. Resolve will use this new timecode as you conform. Hopefully you have very few files total.
Cheers,
KC
-
Hi Ryan,
Definitely don’t let the websites do the compression – we’ve had good results with this formula:
1. Render the material out as ProRes (DNX 175 should also work) and check that you are happy with that render by reviewing it in Resolve on your video display (we view in rec.709 gamma 2.2)
2. Get the maximum quality upload specs from the video site, I strongly recommend Vimeo Plus/Pro over youtube: https://vimeo.com/help/compression
3. Compress that ProRes using compressor or similar program (I’ve been satisfied with compressor, some use handbrake) into H264 using those specs, then bring the resulting h264 into Resolve and compare against the original.
4. If you’re happy with the h.264, upload it and it should not recompress (or at least, only to create different resolution and bandwidth versions).
Definitely consider the presentation on the website. There will be a difference in perception of the blacks viewing against a black, gray, or white background. Also, one thing I like about Vimeo is the ability to limit the interface and customize the colors on the embed.
A lot of people are viewing now on devices with adjustable brightness, so if you are worried about crushed blacks, you might want to have some color bars or other reference that you can check displays with.
Cheers,
KC
-
Hi Andrew,
It should feel more familiar if you use the Alexa LUT as the first node – all the corrections made downstream from there will not be going through the LUT and you won’t have that problem.
Even better might be to use it in the second or third node – then you can make corrections before and after the LUT. Corrections made before the LUT will have a nice soft feel, corrections after the LUT will feel like grading rec.709 material.
Cheers,
KC
-
Hi Sam,
Since the data vs. video levels setting in the video monitoring section is strictly for the video output signal (and doesn’t actually change the values being calculated by Resolve) the trick is to always have it set to whichever is appropriate for the display you are using.
So if you are grading on a rec. 709 projector that is set up for full range values, you will want Resolve set to full range. If you then move to a broadcast display that expects video/legal levels, switch to to video/legal, and your black and white levels should feel the same, excepting the inherent difference in display technologies and the sensation of being in a room with lighting. Then you can do a pass to adjust for those differences (in track mode if you feel it works, or on a more scene-to-scene or shot-to-shot basis).
But if you are grading on a P3 gamut, then your DCP workflow might also include a change between P3 and rec.709, which you might approach a little differently.
Cheers,
KC
-
Thanks guys!
Resolve couldn’t follow an alias created using the OS, but creating a symlink with the ln -s terminal command worked properly.
For reference, I used this command:
Theater1:~ Username$ ln -s (drag the Shared SAN path here) /Library/Application Support/Blackmagic Design/DaVinci Resolve/LUT
-
Hi Lars,
Looks like they made an update to the edge behavior in 9.1.4. I haven’t checked it out but I imagine it applies to this. Per Peter Chamberlain:
“• Improvements for image edges when resizing without the anti-alias filter”
KC
-
If you went into the project settings, enabled “assist using reel numbers” and selected the “embedded” option and didn’t see any reel numbers when you browse to the DPX files, there’s a good chance that metadata is missing. Resolve wouldn’t give you any special warning. You can also edit the Reel #s on a clip in the media pool by going into clip attributes (which supersedes the project setting).
Given that the timecodes all start at 00000000 (a worse crime) it’s too bad you don’t have Reels. Do the timecodes in the ProRes at least appear to all match those DPX? The only thing worse than overlapping TC with no reel #s would be offline media that has mismatched TC.
So to do it right, you’d need to give every DPX sequence a reel # in Resolve, and then go back and give the same to the ProRes. Resolve can systematically extract reel numbers from the directory structure if at least there’s some cohesiveness to the folder structure, otherwise you might have to edit the clip attributes for each clip to add them manually.
Cheers,
KC
-
Hi Glabs,
A couple ways come to mind for getting the media into the media pool, splitting it, and then linking up the EDL timeline to the media pool.
First possibility: Do the DPX files have overlapping timecode? If they don’t overlap, you could reasonably work without adding any reel #s. Bring in all the DPX media to the media pool, then load up the EDL and timeline, which (when it finds no matching reel #s) should settle for finding the right timecode. Note that this won’t split your media, so if you have 10 lab rolls scanned, you’ll need to give each clip it’s own version or use local versions when you grade. This way is a little sloppy if your media isn’t already split up.
Second possibility: Do the DPX files have a reel number embedded in the file (which you can find by setting Resolve to assist Reel #s using embedded)? If they do, you could go back into Final Cut and edit the Reel IDs for the ProRes to match the DPX reel names. If the DPX don’t have a filename, you could use a parent folder, or make new arbitrary reel #s, and then set the same in FCP. Having a functioning EDL this way will also help if you want to “add and split clips based on EDL” which would be helpful if your DPX are whole lab reels.
KC
-