The VFX behind "Äkta människor" (Real Humans)

Update: Just added a bit of info at the end of the post about the Arri Alexa ProRes4444 LogC workflow with Rec709 LUTs in After Effects.


Design for the hubot source code for “Äkta människor” from Jonas Hummelstrand on Vimeo.


Visual effects scene from “Äkta människor” (“Real Humans”) episode 5, season 1, 2012.

Tracked and roto’ed with “mocha Pro” with tracking transfered to Adobe After Effects CS6 via the “MochaImport” script. Adding the photo to the truck was an afterthought during editing, and the side of the truck didn’t have and good features or markers for proper tracking (plus the fact that it has a huge translation, is covered by the actors head, multiple reflections and out-of-focus.) If this was a feature film, I’m sure we’d have had more time to adjust the details it, but being a broadcast drama, we had to abandon it without a lot of finessing…
3D neon sign modeled and rendered in Cinema 4D as 32-bit OpenEXR.
Comped by Nicklas Larsson and Jonas Hummelstrand.
Rendered in 32 bpc and output as 16-bit LogC TIFF.
Original source was LogC ProRes 4444 from Arri Alexa.

As the in-house graphics department of SVT (the Swedish equivalent of the British BBC) we do everything from high-end visual effects for drama and trailers, to motion graphics branding and daily news graphics. We have a wide range of talents that can do anything from 3D modeling to building video play-out software.

In early 2011 we got a request if we were equipped to do all the visual effects and on-screen graphics for an upcoming 10-hour drama called Äkta människor (Real Humans.) The show is something as unusual as a Swedish sci-fi series, set in a parallell universe were the humanoid robot evolution has had the same exponential progress as computers have had in our society. Like all good sci-fi, it focuses on what happens to us if we change one major factor, in this case; what happens when every middle-class family suddenly can afford to have a robotic servant that take care of the chores, and that is human-like enough to get involved with, both mentally and physically…

Although the effects that were planned at that stage weren’t tremendously complicated, it looked to be a big job judged by our broadcast standards. The visual effects budget was really tight, initially only 60 man-hours in total per one-hour episode.

When shooting started in April 2011, we had done quite a number of tests to nail down a look for the recurring effect of the sub-skin lights underneath the eyes that indicate a “deep charging” robot (called “hubot” in the series.) Other than that, there was a very undescriptive list of shots that might need effects, but not much more.

As the visual effects supervisor, I knew that planning ahead was essential if we were to be able to keep to the tight budget, and at the outset I insisted that I was given time to be on set for the shooting of some of the trickier effect scenes. Initially, the plan was to lock the edit for each episode before we started work on any of the effect shots, minimizing the risk of working on effects that were later cut or changed. However, the further the eight-month shooting schedule progressed, delays in shooting and editing started to creep into our allotted time so that we for the last half of the episodes had to work before the edit was locked.

Scene from “Äkta människor” (“Real Humans”) episode 8, season 1, 2012.

Tracked with Imagineer Systems’ “mocha Pro” och brought into Adobe After Effects CS6 with the “mochaImport” as an undistorted and stabilized precomp. Comped in 32-bit float and output as 16-bit LogC TIFF.

Original source was LogC ProRes 4444 from Arri Alexa.


Since most of the show was shot on a Steadicam for quick setup and turn-around, the first stage of our VFX pipeline was almost always a planar track in “mocha AE” that was then brought into Adobe After Effects CS5.5 and CS6 via the excellent script “mochaImport” that lets the artist create a stabilized and undistorted precomp of the areas that will need treatment. This was essential, especially for the sub-skin lights that needed to be tracked to faces that move around and turn. By adding effects to this undistorted shot, and then downstream getting the distortion applied back to the composite made for quick turn-around of shots that would have been impossible just a few years ago. I was so confident in our ability to track everything that the DoP repeatedly asked if I didn’t want the camera locked down for the tricker shots, I always replied that “I actually prefer if you move the camera, it helps sell the effect!”

Scene from “Äkta människor” (“Real Humans”) episode 5, season 1, 2012.

Comped in 32bpc float in Adobe After Effects CS6. Tracked with Imagineer Systems’ “mocha Pro,” 3D neon sign modeled and rendered in Cinema 4D.
Output as 16-bit LogC TIFF.
Original source was LogC ProRes 4444 from Arri Alexa.

