[Tom Brooks]

https://www.videothevote.org/resources/footagehowto?web=1

This link has what appear to be good recommendations. I would recommend H.264, 2-pass VBR at 2000-4000Kbps, with audio around 64-128Kbps, 44.1, mono for YouTube. Now, to extrapolate that to another video service is a guess.

The settings listed on the link are on the lower end of that, but should be good. My calculations suggest that, at 256Kbps upload speed, it will take at least 156 minutes to upload 20 minutes of video if the data rate of the clip is 2000Kbps.

You could probably get away with a data rate around 1000Kbps total and still look very good. This would improve the upload speed. Compare my guess to others you get.

[Tom Brooks]

No. I didn’t think of that. I’ll give it a try. Although, this is a pretty simple comp and the effects don’t go beyond the edges of the group. Maybe a little.
-Tom

[Tom Brooks]

Did you try setting rendering of Motion projects to Best in your FCP sequence settings? Normal setting doesn’t seem to get you certain properties from Motion.

[Tom Brooks]

Noah,
Thanks for the tip. This is the first time I’ve used Motion for a large project and really put it to heavy use.

I went through my project and converted the problem PDFs (basically all of them) to TIFs. In each instance, I’d make the size of the TIF so that its maximum scale in the project was near 100 percent. I could also apply a small amount of gaussian blur in some case and this helped cover up the sparkly, crunchy stuff.

This whole project has pointed out some annoying problems with Motion. First, the program seems to have trouble with animating scale, especially when an object is scaled to a very small size. Objects seem to randomly lose resolution during the scale animation. Something as simple as flying a logo through the camera could not be done without jaggies showing up in seemingly random places.

Motion also had a lot of trouble with its own backgrounds which use generators or emitters to create moving elements. If I used them in compositions with several layers of pictures and various masks and effects, I’d get weird color changes, black frames, partial frames–all sorts of anomalies. Turn off the background and things would behave. It’s likely there’s operator error involved in this, I admit. Maybe I had 3D layers intersecting or some such.

I suspect that Motion also has problems properly dealing with, and making, interlaced fields. I had no end of quality problems with JPEGs that had a vertical or diagonal move applied to them (these were very high quality JPEGs and the same effect was seen with TIF versions). When crunch time came, the quick fix was to simply find a way to make horizontal moves work. This suggests to me that Motion is not producing fields correctly. Turning off field rendering cleaned things up, but this project was DV and therefore interlacing was part of the deal.

Yet another roadblock to using Motion is the very long render times when rendering is set to Best in Final Cut. This appears to be necessary to get motion blur, but in some projects it basically made rendering too long to be practical.

In one instance, I gave up on Motion and did the sequence in After Effects. It had no problem with the scaling and no problem with the interlacing. In general, I can trust the quality of After Effects . In the near future, I’ll be unlikely to use Motion exclusively for compositing. There are things it does well and things that it can do easily that would be difficult to reproduce in After Effects. I think of its emitters, generators and replicators. But its very fundamental flaws will force me to use the slower, render-and-import workflow of After Effects. All in all, I hope Motion is better with progressive material and I hope that some learning on my part with improve my results.
-Tom

Final Cut Pro 6.0.4, Mac OS-X 10.5.5, Quicktime 7.5, Adobe Prod Prem CS3, G5 Quad 2.5, Kona-LHe V6, 8.5GB RAM, Nvidia GeForce 7800-GT 256MB, G-RAID 2x1TB FW800, 6TB RAID-5 (Enhance E8-ML, Highpoint 2322), Panasonic HVX-200P P2. Also MBP 17″ Core 2 Duo 2.5, 4GB, GeForce 8600M GT 512MB.

[Tom Brooks]

I really think it’s a codec issue. Meaning that you have a pretty non-standard timeline there. Let me make sure I understand…your sequence is set to 960×544, h.264, aac audio? If so, it’s a complete crap shoot as to how FCP will handle it. You should get it into a standard editing format first, then cut in FCP.

[Tom Brooks]

Ha ha ha. That was cheery. Yep, I love this little ol’ town.
Yes, my experience was with the MBP. I just blundered in with Qmaster, set it to two instances (duh) and tried it. The batch basically appeared to hang. Turned of sharing and it was fine. It seems to support your post quite well.

Thanks!!!

Final Cut Pro 6.0.4, Mac OS-X 10.5.5, Quicktime 7.5, Adobe Prod Prem CS3, G5 Quad 2.5, Kona-LHe V6, 8.5GB RAM, Nvidia GeForce 7800-GT 256MB, G-RAID 2x1TB FW800, 6TB RAID-5 (Enhance E8-ML, Highpoint 2322), Panasonic HVX-200P P2. Also MBP 17″ Core 2 Duo 2.5, 4GB, GeForce 8600M GT 512MB.

[Tom Brooks]

That would explain my informal findings in the heat of production last week that Compressor worked faster with QMaster sharing off. Quickcluster was set to two and it went nowhere. Turned it off and zoom.

[Tom Brooks]

FCP doesn’t automatically check anamorphic on the Widescreen DV sized files from Photoshop CS3. With the HD 720p files from the previous post mentioned in this thread, the anamorphic check is not needed, but it is for widescreen DV.

[Tom Brooks]

Say it ain’t so, Don. We were just havin’ some fun. ;-D

[Tom Brooks]

OK guys, since my work is done for the day, I had to clear up my own confusion on this.
1. Using the CS3 preset for DVCProHD 720p/29.97, Final Cut recognizes the pixel aspect ratio correctly and the clip plays on the timeline with no rendering.

2. Using 1280×720 square pixels, FCP again interprets correctly as square and scales the graphic to fit. It comes in with a green bar in Unlimited RT. You have to render for full quality playback (my MBP system).

3. If you start at 1280×720 but resize to 960×720 before import to FCP, the clip is interpreted with the HD (960×720) pixel aspect ratio and again plays on the timeline with no render. This would be a good choice for those who have Photoshop 7 or earlier.

I don’t see any obvious differences in positioning, but the real-time advantage is obvious here. Quality looks pretty comparable in all three methods. There are no doubt multi-format situations where it’s an advantage to have a square-pixel version of the graphics. But for me it looks like I need to accept the wisdom of the CS3 presets.

Final Cut Pro 6.0.4, Mac OS-X 10.5.5, Quicktime 7.5, Adobe Prod Prem CS3, G5 Quad 2.5, Kona-LHe V6, 8.5GB RAM, Nvidia GeForce 7800-GT 256MB, G-RAID 2x1TB FW800, 6TB RAID-5 (Enhance E8-ML, Highpoint 2322), Panasonic HVX-200P P2. Also MBP 17″ Core 2 Duo 2.5, 4GB, GeForce 8600M GT 512MB.