Messages

As I use the nightly builds I am in the habit of popping the battery out and in again if I'm not going to use the camera for a while to prevent the 'not shutting down fully, thus flattening the battery overnight' thing.

It's no big deal, I suspect certain combinations of parameters (IIRC, it tended to happen whenever I used 50p and FPS over-ride) get confused. I also recall that the indication this was happening was a lack of sensor cleaning cycle, although there are situations where this doesn't happen even if the camera does shut down completely - perhaps if it's been cleaned recently?

Anyway, the battery thing is just part of my workflow now. I suspect there are higher priorities for the devs, so after it's been reported, just don't worry about it.

Thanks so much everyone. FPS override, will test as soon as it's up.

https://bitbucket.org/hudson/magic-lantern/pull-request/464/fps-override-for-7d-in-raw-recording

Excellent. Thanks Walter.

OK. Another, no doubt often asked question (I couldn't find anything in a search, though). I have many hundreds of DNG files from the last few months. Now that CHMEE's lovely CDNG tool is the latest answer to my latest prayers, I was wondering how to convert these DNGs to CDNGs, as CDNGs are more compatible with a wider range of video apps and I'd like to use them as my standard workflow thingies. I've deleted all my MLVs, and before them all my RAWs so I can't go back to the source.

Any ideas?

Hey all! Been away from a while.. is the ML build up to date to the Tragic Lantern build? Or are still some features missing?

Thanks!

Has anyone managed to get FPS override working on ML for 7D? I've been away also.

Excellent! Thanks audionut.

I have been thrilled with ML since I first installed it on my 7D but when the exciting 5D raw features started to come through it felt a bit like the 7D had been orphaned. I am not complaining, not for a second.  I was, and still am, very impressed. I was booting ML every time I started my camera and accepted the fact that the 7D was old, difficult to work with and probably not as common with ML users as other cameras. I was content and grateful for the work that had been done.
Then I purchased a third party AC adaptor and discovered that loading ML for the 7D while in an alpha stage confused booting via the AC adaptor. I couldn't use ML with the adaptor as it stood and had to forgo using it whenever I wanted to use ML. I accepted this, of course, but one day I was browsing the ML forums to deal with some other issue when I saw mention of TL. I installed it (at my own risk, as is regularly and legitimately stated) simply to allow autobooting. Not really understanding why there was a separate version I became more and more excited as raw recording started to work. I continued to occasionally install TL updates whenever I saw that there were more features I might be interested in and when the testing done by others showed the build to be relatively stable.
After a while I was content and settled on a version. I set the camera to the best possible performance and was very happy with the situation.
But then the audio feature with raw appeared - I was in heaven! Everything I could possibly ask for out of this creaking old system was now possible and I was recording real work with the camera. Again, I settled on a version but came back occasionally to see if there were any performance improvements and/or bug fixes.  In the meantime (January) I was recording raw with backup production audio into MLV rec and couldn't believe my luck. I even decided not to buy an upgrade, I had 90% of what I wanted on my 7D already - I could wait.
That's when I started to see the comments about forks and 'rogue' builds. Although I am naïve about open source code versioning systems, it made sense to me. I saw the work done to merge the code into the main build and, wanting to be a responsible member of the community, installed the new ML builds with the intention of switching over and that would be that. However, the performance was disappointing and there was no audio so I 'downgraded' back to TL. I suppose this is one of the reasons that the ML team wants to keep things together - to keep expectations reasonable whilst keeping the code base consistent but it seems the horse has bolted on the principle here.
As someone who (very gratefully) uses the work of both A1ex (and Gagg0 and team) as well as 1%, I am hugely disappointed that it has come to this situation. I would very much like it if we could keep the TL work within the community until it's possible to merge into the ML stream. Although from what I have read here it's not a simple process and may not ever happen completely.
OK, I understand why it's important and agree with all the arguments but we have a mostly working TL version with excellent features and good performance along with information from people who have tested in quite a lot of detail to tell us what the best settings are. My concern is that all the hard work that has been done will go to waste and we'll never see equivalence on the main ML stream for the 7D. As people are doing this in their spare time I can't imagine what the motivation to improve the 7D build would be, it's an old camera and is about to be declared end-of-line, it doesn't have the x-factor of the 5D III or whatever is going to replace it. The TL version seems to me to be almost there with every possible theoretical improvement (maybe I am I wrong, are there any other magic features being considered for this camera?). A few more of the modules added and a few bug fixes would make it as close to the ML main stream as it's likely to be feature-wise (this is in my somewhat ignorant understanding - I know nothing about coding beyond basic HTML although I think I understand in principle the issues raised by A1ex). Perhaps with these tweaks it then could be frozen?
I'll accept it if I'm told to get back in my box here, I'm sure I don't fully understand the details. But I am very grateful to 1% for all the additional work, without any reduction in my admiration for the main team. I have effectively had two new cameras for free in the last year, first the original ML and then TL. I am once again prepared to settle on the latest, most stable version of TL if I have to but it feels like it would be a bit of a shame.