As the shooting progressed through the summer of 2011, the script kept evolving and the director’s initial trepidation with doing the effects in-house evaporated. This meant that the list of scenes that could and would need VFX grew quickly. At the same time we were doing a lot of logos, symbols, posters and around 15 touch interfaces for disguised iPads used to change settings and firmware in the robots. The interfaces were made for interaction by the actors, so the result was shot in camera to avoid lengthy screen replacements, as were the around 30 screen animations designed for display in computers and TVs. In the end we only did eight screen replacements where the video that was to be shown in the screens didn’t yet exist at the shoot.

One of the biggest effects sequences was the explosion of a store front. The location featured a big parking space with a nicely lit exterior, but no natural opening for an entrance other than a giant garage door with doors that could not be removed. Instead of building a real neon sign and glass doors and that could be blown up (at the cost of around $25,000) we offered to do a digital replacement. With TV production, the schedule is so tight that we couldn’t pre-plan more than really rough camera angles. We devised a single Steadicam shot showing a stunt man walking out from the store, and in sync with a special effects arsenal of propane burners and burning-debris canons, being pulled by a wire into the side of a car. The car was rigged with squibbed windows and a hydraulic jack that caused it to jump at the time of the explosion. The camera would then walk behind the stunt man as he turned around and looked at the devastated and burning store front, at the time represented by a semi-lit green screen behind lots of smoke.

Scene from “Äkta människor” (“Real Humans”) episode 5, season 1, 2012.

Comped in 32bpc float in Adobe After Effects CS6. Wire removal with the “Remove” module in Imagineer Systems’ “mocha Pro.” Dust particles created with Trapcode’s “Particular.”
Output as 16-bit LogC TIFF.
Original source was LogC ProRes 4444 from Arri Alexa.

The shoot went fine, but we quickly realized that this shot required more advanced tools than were available with the combo of “mocha AE” and After Effects. The large amount of smoke and parallax from the moving camera made the manual wire removal painstakingly complex and slow. After a few days we gave up and bought “mocha Pro” to get access to the “Remove” module that made it possible to finish the wire-removal the same day we got the licenses. When it came to isolate the stunt man from the smoky green screen, “mocha Pro’s” tracking-assisted rotoscoping capabilities and variable-edge feathering also helped to save the day.

Everything was shot on Arri Alexa at 1080p25 to Apple ProRes4444 codec as LogC. We processed everything as a 32 bpc workflow with Rec709 LUTs used only for viewing, and output offline copies for the editors as DNxHD 36 or 120 with baked-in Rec709 for import into their Avid MediaComposer 5 suites. The final delivery was done as 1,920 x 1,080 16 bpc LogC TIFF sequences that were conformed in a Nucoda grading suite.

Visual effects scene from “Äkta människor” (“Real Humans”) episode 5, season 1, 2012.

Comped in 32bpc float in Adobe After Effects CS6. Wire removal with the “Remove” module in Imagineer Systems’ “mocha Pro.” Dust particles created with Trapcode’s “Particular.”
Output as 16-bit LogC TIFF.
Original source was LogC ProRes 4444 from Arri Alexa.

Almost all of the 290 VFX shots in season 1 was comped entirely in After Effects, with a few exceptions that were done by a freelancer in The Foundry’s “NukeX.” Being in broadcast, After Effects is the natural choice and our go-to software, but we have now bought a Nuke and a NukeX license that we will start using more for similar projects in the future.

The first season (9.5 hours in total) has 26 minutes and 24 seconds of visual effects in it. All in all, we’ve spent on average 50 minutes per delivered and aired second of visual effects, which I believe is pretty effective even by broadcast standards. Add to that around 100 hours of on set supervision, and a few hundred hours for the other assets we’ve created, such as iPad GUIs, screen animations and other UIs such as web pages.

It’s pretty amazing to be able to do high-end, seamless and photo-real HD effects on hardware that costs less than $2,000 per machine and with software for even less than that. Being an in-house shop, I know that we need to be extremely efficient, but I know of similar TV shows that have paid more for one tenth the amount of similar effects done “on the outside.” And ten years ago the budget for this would have to have been ten times that!


Test for charging indicator for “Äkta människor” from Jonas Hummelstrand on Vimeo.


Test of eye blink removal “Äkta människor” from Jonas Hummelstrand on Vimeo.


