[Marc Brown]

CS4 has given me nothing but headaches. Can’t help you figure out what the specific problem may be, since your issues differ from mine, but I can recommend grabbing CS3 and seeing if it works better.

One thing I might suggest is this: Hide your assets from PPro, and see if it can load. Then replace the missing assets one by one. You may have a corrupt file, or one which PPro simply doesn’t like. I determined recently that a certain DV capture I had, which worked fine in CS3, was causing PPro and AE to freak out, forcing a process kill. They evidently generate peak files differently.

[Marc Brown]

In the grand tradition of solving one’s own issues, I present the solution for anyone who finds themselves Googling this one. And it comes as no surprise to me that the solution makes as little sense as the problem. Good old Adobe.

The shutter angle. Even though it displays 180 degrees by default, it most certainly does not use that setting (at least not in my case) UNTIL you switch to “manual”. If I were to make an estimation, it may have been using an angle of 2400 degrees, to match the speed I picked for the time warp. Obviously nonsensical, but it’s a good match for the roughly one full second of “motion blur” I was getting.

This is for After Effects, which I am using preferentially because when you tell it you want 59.94fps, it gives you those 59.94fps, even if it has to bob interlaced footage. Premiere Pro hasn’t gotten that right in something like a solid decade of existence.

[Marc Brown]

Thanks for the reply.

Far as I have read and observed, Time Warp is identical in both applications. There is an alternative time stretching effect, but Adobe themselves point out that it’s not meant for dramatic framerate modifications, and they recommend Time Warp.

The thing is.. I have seen plenty of examples of the effect I am after. Clearly, the folks who generated their time warps found the magic, arbitrary settings which gave the result they wanted (and which should have been the default), while I have had no such luck.

[Marc Brown]

This one is still stumping me. If this was Avisynth, I could get it done in two or three lines of text. But I do not have that option. AE’s render queue does me no good specifically because it can still only do single-pass H.264 encodes.

It looks more and more like I’m going to have to get more HDDs and render this off as raw, because if I’m not getting a “generic error”, I’m getting an unknown error from Dynamic Link. The whole thing is just so gosh darned buggy (I have never gotten a cross-application function to work!) that I marvel that people largely consider it usable.

Meanwhile, I’ll still be hunting around for a way to get this done. A “reinterlacer”, if you will.

[Marc Brown]

Yeah. That’s sort of how I remember encoding the project a year ago, when I finally determined that Encore was never going to successfully encode the AE project, and let AE do its one-pass job.

However, in this case, what I need to end up with is a composition (not a render) which is 60i, so I can then import it into Encore, or Media Encoder, or Premiere Pro, as they are all supposedly capable of importing AE compositions. (Knock on wood.)

It’s probably something equally simple. Whatever step I’m missing which will allow me to interpret the “footage” of a new composition as 60i, while preserving each pair of frames as separate fields.

[Marc Brown]

Thanks for the quick reply (before I get some sleep myself).

Encore CS3 was probably meant to be able to render two-pass from AE, but really, I’ve played around with AE6, AE7 and AE CS3, and each one of them has had its own bugs which I’ve needed to develop workarounds for. Indeed, just toying with CS3 recently, I started getting a “not enough storage to complete this task” bug, at roughly the same spot in a given project, every time I tried to render it. I did have enough storage, needless to say. My solution was to render it in two (raw) parts and splice them together in a separate project. Just an example.

Adobe Media Encoder, eh? Intriguing. I’ll be certain to check it out. Hopefully it will be able to render an H.264 encode which Encore likes, and can use in a Bluray image.

But a standalone media encoder won’t let me add chapter marks (specifically, i-frames where I want them). Or perhaps it will? Premiere Pro had a good system. And then there’s the matter of the second audio channel. I guess I’ll have to wait until I can sit down in front of CS4.

[Marc Brown]

I’ll offer my thoughts. I just wrapped up a two hour video. Here’s a brief rundown of the steps I took to achieve some of the quality I was after:

1) Imported the videos into PPro 2.0 via Avisynth, as RGB (converted from YUV). This was done because PPro will chop off any luma above 235 or below 16, the moment you add an effect or adjustment. This is a bit hair-brained, because I have yet to encounter a digital camcorder which does not utilize all of the luma range above 235. Adjustments must then be made to the videos in order to compensate for the fact that their actual luma range is 16-255.

3) Did my edits.

