Topics

I found out about a video competition a few days before the deadline (end of September), so I was up last few nights working on it. The rules are: it has to be less than one minute long, and the theme was "the eye."
Here is what I submitted:

https://vimeo.com/184777117

Starring my wife and myself

I wanted to do a sound mix and a million other things- but there was no time/budget.

Gear list on the description page

Is there a link somewhere to a primer on what crop mode is - and when one might want to use it?

More specifically - I'm wondering if it will help me turn a non-macro lens into a sort of macro lens, e.g. to get imagery which is closer to the subject and still be in focus than would normally be possible?

When developing a module, is it necessary to compile, install onto card, boot card onto cam, enable module, run it, rinse and repeat?

More to the point- I'm wondering how the devs approach iteration and testing. Do you build a bunch of tests and hook them up to buttons or whatever and then run it? e.g. write larger chunks at a time, test more rarely?

Basically just looking for advice about iteration and approach to coding ML for more fun

Would it be possible to read custom remote control codes?

I'm thinking for example to setup an Arduino with IR output, and be able to send new commands the the 7D to be processed with ML. Not the builtin record/stop/etc.

For starters- even just to log "hello world" or whatever.

My ultimate goal is then to use that as some sortof syncing mechanism to do genlock (http://www.magiclantern.fm/forum/index.php?topic=1596.msg171511#msg171511) but one step pulse at a time!

I feel like a dunce, but I can't find that button anywhere... can someone post a screenshot? 

When shooting 1:1 RAW I understand the VAF filter is a huge bonus. If shooting in crop mode, should the filter be removed?

I'm hoping to shoot 3D soon for a project on the GearVR. My camera budget is around $1000-$2000. I'd like to get advice from y'all about the wisest route to go... mind you I live in Israel so shipping fees, customs, headache, are all relevant factors. Here's what I'm thinking:

I currently own a 7D. It's served me well over the years and I am incredibly grateful to the Magic Lantern devs for breathing new life into it and helping me pimp it out to squeeze best possible footage out of it. My kit is ML Raw, Komputer Bay 64GB cards, and VAF-7D Mosaic filter. I have two lenses with it, the sigma 30mm f/1.4 ex dc and a jupiter-9 with m42/ef mount.

So this gives me a few options:

1. Buy a second 7D and matching lens to the Sigma (will be easiest to match focus and predict close look) and a genlock signal generator. My question is - does ML support genlock on the 7D somehow?

2. Buy the DualHero rig and two GoPro Hero+ 3 Blacks, as well as the 5.4mm lenses for each. Probably don't pay for the "pink fringe" removal. This is a pretty niche area so feel free to ask if google doesn't tell you what some of this is about. This would basically just be for this project and then not really do much for me past that.

3. Buy two Black Magic Micros, genlock sync generator, and lenses. This goes way beyond my budget - so this option means selling my 7D kit

Note that I do want to still be able to shoot regular stuff from time to time. Option 1 and 2 allow me to do that with my 7D. Option 1 will let me do dual-camera shoots even. Option 3 means I shoot regular stuff on the Micro.

I'm bringing this question to the Magic Lantern community, mostly to see if option #1 is viable. If it is, I think it's probably the way to go, since the larger sensor size of the 7D will give me more freedom in general than the Micro....

This is the workflow I used for the film @ http://www.magiclantern.fm/forum/index.php?topic=12720.0

Resolve (offline edit) -> FCP XML & Then -> After Effects (color grade w/ Film Convert) -> PSD Sequences -> Premiere (online "edit", not really much- just fade ins/outs) -> DNxHD -> Adobe Media Encoder (H264).

Why use Resolve for an offline editor and then after effects for color grade?! And if I'm already doing that, why not just do the online edit in Resolve as well?? Glad you asked

1) Resolve as an offline editor.

The real question is why *not* just use premiere. The answer is that Premiere sucks at playing DNG natively. By using Resolve as the offline editor, I could bring in the DNGs and edit natively with perfect playback, and even play with rough sample grades, without affecting the original DNGs or creating any intermediates.

2) FCP XML
I tried exporting to regular EDL and others and it was hard to read. FCP XML is easy to read and I can reference the clips by their actual file number. I.e. the <in><out> tags. That way in After Effects, I can match those numbers and export exactly what I need for online edit only

3) After effects as coloring

MLV did not look nice in Resolve natively using BMD Film settings. I know I could bump up the exposure in raw interpretation, but that just feels wrong, and didn't quite look as good as ACR

Furthermore, in after effects I use ACR/Bridge just to get white balance and exposure (so no flicker)- as well as denoise and sharpen. Since this is all happening on the native DNG, I can undestructively go back and tweak, using the benefits of ACR and raw

On top of this, there is then the entire arsenal of after effects to tweak the look/color. I used FilmConvert and Color Finesse.

I'm also very used to / fond of Color Finesse. I tend to get better results more easily with that than any other tool, especially when the intended look is natural (can then use other effects in after effects to get unnatural looks anyway... such as magic bullet or whatever)

4) PSD Sequences

I've just had battle scars from doing anything else in the past. I know that when I export a PSD, it is what it is, and don't need to worry about if any program is going to chop the bit depth or shift the gamma or whatever. PSD sequences work perfectly in Adobe products...

5) Premiere

As I said, PSD work perfectly in *Adobe* products

6) DNxHD

This is the one part I'm not happy about. Adobe won't export a DNxHD 444 at 48/64 bit depth without it getting all messed up. Google the problem, and you'll see the consensus is that DNxHD ignores this setting and simply uses it's own determination to do 10 bit instead of AME's settings. I've tried it, and indeed bringing the exported DNxHD back into After Effects for comparing shows it as "Trillions of Colors" and no visible difference really. So, reluctantly I'm trusting this and anyway it looks fantastic

7) Premiere crashed when exporting from timeline, so I brought the DNxHD into AME. No big deal. Rendered at High Profile 5.1 maximum quality and depth

[EDIT]

Youtube version:

Vimeo version:

Uploading DNxHD master now... will update and share here as long as y'all promise not to kill my bandwidth charges If the upload doesn't time out, should be ready in a day or so.

Okay, so what's the story with the "Weird Workflow"... well.. it went:

[ EDIT, moved to Workflow forum @ http://www.magiclantern.fm/forum/index.php?topic=12723 ]

What's the difference between those?

Just thought of a funky idea... not sure it would be much use, and not sure if I'll break my head over it just yet, but for now I'm wondering it the scripting capabilities would allow it

1) Display overlay image for liveview. Specifically this overlay image could be, for example, a drumset placed over the lower 1/3 of the screen (most useful if via external monitor)

2) Detect collision between predefined object (say a green square) in real life and overlay (this could be simple detection- bounding boxes)

3) Play sound file which was stored on card at volume determined by program (ideally, velocity of collision), or generate real-time wav for playing through headphone jack.

The possibilities here could be really cool, since the high definition of cameras ML runs on, you could get some really detailed control/movements....

Is this possible? If not, why not?