Paul Carlin
2
After Effects Coins
4
Apple Final Cut Pro X Coins
31
COW Coins
6
DaVinci Coins
6
Premiere Pro Coins
Forum Replies Created
-
To add my $.02, I would agree with those who are suggesting a proxy workflow… No matter what system you edit on. And since you plan on finishing in Resolve, why not edit in Resolve and avoid the entire conform process. Begin with Page 17 of the Resolve Manual “New Proxy Media Workflow” to see the advantages.
-
A hard drive docking station and some “archive” rated SMR hard drives are an economical solution to backing up your data. Make two copies if you’re paranoid. Instant access to the data when you need to restore it. I take a snapshot of the directory structure and save them as text files so I can easily search for what I’m looking for and know which drive to plug in. I also recommend a hard drive storage case like the Orico 20-bay suitcase. Grab and go!
-
Make a QT reference with the record timecode burned in before you make any changes. Then drop this QT reference onto a layer of your timeline and chop it up along with any changes you make. You can always SEE the previous timecode as it will be in the window burn of the QT reference.
-
I was curious to know if the latency of accessing the PostgreSQL database remotely was an issue for the colorist? How does it compare to a local PostgreSQL database?
– Paul
-
I know that this is an old thread, but perhaps this may help someone…
As Leon eluded to in the original post, the R3D files had already been debayered to DPX Log file sequences. The proper way to handle them in AE is to setup the project as 16 or 32 bit, color managed and to interpret the files as Universal Camera Printing Density. As long as you use this same Universal Camera Printing Density profile on the render, you will get output that matches the input. In the end, the rendered media should match exactly to what they gave you to begin with. You can check this in a non-color managed project.
To “normalize” the working space, simply check the “Linearize Working Space” checkbox in the project/color settings. The workspace can be anything you want. If you have the color management setup correctly and you interpret all of your elements correctly, there is no need to ever apply a LUT in a comp.
You should NOT be using the original R3D files as your debayer settings will not match the rest of the project’s debayered settings, unless you are debayering everything to begin with, or you have the exact recipe for the debayer. You should never do a “one light” as this bakes in YOUR color into the shot… which the colorist will not appreciate at all. You should ask for “vfx pulls” for the shots you are working on and use those (DPX or EXR sequences) and give back exactly what they gave you. “They” would be the people who are putting the conform together. If They can’t provide this to you… then good luck to you… you are working with amateurs, and they probably won’t even notice the color shifts anyways.
It would be nice to see AE support RED as much as they support the ARRI profiles. There are even output profiles for Sony S-Log. But as long as the input matches the output, you should be fine.
To answer Greg’s question, why DPX? DPX sequences make for a great way to preserve all of the information required to represent a film scans resolution and dynamic range (as well as tape and timecode and keycode, etc.). It just became an industry standard since Kodak created the Cineon format back in the day. EXRs have mostly taken over as they can better represent the very wide dynamic ranges required for HDR.
-
I know this is an old post, but so is my guide to exporting EDLs.
https://www.sparkwizdom.com/2010/02/how-to-output-edl-from-avid-media.html
-
Seems I was stuck in some destination track selection hell.
As you can see, the top timeline doesn’t allow me to enable or disable the source side, as it does allow on the lower timeline. After clicking a lot of things and pressing a lot of shortcuts, I got it to behave as expected. Not sure how I got stuck in that limbo.
-
Obviously, 30 frames don’t fit into a 24 frame timeline. Every fifth source frame will need to be lost, causing jerky movement.
Nearest does exactly that. It uses the nearest frame. 30/24 = 1.25, therefore the source frames needed would be 1.25, 2.5, 3.75, 5. Rounded to the nearest frames, that would be 1, 2, 4, 5. Notice it skips frame 3. This sequence repeats as the fraction 30/24 can be reduced to 5/4.
If you choose Optical Flow, then temporal motion estimation attempts to calculate what a completely non-existent frame 1.25 would look like by analyzing the movement in the images from one frame to the next. This is not perfect science, especially in areas of conflicting motion (think “car driving past a picket fence”). So the results are sometimes good, but mostly crap, and you would not attempt to rely on this for an entire project. Blend simply blends the two adjacent frames. The third frame (2.5) would be a 50%/50% mix of frames two and three, which also would look like crap (unless your entire project is a documentary of slow moving snails).
If you don’t want your project to look like crap, then set your timeline to the same frame rate as your source footage.
P.S. There are third-party tools for dealing with 30i (59.94) and 30p footage inside a 24p timeline. Check out FRAMES by Red Giant Software, part of the Shooter Suite. However, for all the same reasons above, you are still going to end up with different degrees of crap.
-
If I was an up and coming YouTuber, the new 4K pocket camera and the included Resolve software would be very interesting to me. I appreciate BMD for putting so much effort into improving their products (listening to users) and see their long-term strategy to be wise and disrupting in a good way.
-
23.976 and 23.98 are the same thing. 23.98 is the rounded off version of 23.976. The difference between the two would not explain 3.8 seconds of discrepancy.
The difference between 29.97 and 30 fps over the duration of an hour is 3.6 seconds. The same would be true for 23.976 vs. 24 fps as these are derivatives of 29.97 and 30 fps. Meaning that 23.976 only exists because 29.97 exists.
Drop Frame timecode is a way of dropping frames to accommodate for the 3.6 seconds lost over the course of an hour of real-time playback. In broadcast video (AKA color NTSC and everything that followed), there is only ONE playback speed and that is 29.97 frames per second. To have the timecode match “real time” as close as possible, certain designated frames were “dropped” from the time code counter (NOT from the actual playback video!).
To get to the point… it appears that your problem is related to the interpolation of the playback rate in the different applications. The difference between 23.976 and 24 fps would explain the 3.8 seconds of discrepancy you are experiencing. The actual playback is a completely different story.
The real question is, what is your deliverable? How will the video actually be played? And what is the true frame rate of the playback mechnism?
