Jump to content


  • Content Count

  • Joined

  • Last visited

Everything posted by Sen

  1. http://motionworks.net/new-features-in-ae-cc-13-5/ • Interactive performance seems to work great (but very simple one layer scene in the video) • Friendlier handling of expression errors • Multi-processing & brainstorm have been removed
  2. OS X ready for NAB 15. https://www.blackmagicdesign.com/videos/fusion
  3. Sure! In half float (16bit) EXR you can go up to ±16 stops above white. You can measure the RGB values of overbrights in the C4D picture viewer while pressing command/windows key and hovering over the brightest spots. btw. On linearized camera footage the highest values I've measured (reflection glints from the sun) are around 60 (around 6 stops over 100% white) but it depends on the latitude of scene, camera chip or film stock.
  4. Did you try EXR half float (16 bit)? That should cut your file size in half. 32 bit is only needed when you have values greater than 65504. This can happen with tech passes like world XYZ or Z-Depth. ZIP16 gave best compression with CGI when I did a test.
  5. Color management (which the Color Profile Converter is part of) in AE is based on ICC profiles. This is a somewhat different setup to most other VFX packages that are based on a 1D LUT (similar to a curves effect) or a 3D LUT to do their color transformations. OpenColorIO is going to be the upcoming opensource standard. The main advantage is that it will work identically between programs (ex. Nuke) and your color transformations will look the same regardless of the program used. NuclearLUT is 1D LUT converter and provides a lot more transformations than the Color Profile Converter. I'm actually working with Sony F55 footage right now in Slog2 which is available in NuclearLUT. You can do a complete manual color management workflow in AE: always enable "preserve RGB" on in- and output files, pre comp the footage and convert to linear with one of the plug-ins. Disable 'Use display color management' under View and add a guide layer with a linear to sRGB transform. This is like the old eLin days in 2004 when there was no 32bit in AE but it's still a valid workflow and it can clarify what AE is doing under the hood.
  6. I agree but it's not that AE messes up. Some tools are expecting a perceptual, gamma encoded color space. Best way as you describe is to sandwich the effects between color profile converters and convert from linear to sRGB/log and go back to linear. Inverting an image and adding grain are also two examples that are nicer in a gamma encoded space (log). http://prolost.com/blog/2006/6/4/know-when-to-log-em-know-when-to-lin-em.html Alternatively, you can convert between gamma/color spaces with Fnordware OpenColorIO or QuarterLight NuclearLUT. Gamma 2.2 is not 100% equal to sRGB. The gamma 2.2 curve approaches black at a lower angle than the sRGB spec so there's some slight crushing of the blacks.
  7. If you're working in a linearized color space and want to input gamma encoded sRGB values you can apply the “Color Profile Converter” effect to solids, text and shape layers. If you set the Input Profile to sRGB, it will convert the gamma encoded colors to linear color values and it will match up visually. AE can’t guess if you want to input color values in a linearized scene-referred space or in a gamma encoded space.
  8. Apple Prores codec got reverse engineered somewhere in 2011. That's why it's available in ffmpeg/ffmbc. There exists a GUI http://www.fallenempiredigital.com/blog/prhelpersoftware/ Cinemartin could also license it from Apple (like Arri Alexa, Aja and BM do).
  9. The MTV piece appears to be ripped from Umeric. (1m34s)vs https://vimeo.com/40592892
  10. New page is up: http://www.apple.com/mac-pro/ All custom chips apart from the RAM maybe? I guess it'll be rather difficult to upgrade/expand. Cooling seems very efficient.
  11. Size is incredible. I hope they add Nvidia support (CUDA apps)
  12. Did you try turning off "Use Display Color Management" in View menu? That should fix the double gamma compensation in the composition window. You could use any 3D LUT box to apply an sRGB/Rec.709 gamma curve to your HDMI/SDI output. ex. http://www.lightillusion.com/eecolor.html IMO, even the simplest operations (opacity fades, image blends, soft edge masks, motion blur etc.) look more photographic in a linear float workflow. I agree that you need log footage (cineon, dpx) where there is a set relationship between real ratios of light in the scene and the pixels in the file or 3D CGI (OpenEXR) with overbright values to really appreciate using this workflow.
  13. I like nvALT (a fork of Notational Velocity). It supports tagging, Dropbox sync and Markdown syntax.
  14. Texturing soft is Mari. http://www.thefoundry.co.uk/products/mari/
  15. Yeah, it's quite a difficult problem to solve. If you anti-alias your Z-depth pass, any anti-aliased pixels would indicate wrong depth values and result in artefacts. In case of aliased Z-depth pass and anti-aliasing on the image, the passes don't align exactly which result in another type of edge artefacts. You can deal with the second problem by rendering your aliased Z-depth pass at double resolution. Add the oversized Z-depth pass and the image pass (scaled at 200%) in a pre-comp. Apply Lenscare and then scale the pre-comp back to 100% in another comp. You'll notice that Frischluft is really slowing down because of the oversized comp… :-) If you're still seeing some edge artefacts you can try applying the minimax effect (operation: maximum, radius: 1px) on the Z-depth pass. In case of obscuration artefacts, you need to render in layers (FG, BG) and blur separately because you want the pixel information behind FG objects. Deep compositing data (Open EXR 2.0) is the ultimate solution, but it's going to take a while to arrive in AE and C4D, I suppose. http://www.deepimg.com/
  16. ProEXR provides automatic comp creation and layered file export. Speed is the same as OpenEXR afaik.
  17. AE does not include the latest version of the OpenEXR plug-in (v1.5). There were some speed improvements in v1.6 and v1.7. Especially when reading layers with EXtractoR. Try downloading that first: http://fnordware.com/OpenEXR/ Storage hardware also makes a difference. I'm using a 4xSSD RAID0 (1200MB/sec) but if you really want to scream check out Fusion IO cards. http://www.fusionio.com/products/iofx/
  18. I’m using ZIP16 for mograph and PIZ for HDR photos. The difference between float (32bit) and half-float (16bit) is considered mostly academic. I would avoid B44(A) in production because they use a JPEG like compression scheme. Good for client review though. There is an excellent overview in the manual of ProEXR, see pages 30-32: http://fnordware.com...oEXR_Manual.pdf ps. Quite blushing, all the big guys looking at me. ;-) I'm a motion designer living in Brussels, currently finishing VFX for a belgian documentary.
  19. It was added in a R13 update and hidden in the options dialog (16 bit float).
  20. Certainly possible, render out material UVW + object mattes (front, back side) Use http://www.revisionf...products/remap/ or http://aescripts.com/ft-uvpass/ plugins to map your textures in AE.
  21. This tutorial could be helpful: http://www.c4dcafe.com/ipb/index.php?/files/file/716-rotating-image-slices/
  22. Katrin Eismann's Masking & Compositing and Advanced Masking Techniques are also good resources on channel based masking.
  23. There is indeed an annoying gamma shift when playing back .h264 encoded material due to a bug in Core Video. Maybe they will fix this in upcoming Leopard? More info: http://sprynthesis.com/blog/?page_id=51 http://lists.apple.com/archives/QuickTime-...l/msg00007.html http://lists.apple.com/archives/quicktime-...l/msg00018.html p.s. New here, lots of internal server errors when browsing the forum.
  • Create New...