Film Gimp

gosand writes "DesktopLinux.com is running this story about Film Gimp. It is a movie editor based on The Gimp that movie studios have been developing for their own use for a while now. The article is an interview with Robin Rowe about Film Gimp's use, and includes some interesting info about the film industry's use of GNU/Linux desktops. One quote worth noting: 'Studios have become the leading desktop users of Linux. Three hundred Linux desktops at Dreamworks. That's amazing! While the MPAA is campaigning for new restrictions on content, the artists at the studios are using and helping create open source. Having Linux and open source as a crucial part of studio operations may help executives rethink their corporate position on open source and Linux issues.'"

Comments

Not just better playback

[UberLame]

Film gimp adds lots of support for superior playback. However, the biggest and most importanted different is that it uses 16 bits per channel instead of only 8 like the regular gimp. That means that instead of roughly 16 million colors, you get 16 million squared colors. This adds much less chance of rounding errors on compositing, and gives you more room to play with when adjusting brightness and color balance over 8 bit images.

The downside is that film gimp is based on an old version of the gimp, and it doesn't really look like that is going to change soon. But at least they are talking about syncing up a bit before 2.0 whereas before they seemed to be planning on waiting for the Gimp 2.0.

Re:Not just better playback

[ajs]

It uses 16 bits per channel instead of only 8 like the regular gimp. That means that instead of roughly 16 million colors, you get 16 million squared colors.

And since that math doesn't work, let me explain for those in the cheap seats:

8 bits *per channel* means a total of 24 bits, which yields 16777215 color combinations. When you hear people say things like "32-bit color", they're talking about storage, not bits-per-channel (which will still be 8, and the excess 8 bits is used in various ways, including as an alpha or transparency channel that desribes the opacity of the resulting color).

This adds much less chance of rounding errors on compositing

That's quite untrue. What it does is reduce the impact of rounding errors. The errors persist all the same, and really the solution is to film digitally, composite digitally and master digitally. At which point such problems become insignificant even at 8-bits-per-channel.

Re:Not just better playback

[Pseudonym]

That's quite untrue. What it does is reduce the impact of rounding errors.

Which is highly misleading. What 16 bits per channel really gives you is more dynamic range.

We CRT users are used to going from black (electron gun turned as far down as possible) to white (electron gun turned up as far as possible). "White" for us is a certain point on the black body radiation curve at a certain intensity.

Film users know that this is not the maximum number of photons which are available. In real life, you can always add more. On film, you can usually only add about 20 times more before it's fully exposed, but it's still more, and it's noticeable on specular highlights, such as on chrome or water.

Cineon digital negatives specify 10 bits per channel in logarithmic space, which appears to be sufficient for capturing the high dynamic range. Converting this into linear space (which is more convenient for manipulation) corresponds to about 16 bits, about 12 bits of which (i.e. the range 0-4096) are what we think of as the black to white range. The rest is "headroom" for those times you need the full dynamic range.

not really surprising

[circusboy]

from working two years trying to sell a piece of technology to the members of the entertainment industry, I have come to realize that there is no group more interested in getting something for nothing than the entertainment industry.

as a result, I'm not at all surprised to find OSS in the major studios, being used to create stuff.

places like ILM exist successfully largely because people give them hardware for the joy of being known as the hardware that ILM chooses. then people ignore the fact that the reason they choose that hardware is largely based on it being free.

Digital Domain moves ALL workstations to the Linux

[Anonymous Coward]

Check this out:

"Digital Domain is transitioning all of its 2D and 3D production workstations to include NVIDIA Quadro4 XGL professional graphics solutions, NVIDIA's Unified Driver Architecture (UDA), and the Linux operating system. The company is also deploying NVIDIA Quadro4 graphics hardware and Linux software drivers in its software development, digital content creation studio and systems administration departments."

Why Gimp rejected Film Gimp

[burgburgburg]

According to the page, the rumor is that the committee saw the Film Gimp effort as the prototype, "the one you throw away" and decided to put their efforts into gegl.

Re:Well, that's all good 'n' all

[MoneyT]

depends on if they're distributing the program or not. IIRC one of the reasons people bitch about some of the corporate opensource licences (like Apple's) was because they made it a requirement that you released the source code. Correct me if I'm wrong here, but that's the impression I got.

Re:Dumb Question.

[Anonvmous Coward]

"If a conventional monitor can't display the colors at that depth then how does the film editor know exactly what will end up on film after printing?"

I think you're looking at it the wrong way. It's not about what's displayed on the screen, it's about having enough color information to prevent color banding when doing things like brightness and contrast adjustments. The color banding is a result of rounding errors that stair-step the color values when adjusted too heavily. 16-bit images have a greater degree of accuracy, so rounding errors are reduced, as is the resulting color banding.

The end result is dithered down to 24-bits, but anything can look good at 24-bits. It's not a problem until you need to tighten in on information. It's kind of like resizing an image from 320 by 240 to 640 by 480. The image looks great at 320, but there are artifcacts to blowing it up to 640. If there was subpixel information in the original image, then the expansion to 640 would go a lot smoother. Try to imagine that in the color space.

*hoping I expressed that in a way that makes sense*

Re:Dumb Question.

[Thagg]

> If a conventional monitor can't display the colors
> at that depth then how does the film editor know
> exactly what will end up on film after printing?

Not a dumb question at all -- unless by dumb question you mean one that will start hour-long religious arguments that have no resolution :)

Seriously, one typically sets up one's monitor or display software to show a 'window' into the film's dynamic range. You can choose where you want to clip the bright values based on what part of the scene you're working on.

In the end you can get a good enough idea of what will show up on film that you are rarely too surprised -- and if you are surprised, you make changes based on your experience, and film it out again.

thad

Re:CMYK, and GIMP UI vs. drop-down menus

[mgkimsal2]

I can't draw a straight line with it.