Crop mode becomes available when you press the zoom button to x5. I don't know if it's been fixed but I switch off x10 zoom in the ML settings as it used to freeze.

So, press the zoom button once, for x5.
(While still zoomed) open the ML menus and go to the video menu where you will see new options for frame size under RAW (if switched on) under the 'Q' button.
Also enable FPS override here, it should be in the video menu.
I also switch the number of buffers to 1, based on advice here. It gives a whisker more performance, which you'll probably need for the higher non standard frame sizes. I've managed 2048x1152 at 25 FPS using MLV rec with audio off (from the audio menu), although I had buffers at 2 at the time. I also switch the photo size to JPEG only, small size (RAW off).

Wow, M. Maurin. Extraordinaire!

My understanding is that Ginger HDR is based on a dematrixing wrapper. It loads raw files into the wrapper that can then be viewed in prempro and AE. There is a utility provided that can be used to create the wrappers outside of these apps as well.
Ginger HDR also comes with a couple of plugins that allow tonemapping. I believe these were specifically designed for HDR time lapse but can also work well on raw files which can have values outside the standard range.

The plugins are also an interesting, if number heavy and intricate, way to colour correct your raw (or any) clips. They can do things many standard colour correcting tools can't such as faux HDR and adjusting just the shadows, midtones or highlights in a way I haven't seen outside of a heavy duty cc app like Resolve or Speedgrade using qualifiers etc.
One of the compositors I came across at work sometimes uses Ginger HDR on computer generated high dynamic range renders which have been output in EXR format, particularly if he doesn't have access to Nuke.
One problem I've seen is that if Ginger HDR is installed in AE, it takes over ownership of raw files so the ACR workflow is not possible. There is a separate plugin folder for PP where it can be installed which allows you to use Ginger HDR in PP but leaves AE with the ACR method.

Note, this is a different location to the standard media core plugins, used for both apps and Media Encoder, for Win 7 the main folder is found at:
C:\Program Files\Adobe\Common\Plug-ins\7.0\MediaCore which installs Ginger HDR to both AE and PP.

Whereas the PP only Win7 folder is:
C:\Program Files\Adobe\Adobe Premiere Pro CC\Plug-ins\Common

Hey Tonybeccar. I hope you're enjoying your holidays. Just for the record, I don't use the proxy function, I bring the DNGs straight into AE (via ACR tweaks) and do some basic CC before rendering into an edit format from there. My machine is a few years old but I don't have problems with performance using the original files or transcoded versions so don't have the need for proxies. Others may differ, particularly if they go straight to editorial.

Further to Ted's comment above, now that the latest Magic Lantern (as opposed to TL) has the MLV sound module included, what are the experiences? Is it ready for use yet (with the standard caveats about it being a work in progress)? I'll do some testing on the weekend but if anyone has any info I'd be grateful. I'm looking forward to coming back into the ML fold, MLV audio has been the thing keeping me out.

[EDIT] Couldn't resist, took some time off work to test. ML Feb 20 build seems to work quite well. I managed 9 seconds of cropmode 1080p with MLV sound on. That's the best I've had. I'll try it with sound off later, along with more extensive testing.