Workflow: We have now (November 2013) delivered two seasons and a total of 67 minutes of 1920 x 1080 VFX for “Real Humans” using Arri Alexa ProRes4444 LogC with 98% of the shots comped in After Effects. Since on-set-viewing was all done with Rec709, it was straight-forward to comp in 32 bpc under a “LogC2Video” LUT, but deliver as LogC, so that all VFX clips matched the rest of the footage.

  The workflow has been very straight-forward:
  1. Import ProRes4444 with default settings (no color management, no ‘Preserve RGB’ or anything.)
  2. Create Adj. layer with “Apply LUT” and choose “LogC2Video” LUT created in Arri’s LUT Generator, to match the on-set look of Rec 709.
  3. Set the LUT Adj. layer on top of every comp, and set it to a Guide layer so it doesn’t render.
  4. Any RGB sources get a “Apply LUT” with a “Video2LogC” LUT applied to them.
  5. Comp in 32 bpc and view everything with the “LogC2Video” LUT.
  6. Render out to ProRes4444 at 32bpc. Since the “LogC2Video” LUT is a Guide layer, you just “Add to Render Queue” and hit Render.

We’ve done this for everything, and haven’t had any problems. Overbrights are kept, rendering is fast, the dynamic range has been fine, and we can easily render out with baked-in Rec709 by just setting the LUT layer to “non-Guide layer” so that it renders.
Grading is done on a Nucoda system, with Rec709 as a rough starting-point.

I’m sure that the 10 people in the world that understand Ae’s color management would say that there’s a better way of doing this, but since I’m not one of them, we’ve been very happy with this workflow.

If there is a season 3 made, we will investigate using the OpenColorIO plugin for After Effects, thanks to Simon Björk who has pointed me in the right direction.



Äkta människor: Official site

VFX by SVT Grafiskt Center

See also Nicklas Larsson’s blog for more breakdowns

CasparCG Server 2.0 Released

The professional graphics and video play-out server for broadcast just got a major update.

Here’s just some of the awesome new features:


NTSC, PAL and lower-field interlacing

NTSC, PAL and lower-field interlacing in SD and HD

With the added support for NTSC, CasparCG Server 2.0 can now play all the current broadcast formats in both SD and HD, and output them to one or several video cards and/or screens simultaneously. Choose between progressive, upper- field or lower-field interlacing, it’s up to you!


Multiple video cards in one server

Multiple video cards in one server

You can now use several video cards in the same machine, no matter if you use cards from BlackMagic Design or Bluefish Technologies. You can even use cards that have multiple channels, letting you use smaller cases that are easier to fit in a small rack.


Much, much higher performance

Much, much higher performance

By rewriting the software for multi-threading, CasparCG Server 2.0 can now take advantage of all your processors by performing many tasks in parallel. Other tasks have been moved to the graphics card, further increasing the performance and enabling real-time functions like scaling. Reading files from disk is now asynchronous and the buffering has been greatly improved.


Built-in video effects (DVE)

Built-in video effects (DVE)

All playing media can be manipulated in real-time. You can move, scale, mask and change the opacity and video plus audio gain with the new real-time Mixer feature. Take advantage of the built-in animation types, just like with transitions. The Mixer can even de-interlace video, transform it and then re-interlace it correctly.


Generate alpha channel from a separate file of any type

Generate alpha channel from a separate file of any type

While you can still use the embedded alpha channel of videos, Flash templates and images, you can now also override the alpha channel with a separate file. You can, for example, cut a hole in a video file with an animated Flash template, or even an image that you can
This is also useful for HD video where the alpha-capable codecs tend to generate very big files. Now you can play two small files, such as H.264; one for fill and one for alpha.


Play all common formats

Play all common formats

Previous versions have always been able to play many video formats and codecs, such as DVCPRO, DVCPROHD, MPEG-2 and many AVI and QuickTime-wrapped codecs such as Animation+. That support is now extended with codecs such as DNxHD and H.264. And, now that you can generate an alpha channel from any media type, it’s easier than ever to output good looking HD without having to use large files.
The new version of CasparCG Server can also display most common bitmap formats directly.
Grab input video and overlay graphics

Grab input video and overlay graphics

Video and audio input into a DeckLink card can now be used just like any other media. Use it as one of several layers (each of which can be transformed with the Mixer feature) or just overlay other media such as video and graphics on top and then output it. You can even cut a hole in the input video with another video, a Flash template or an image.


