Premiere Pro Clip Markers – Solved!

I’ve written a lot about Premiere Pro and metadata. Surprisingly, almost as an after thought, the solution was shipped in the October 2014 release of Creative Cloud applications. It is hiding in the preference panel under the “Media” tab, and is called “Write clip markers to XMP”. I believe this is one of the most important features that was added to Premiere, because it addresses the issue of backup, archive and project sharing and also helps with on-line/off-line workflows.

see more

Nuke Studio For After Effects

When I first saw the demonstration of Nuke Studio shown on NAB, with each new feature I was thinking: yes, this is exactly as it should be done. This is precisely how Dynamic Link should work from the get go in Adobe applications. Congratulations to The Foundry for making it happen. I wish I could afford your tool 🙂 see more

The Strange Maths of After Effects Timecode

For past few weeks I’ve been developing for That Studio a number of scripts for After Effects which are supposed to make one’s life easier, when dealing with an edit in Premiere that requires handling of more than a couple of VFX shots.

One thing that surprised me is that you can’t access layer’s timecode directly using scripting, even though you can do this using expressions. At first I thought that I can use the Time Remap effect to do this, since its values are supposedly shown as a timecode. But a brief experimentation shows that it’s not the case. Time Remap values are given in seconds from the start of the layer regardless of its timecode. Ouch.

Therefore one has to resort to a brutish hack to obtain the layer’s starting timecode: create a text layer, add an expression that reads the layer.sourceTime value and assigns it to the text, and then read the source text in the script. That’s hardly an elegant solution, and it would be so much better, if the layer.sourceTime was supported not only for expressions.

Perhaps I could live with it, but on top of it there is a nasty bug hidden in AE, and it can bite you rather hard, if you’re not careful.

You might be familiar with the fact that when you precompose a single footage layer, and choose to leave the effects in the main composition, the precomp will have the length of the whole footage file, and will inherit the exact timecode… more or less so.

You might not be aware, that if you are using 23.976 fps frame rate, the timecode might be off by a frame or even two when you attempt to render this layer – which you will notice only if you open the file, because even the render queue will show the correct timecode value. You can mitigate this by manually entering the starting timecode in the composition settings window.

The real problem begins when you are trying to set this timecode via scripting. Let’s say you want the composition to start at 20:20:51:22, which at 23.976 fps translates to 73325.2419085752. When you assign it to the desired composition, the internal AE procedure will truncate it to 73325.2421875, which will result in the timecode which is not frame accurate, and even though it is displayed correctly in the composition, it is incorrectly rendered and written to file. You can check it yourself by running the following script on a selected composition (make sure to first enter 20:20:51:22 into the Start Timecode of the Composition Settings dialog):

c = app.project.activeItem;
a = c.displayStartTime;
alert(a);
c.displayStartTime = a;
alert(c.displayStartTime);

Note, that the value changes by the sheer fact of assigning the code to the composition start time. My supposition is that the internal workings of AE convert the double value into a float, thereby reducing the precision at higher values. That’s definitely not nice. Depending on how critical the timecode is for your application, it might rule out automation of certain tasks, and make it more troublesome to render – you have to always manually enter the starting timecode, even if you precompose. I have not found a way to reliably code around this bug.

The conclusion so far is: forget about accurate timecode renders from AE, especially if you are using 23.976. 29.97 or higher fps and do not start from a full second. Lower integer frame rates seem not to be affected so much, as are easy dividers (12 for 24 fps, etc.). Unfortunately, even if this is fixed in the future releases, it still means that you can’t use scripts that rely on this feature for earlier versions of AE.

How To Fix Hiero XML Export For Premiere… Sort Of

Some time ago, if you ever tried to export XML from Hiero to Premiere, you got a generic importer error message and that was it. This issue – audio related – was thankfully fixed in later versions of Hiero. However, another hydra reared its ugly head, and this time attempted XML interchange resulted in something along these lines: see more

The Inevitable Convergence – Episode II

In the aftermath of IBC 2013 I wrote about the inevitable convergence of various software packages. It was easy to see how various vendors began expanding their packages into areas beyond the primary intended roles. NAB 2014 confirms this ongoing trend, and breeds more and more interesting solutions at various price ranges.

Let’s quickly sum it up: BlackMagic Design gave Resolve a serious boost in the editing realm and collaboration, The Foundry announced Nuke Studio, bringing Hiero timeline into Nuke – or another way around, if you prefer – upping the VFX management expectations for everyone and aiming towards the on-line market. Autodesk enhanced real-time timeline capabilities in both Flame and Smoke, while Adobe is constantly tightening the interaction between its various applications to make them work seamlessly as one. The case can be made that Avid is also attempting to do precisely that, gathering all its offerings in Avid Everywhere platform mirroring Adobe Anywhere though with proxy workflow instead of real-time server rendering.

All in all, this expansion outside the primary areas suggests that the applications are mostly mature, the toolset required to fulfil the primary functions is pretty much there, and the software companies are aggressively attempting to widen the user base. This is the case especially with grading packages, where the competition is relatively intense, and the high-end segment stops being perceived as the only viable support. Witness Digital Vision licensing its precision control surface to SGO Mistika, and going software-only route with its Nucoda, dropping its price in a clear attempt to widen its reach.

Which breeds the question – is specialized software doomed to fail in the long run? Will the likes of Baselight eventually run off of the resources to sustain themselves? Certainly, there are some comfortable niches where individual applications do and will exist – Mocha for planar tracking and Silhouette for rotoscoping seem to be pretty good examples. But they thrive in the space where they have no competition, protected by patents or relative obscurity. It’s a very cozy place to be in, but there are not many like these. How will Nuke fare against Mamba FX, now that it has Mac version? How will Premiere, Avid and FCPX survive the BlackMagic incursion?

Today for pure editing still nothing beats dedicated NLEs. I bet it might be a year or two before somebody attempts to do a larger editing project in Nuke Studio or Resolve. But I can easily see how shorter forms might resort to these tools, especially to Resolve for its unbeatable price point and relative ease of use, and Nuke Studio will comfortably find its place in the VFX editorial and possibly finishing.

Lastly, there is the problem of feature bloat and discoverability. When software starts to expand into areas not envisioned from the moment of its conception, the risk of hitting a development wall is pretty huge, since the base code and the user interface was not optimized for these additional tasks, and the forays will most likely appear clumsy to the eyes of the users of specialized packages. Nuke will never be as good roto software as is Silhouette, and I highly doubt it will outclass After Effects in motion graphics.

Will the convergence happen though? Will there be enough overlap between Adobe Creative Cloud, Nuke Studio, Autodesk Flame, and daVinci Resolve that the choice will come down to user preference and – gosh – pricing? Not unless BlackMagic partners with SGO, Eyeon or takes over Toxik from Autodesk. If that happens, all bets are off.

As for now, we can happily choose any tool we deem appropriate for the job and out budgets.

Showcasing an Adobe Workflow

Recently, after I was forced to go all freelance, I’ve been doing quite a lot of editing and VFX work for Kanen Flowers and his company That Studio. You might have already seen the teaser that we did for his new upcoming movie with the cryptic title “18:20”, but perhaps of more interest might be the behind the scenes that showcases the workflow used to manage the assets and coordinate the project – and a tiny bit of yours truly. see more

ProRes Encoding on Windows Finally Feasible

For quite some time one of the holy grails of multi-platform support has been the ability of encoding ProRes codecs on Windows. Or the lack thereof. True, there were makeshift solutions based on ffmpeg encoder that required either a knowledge of command line scripting or the use of other open source applications like AnotherGUI. see more