Lossless means lossless compression which is exactly that. The files are compressed but no data is thrown away. Consider what happens when you zip or rar a file. If they looked any different on playback the compression would be lossy.

Hey tonybeccar, excellent job. Much appreciated. At this risk of sounding like an ungrateful knob (I'm not - ungrateful that is, although all my 'friends' will tell you that I'm a knob) will it ever be possible to use a different file browser within the app? I can get around fine with the existing one but with the  version I see in some apps I can type the path by hand if it's easier than browsing through a folder dropdown. Or I might copy and paste the path into an address bar at the top of the file browser window. In the current version, sometimes I have to click all of four or five times (!) to select the folder used for DNGs etc. Yes, I know. It's a difficult life :-).

No priority, the work you've done is already making my life much simpler. And I understand if you are stuck with whatever you are given by the environment. I've seen plenty of paid-for apps which are less useful than yours that use the same disclosure/dropdown-only file browser, it's not a show stopper.

No worries on the cineform, it's no longer a priority for me at this stage of my workflow. And as I said, I think you have to use the paid version to get the best quality (cineform RAW, or 444) anyway. I love this tool, thanks so much for creating it. It's really been useful for me and I'm enjoying coming back every few days to see yet more useful features.

This is probably not going to be possible, given that the codec needs to be purchased but as you're asking for AVI codecs - Cineform. I love the active metadata feature - Cineform RAW would be heaven.

Rawanizer works with Cineform but I've not been able to get the Cineform transcode working on my machine (I'm long overdue a rebuild) and, at last check, Rawanizer didn't handle MLV or audio.

Saying that, I've not used CF as much since the birth of RAW as I've been using ACR for much the same purpose - an active basic grade. However, I only work on short sequences so I'm fine loading RAW into After Effects and either rendering from there or transcoding to Cineform for use elsewhere.

Anyway, please don't consider this a moan or a begging letter. I really love this tool and have moved over completely to MLV (with audio) since starting to use it.

Downloading version 1.6 now.

Ok, I've done a 2k (2048x1152) test. Tragic Lantern 7th Jan, photo RAW off, small low quality JPEG (the icon with the blocky curve). 25FPS crop mode - Canon 100 Macro IS L 2.8 lens, buffer 2. MLV module with the audio switched off.

Non continuous recording (although 8-9 seconds at 25 FPS is fine for most things I do).

I'm gobsmacked. On a 7D, a 4-5 year old camera. After Effects standard VFX stabiliser (1.6% scale for the slight movement but I also scaled to 1080p for Vimeo). Recorded this on a monopod outside my studio window in Brighton. Used Adobe camera raw, tweaked the parameters quite a bit (boosted shadows etc) but as the light doesn't change no problem with flicker. Some very gentle CC using curves and a whisker of sharpen.

Fantastic. Now that we have audio recording, I'm switching completely to .MLV. This utility is the final piece of the puzzle for me, the only reason I was sticking with .RAW was due to the  number of steps to convert to  .DNG.

Actually, and I'm sorry if I'm alone here, I like the system where RAW is viewed when the camera is used for RAW and standard images (photos and H264 video) when it's not. What I've been doing is use one card (Lexar x1000) for raw video and an older card (Sandisk Extreme 60MB/s) for other stuff. I understand this might financially be a problem for some people but for me it means focus on what I'm doing. All it requires is a boot and card swap, easier even than going into the menus from an idiot proofing point of view.

And jman, nice benchmarks. I've not concentrated on benchmarking my camera, did it once ages ago and didn't pay much attention. I suspect any trace of the benchmarks are long gone, it was several versions ago and I've reformatted my card and reinstalled several times since then. Does it help, or is it for information only? I haven't been unhappy with performance at all. I switch global draw off, use a 128MB startup file, and only use frame override when I'm playing with crop mode. I was using MLV for a while but I like Rawanizer for getting the files to my HDD so I've switched back to old fashioned RAW as Rawanizer doesn't recognise MLV. I also have more options for the odd occasion my files are a bit finicky, I have a couple of MLVs that aren't read by mlv2dng and it's the only option on PC that I know about.