Embedded SDI audio

Embedded SDI audio

Both the Bluefish and DeckLink video card families can now embedded audio in your SDI output. Some DeckLink cards also output audio in the HDMI and AES/EBU connectors.


Output to computer screens with DVI and HDMI

Output to computer screens

The fully GPU-accelerated Screen Producer can be used to display playing media on one or several computer screens, either fullscreen or windowed. Choose if you want your content scaled to fit the screen or displayed at the native resolution. If you have several screen attached you can either choose to mirror them or use each as a separate output.


Custom loop points and pause

Custom loop points and pause

Video and audio files can now have a custom in-point for the built-in looping functionality. You can also choose to pause / freeze any playing media, even Flash templates.


Monitor real-time performance

Monitor real-time performance

With the new Diagnostics window you can see a graphical break-down of how long each step in the process of outputting a frame takes. This helps you optimize performance and trouble-shoot lagging video by seeing, for example, how long it takes to read a file from disk.


Stack multiple media freely and re-order them on the fly

Stack multiple media freely and re-order them on the fly

All the media you add to an output channel is composited in real-time, using the alpha channels. Play videos, Flash templates, audio and images at the same time and choose the stacking order freely. CasparCG Server 2.0 can even reorder media while it’s playing!


Grid-view of all playing media

Grid-view of all playing media

With a simple command, you can output all your playing channels as a grid view to a screen and/or video output. Combined with the feature that lets you grab the input video from a DeckLink card, you can build yourself a monitoring station!


Real-time, animated transitions between all media types

Real-time, animated transitions between all media types

Create a instant transitions between playing media, no matter if it is video, Flash templates, audio or images, by using the built-in functions. Mix between a video loop and a graph, or wipe on a chart on top of an image, for example. You can use a number of easing curves for transitions of all types. Transitions now support both interlaced and progressive output, and audio is mixed during the transitions.


Works on Windows 7 and Vista

Works on Windows 7 and Vista

CasparCG Server 2.0 now fully supports modern operating systems and can be installed on both 32-bit and 64-bit versions of Windows 7 and Windows Vista. Take full advantage of faster disk access and enhanced performance when upgrading from Windows XP (which is still supported.)
All configuration has moved to a text file instead of using the Windows Registry, and we don’t even force you to use an installer anymore. You can move the CasparCG folder freely and place it wherever you please!


System Overview

CasparCG Server 2.0 Overview

CasparCG Server 2.0 Overview(Please click to view full size)


Other notable features and enhancements in CasparCG Server 2.0a

  • Video, audio and image files can now be placed in subfolders, just like Flash templates.
  • Smoother animations and better interlacing in the Flash Producer.
  • Outputs (Consumers) can now be added and removed at run-time, without having to restart the server.
  • Swap producers between layers and channels during run-time.
  • Parallel decoding of audio and video.
  • Real-time commands viewable in the new Console window.
  • Improved logging with full exception details, and a new log file for every server start.
  • Backwards-compatible AMCP protocol supporting all new server features.
  • Find out the running version of the CasparCG Server and the TemplateHost via a network query.
  • DeckLink output is now using the card’s hardware clock.
  • Bluefish drivers are now loaded on-demand, making it possible to have a single Server application.
  • Per-sample mixing between source and destination clips during video transitions.
  • The Flash Producer now uses DirectDraw access, resulting in improved performance.
  • All new GPU-accelerated internal pipeline with the following functions all being done on the GPU: Compositing, Colorspace transforms (RGBA, BGRA, ARGB, YUV, YUVA, YUV-HD, YUVA-HD), Interlacing, Per-layer image-transforms: opacity, gain, scaling, clipping, transform, plus the Color Producer.
  • Major code refactoring for improved readability and maintainability. 66% of legacy code removed.
  • Rewritten to take advantage of common standardized libraries instead of custom solutions.
  • Work done to facilitate cross-platform porting.
Cross-platform demo client as an Adobe AIR app

Try out some of the new features with the included demo client. The ready-to-run Adobe AIR app is also available as a source project if you want to learn how to build your own client.


To download CasparCG Server 2.0a for free, go to the official CasparCG site.


- Jonas

Great Time to Buy Software During NAB

Here’s a round-up of all the NAB discounts I’ve found so far. Remember, these deals will normally end when NAB ends, so be quick. If you find others, please post a comment!