3) Exported as RAW RGB. All of it. And the whole point was to shift luma from its original 16-255 to 0-255 (or 16-235, as the case may be) without simply chopping it off.

3) Imported the lot into AE. As Gabe Cotto helpfully points out, editing in AE is possible but terribly impractical. It’s best to be very sure of one’s edits before importing the footage. AE is very slow at previewing video (RAM or otherwise). I would say this is the #1 reason behind PPro’s existence. It even feels engineered that way. And let’s face it: The only possible reason why AE can’t preview audio without loading whole chunks into RAM is so AE doesn’t step on PPro’s toes.

4) Did my compositing.

5) Currently encoding with AE. I’ll be using Encore to create the disc image but that’s all I’m going to touch that app for. Not only does Encore fail to support multiple cores for encoding (a conspicuous lack, in essentially the only disc-authoring app that can boast Blu-ray support), but it’s so buggy that one is literally risking days of processing time in the hope that Encore won’t decide – the moment it finishes encoding – that it can’t find some file or other, and instantly lose everything it just did. Out of the dozen or so test disc images I had Encore try to make, two-thirds of them crapped out in the fashion I described.

The #1 thing that would speed this workflow up would be for Premiere Pro to begin handling YUV video more intelligently. Currently, it’s atrocious. Try importing some DV video, turning on the YC monitor, and adding a brightness/contrast adjustment. Watch your video’s waveform get crushed. Yet you don’t see this happen in PPro’s preview window because said window is _already_ crushing those details, underscoring the inexcusability of PPro’s YUV handling. What’s probably worse is that PPro never indicates that it’s performing this destructive process.

AE has similar shortcomings. It won’t indicate, for example, what it’s doing to a 59.94fps video when said video is imported into a 29.97fps composition. AE interlaces it. But was that what the user was after? Without some sort of heads-up, how can they know? It has to be tested. That’s pretty weak.

[Marc Brown]

3.45 hours is a lot of video for a single dvd, even a DVD9. Depending on the noise level present in the original video, the results could be horrid.

Having said that, I personally have not yet created a disc with Encore. I know enough about it now to realize that the moment there’s an alternative, I’ll be using something else to develop Blu-ray discs. But for regular DVDs, I’ve always used the combination of Nero and some authoring program.. usually DVD Lab for its quick and relatively painless building options.

[Marc Brown]

You, sir, are my savior.

For what it’s worth, the happy magical formula seems to be this:

For a (hd) composition at 59.94p, drag it to a new comp at 29.97fps. AE will treat it as a proper interlaced comp, rendering every pair of frames from the original comp as a pair of fields in the new comp (without telling you, or even providing any sort of indication). When exporting, open the “best settings” and pick upper field first. Make sure to pick it also in whatever renderer you use.

So in the end, my complications stemmed from two factors: First, AE’s lack of indicators as to what it’s doing with the video. Second, AE’s tendency to have video options secreted away in about a half dozen different settings menus, and my corresponding inability to divine such workings.

But frankly, AE is a breath of fresh air next to Encore. Now.. I just need to hope that AE doesn’t essentially lock up once it’s done rendering the video, as it has about 40% of the time. (Unless I really do need to wait longer than five solid minutes before AE ties up whatever it needs to do at the end of a render.)

Oh, before I go… Assuming the day comes when I have another two hour video project that needs to be burned in high-def.. How can I get MainConcept’s H264 (or even MPEG2) codec to render in 2-pass mode in AE? Encore has this option. In fact, it has tons of presets, while AE’s version has almost none. And the presets are not inter-compatible, sadly. A single pass render is all I have time for, this time around, but in the future, it would absolutely be handy to have better quality options.

[Marc Brown]

Thanks for the suggestion.

The MainConcept codec does give the choice of field order when one uses an interlaced preset. In fact it doesn’t give the option of not picking a field order. The problem seems to be that AE will not feed an interlaced signal to any codec, under any circumstance, including one where the user has gone out of their way to ensure that the output is interlaced.

A lot of headaches would have been saved had AE provided users with the ability to interpret nested comps the way one can interpret footage. Really, AE’s support for interlaced video is just one tiny notch above PPro’s.

I would use Encore to encode the video if it made any difference, and if I could trust the da** program not to give me an error. Test encodes more often than not return a “not found” error just after transcoding is complete, and when that happens, the only thing left to do is start over. Imagine the HORROR of having that happen at the end of a 50+ hour encode.