Yep, 1%. I've tested it, too - removal of VAF brings back nice sharp video using crop mode. I can even get a second or so of 2.5K - sharp as a tack and after quick reviewing in AE, haven't noticed any hideous moiré. Once again, I am gobsmacked at the work you guys are doing.

Thanks for the comment on the VAF-7D, 1%. What I was thinking -
(note, hypothesis on my part)
Mosaic filter is designed to work specifically at the 1728x972 pixel (not photosite) resolution, which is achieved by a particular formula of line skipping. From what I have read when the 7D records standard 1080p it is upscaled from this res anyway. Frankly, I'd rather do any upscaling myself, save a few CPU cycles during editing/post and choose exactly which algorithm and sharpening to use if I want 1080p. I'm unlikely to be producing video for anything other than the web or computer screen for the forseeable future anyway so I'm actually more likely to downscale to 720p. Doing this, with the raw files, I'm avoiding the upscale/downscale and going straight to the downscale from 972p!
Crop modes are closer to/(exactly?) a 1:1 pixel correspondence (probably not strictly true as we're talking about photosites but I'm simplifying by concentrating on the output, not the pre de-bayer) so the VAF-7D, which is effectively a clever form of blurring, is fighting me. Reducing image quality.
Does this make sense?
I'm risking being one of those people making ignorant statements about topics they don't know that much about but all the above certainly seems plausible to me.

Thrilled to bits with the new/old camera I have with 972p raw, I decided to push the envelope a bit so I've tried the crop mode. I was getting a reasonable number of frames at 1920x818 with a Lexar 1000x card. However, I have been disappointed with the lack of sharpness compared to some of the examples I've seen on Vimeo. It just occurred to me however - I use a VAF7D (which I highly recommend for the 1278x972 or standard formats) and my limited understanding of the filter is that it's designed specifically for the 7D system spitting out 1080p using the default Canon setup. Perhaps the filter is not agreeing with the crop mode? I'll test (probably next weekend) but I was wondering if anyone else had experience with this? It's not the lens - 17-55 2.8, which is nicely sharp as long as I get the focus right. Maybe it's just my rubbish skilz?

While I'm here (yes, yes, off topic again but I'm already adding to the noise with the stuff above) I've been sneaking in a few of the ACR basic adjustments again as I prefer what I can do with my images compared to the other tools I have. No flickering but the examples I've been working on haven't had significant changes in lighting throughout the shot. And I've been very moderate. I just haven't been happy with what I've been getting from Cineform Studio Prem (using matrix numbers from this forum), or Speedgrade, or Resolve Lite using RAWs or DNGs - so far. Again, could be user error. But there's something to be said for getting pleasing images with minimal effort/talent. Well, relatively minimal.

Er, I've been a numbskull. Should have been putting the post-pro posts in a different thread. One of the raw post-production ones :-) Apologies for the noise, the excitement got the better of me.

Cheers DFM. Excellent info on the ACR noise and sharpening. As recommended, I'll keep it to a dull roar and use AE for anything else.

Thank you for that DFM. Much appreciated. I feel much more confident now that I know that as long as I'm working in 32 bits in AE that I can still 'get at' all the data. I'd hoped as such, but without proper testing and your comment, it was a guess. So, I can confidently use video tools to do the problematic stuff I was doing in ACR. Lovely.
Your statement about NR - 'a tad of filtering but not too much'. Is that a comment on aesthetics or is there something that doing NR in ACR (as opposed to, say, Neat Video, or MB Denoiser) that means it's a worse choice?
Which also leads to my next question - Is the sharpening in ACR safe to use as well?
My trusty old 7D is exciting me again, and although there's workflow changes to go through, the results are worth it when you need them. I keep thinking of old projects I could have really used a few raw shots in :-)
When I was visiting the agency the other day (I just happened to have a CF card with the raw files on them :-) I had a play around in Nuke with the J_Ops additions. It's likely going to be a result of my limited skills with Nuke but I found the Adobe process more straightforward and I could get better results faster.
JZ