  • Zaxwerks offer up to $200 off on AE plugins and stand-alone 3D apps
  • RedGiant Software offer 30% off on all software for Facebook fans, and 20% off for all others. That’s Colorista II, Magic Bullet Looks and all the Trapcode filters as well!
  • Knowledge at your fingertips‘ After Effects Script 50% off, use NAB11 code
  • Fusion 6.2 for $995 incl. 1 year subscription
  • Cinema 4D can be bought for 25% during April
  • GenArts offer 50% off on Monsters GT for AE and free particleIllusion for AE when you buy Saphire. Or 10% off all purchases (can’t be combined with other offers)
  • All DigiEffects AE plugins are 50% off with the code 50NAB11
  • mocha AE and mocha Pro deals
  • ft-Toolbar AE script 50% off, use code TOOLBAR50NAB at checkout
  • All NoiseIndustries’ FX Factory plugins for AE, Motion and FCP 20 % off
  • Artbeats offer 30% off on stock footage
  • Automatic Duck offers $100 rebate on any purchase
  • Boris RED for $595 ($295 upgrade) with a free upgrade to version 5 on release
  • Many of these specials can also be found at ToolFarm’s NAB specials

Show Reel 2011 Posted

A quick collection of some of my broadcast work for SVT, including motion graphics, design, 3D, time-lapse photography, set design, virtual sets, visual effects and more. Make sure you view it in full HD resolution!

- Jonas

>Just Released: Free and Open Source Broadcast Graphics Server


Free and Open SourceWe’ve finally released the new version 1.8.0b4 of the CasparCG system, a broadcast graphics system that plays QuickTime and Flash out to SDI, HD-SDI, HDMI and DVI.
Download installers, source code and tools from

>Great stroll down After Effects’ memory lane

>Thanks to Matt Silverman for posting this great video with an informal talk by three of the engineers of Adobe After Effects.

SFMOGRAPH – After Effects Team from Matt Silverman on Vimeo.

- Jonas

>Patches for After Effects CS4 and Premiere Pro CS4


Don’t miss the updates!

After Effects 9.0.1 patch for Mac OS X
After Effects 9.0.1 patch for Windows
Read me file (PDF)

Premiere Pro 4.0.1 patch for Mac OS X
Premiere Pro 4.0.1 patch for Windows
Read me file (PDF)

Here’s the goodies from the read me files:

Notable fixes in the Adobe After Effects CS4 9.0.1 patch:

  • Miscellaneous performance improvements with MediaCore video formats.
  • General improvements to Dynamic Link workflows.
  • Mocha AE doesn’t read v210 QuickTime files correctly. (49015)
  • On some systems, the application may suspend rendering when the display turns off due
    to power management or screen saver settings. (49146)
  • All Render Queue items are removed when opening or importing a saved project in which the first item has no comment, but other Render Queue items have comments. (49186)
  • Crash using “Open Layer” when multiple layers are selected if one of the selected layers does not have a source or has a collapsed transformation. (49207)
  • Auto-save is marking a project as saved; user can close project without being prompted to save changes, and multiprocessing can return old frames. (49209)
  • (Windows only) Custom eyedroppers in keying effects pick up wrong colors when clicked in custom thumbnail. (49248)
  • Exposure effect: negative offset values render garbage/underflow in 8bpc and 16bpc. (49294)
  • Puppet Tool: For a 3D layer, results are being clipped to layer bounds. (49267)
  • Importing a still sequence via Recent Footage imports only a single still instead of a sequence. (49265)
  • For some OpenGL effects in 32bpc projects, overbright colors are getting clipped.(1893997)
  • When the “Write XMP IDs to Files on Import” option is selected, MediaCore playback can make some files appear out of sync. (1880742)
  • Application may appear to hang for several minutes “Reading XMP markers from footage” (status message) with files that contain deeply nested metadata. (1874477)

New in the Adobe Premiere Pro CS4 4.0.1 patch:

  • OMF export
  • Final Cut Pro import
  • Edit in Audition (Windows only)
  • Surcode support
  • AAF import and export
  • High quality motion
  • Third-party support

Notable fixes in the Adobe Premiere Pro CS4 4.0.1 patch:

  • Attempting to trim locked tracks results in a crash (#1888355)
  • Crash on Win and error on Mac when using Numpad-Asterisk keys to generate marker when
    playing in Program Pane when Effect Controls pane is closed (#1890626)
  • Crash when the Metadata panel is brought into focus when a DVCPro HD clip is selected
  • When a filter/search is applied in the Metadata Panel and then a different asset is selected, the
    filter does not update (#1876954)
  • Custom schema causes Adobe Premiere Pro to crash performing many basic operations,
    starting with New Project (#1888938)
  • Render of transition in imported EDL-based sequence fails; on Mac, causes application to hang
    indefinitely, user must be force-quit application (#1873434)
  • After ASND is saved, an effect is always ON if it turns OFF while playing (#1864978)
  • Crash when attempting to load a project that contains a composition if headless Adobe
    Premiere Pro is serving Adobe After Effects (#1871578)
  • Exporting video as MPEG 4 will create a video file that imports at the incorrect length
  • Trimming both ends of a clip, then moving the clip will cause looped playback to be silent, or
    only playback portions of the clip (#1868409)
  • When exporting a timeline containing cuepoints, deleting a cue point via AME Batch-> Export
    Settings, the cue point is still present in the output file (#1866690)
  • Playing an audio-only file in the source monitor will cause the application to lock up (#1887330)
  • Exporting any MT session as 24bit WAV leads to garbled audio – Mac only (#1881684)
  • Adobe Premiere Pro allocates too much memory and reads too much data from disk when
    loading project files on the Mac (#1890579)
  • With camera connected, opening clip from Timeline into Source and selecting File  Debug
    results in a serious error (#1889197)
  • Export to tape is active from DVCPRO HD sequences (#1863718)
  • Application intermittently crashes when exporting a movie (#1868405)
  • Render and Replace of offline clips takes over 20 minutes (#1885273)
  • Application crashes when cancelling “Normalize Audio Track” dialog (#1887615)
  • When an AVCHD clip is inserted into a DV timeline with Scale to Frame Size applied, rendered
    previews of the clip show artifacts during Timeline playback (#1863988)
  • Using scale to frame size in a timeline lowers output quality and softens detail and sharpness;
    typing in scale attributes manually or using Adobe After Effects produces much better results
  • Multiplexed MPEG files exported to HD size and then imported back in to Adobe Premiere Pro
    show up as 601 color space instead of 709 color space and hence the colors of the imported
    files are off (#1852977)

Known issues in the Adobe Premiere Pro CS4 4.0.1 patch:

  • Important note: if you are using the Elemental Technologies RapiHD plug-in, you must update
    to the latest version in order to use Adobe Premiere Pro CS4 4.0.1. You can update the plug-in
    by visiting the Elemental Technologies web page:
  • Re-installing Adobe Premiere Pro CS4.0.0 after it has been patched will generate an error
    dialog at the end of the installation. This dialog can be ignored.
  • Uninstalling Adobe Premiere Pro CS3 after CS4 is installed will result in AME exports in CS4 to
    fail with ‘Could not read from the source’. Creating a shortcut named Premiere alongside the
    dynamiclinkmanager.exe that points to the Premiere.exe will fix the problem.
  • When converting Final Cut Pro project with Color correction applied, the color correction will be
    incorrect within Adobe Premiere Pro.
  • When converting Final Cut Pro project with Bins, the bins do not appear in Adobe Premiere Pro
    unless selected in the Final Cut Pro xml export.
  • Multicamera enable sequence with HDV clips does not playback as normal; severe drop frame
    or stopping playback occurs
  • Audio in a nested 16 channel sequence does not simultaneously playback all of its audio tracks
    in the Source Monitor.
  • File>Adobe Dynamic Link>Send to Encore selection is available when a 16 channel sequence
    is selected; however, Adobe Encore does not support 16 channel sequences.
  • File>Export>Export to OMF… selection is grayed when the Audio Mixer Panel is selected on the
  • File>Export>Export to EDL… selection is grayed when the Audio Mixer Panel is selected on the
  • 5.1 sequences use output mapping from the stereo channel column of the Preferences>Audio
    Output Mapping dialog when the audio device selected from the Map Output for menu has less
    than 6 channels.
  • All 16 channels of a 16 channel sequence are summed to mono without attenuation when
    exported as media when Audio>Basic Audio Settings>Channels is set to Mono.
  • Both odd and even numbered channels of a 16 channel sequence are summed to both stereo
    channels when exported as media when Audio>Basic Audio Settings>Channels is set to Stereo.
  • 16-channel selection does not appear in the Export Settings dialog Basic Audio Settings
    Channels menu when a 16-channel sequence is selected.
  • All channels of a 16 channel sequence are summed to Ls, Rs and Center channels when
    exported as media when Audio>Basic Audio Settings>Channels is set to 5.1.
  • Setting the Optimize Rendering preference to “Memory” is highly recommended when Max
    Quality or Max Bit Depth rendering is on. These settings are designed for high performance
    systems. Please check Adobe Online Support for recommended system requirements.
  • Loading an existing project with a P2 sequence that is missing media may cause a crash.
  • Adobe Premiere Pro crashes when importing AAF from Avid if it contains references to clips
    with OMF audio (WAVE or AIFF).
  • AAF may not provide consistent results for complex projects in formats other than DV or HDV.

- Jonas

>RED Finally Cuts in Adobe Apps

While RED has published a picture of the Premiere Pro 1.0 box on their RED Adobe CS4 Installer (beta) page, you’ll actually need the newly released version 4.0.1 of Premiere Pro CS4 and version 9.0.1 of After Effects CS4.

The system requirements aren’t exactly designed for the famed “soccer mom’s” of this world; you’ll need to have a 3.0 GHz quad-core system, at least 8 GB of RAM and a 64-bit OS such as OS X or Vista 64 if you want to work comfortably, but using lower resolutions with lesser machines will work as well.

Dave Helmly has posted a video workflow (but could someone please help the guy with the design?)

- Jonas

>RED’s Latest Line-up

>The December 3rd updates to RED’s earlier announcements seem to be moving in the right direction. However, I can’t really figure out if we will be able to get a 2/3-inch Cinema Scarlet sensor with interchangeable lens mounts.
If we can mount 3rd-party lenses to it, with a REDhandle, a viewfinder (BombEVF?) and a recording module the price would presumably be in the vicinity of the $3,750 of the 2/3-inch 8X Fixed Zoom complete kit.

- Jonas

>Why the RED cameras won’t be seen at pool parties

>If RED was aiming their new Scarlet to be a “DSLR killer,” they missed the mark, unfortunately.

I love what RED is doing to the dinosaurs of the camera business, but I was hoping that a Scarlet would replace my prosumer gear and give me new possibilities. Unless the Scarlet with the fixed lens is dirt-cheap I don’t see me getting it.

Instead RED’s proposed system may well be a revolution for motion camera design principles with the sensor being decoupled from the rest of the gear, but this current design won’t mean anything for me based on these factors:

  1. Stills resolution
    For still images the 2/3″ Scarlet’s resolution is just 4.9 mega-pixels where at least 8 MP or more was needed to become a usable alternative to any cheap DSLR. You need to go up to the S35 sensor to get 13.8 MP.
  2. Movie resolution
    No doubt this is where all RED systems excel, and the 2/3″ sensor’s 3K sounds like a sweet spot where the data rate is still manageable while you will still have the ability to crop or down-sample to HD / 2K. However, I’m pretty happy to get 1080p out of my cheap rig or from a Nikon D90 or a Canon EOS 5D Mark II (which are just the first versions of what will surely become standard features of traditional DSLRs.)
  3. Frame rate
    The great 120 fps frame rate available in the 2/3″ sensor drops down to a mere 30 fps in the other Scarlet sensors (unless you crop in on the sensor and loose resolution and DOF,) which is a bit of a disappointment. Being able to over-crank to get beautiful slow-mo has always been a dream of mine…
  4. Lenses
    You also need to go up to the S35-sized sensor to be able to attach readily available still camera lenses from Nikon and Canon, which suddenly brings us to the next point:
  5. Price
    Just the S35 sensor and mount for still lenses bring the price up to $7,000, without a viewfinder or any accessories. That’s a far cry from something that a prosumer could justify spending on a camera body, considering Canon’s and Nikon’s bodies which go for under $2,500. If you get the 2/3″ sensor you need to get new lenses with special mounts.
  6. Shallow depth-of-field
    The 2/3″ sensor is so small (see Stu Maschwitz’s sensor cheat sheet) that you won’t get the cinematic shallow depth-of-field and bokeh that we all love, without putting a DOF adaptor and additional lenses in front of the 2/3″ Scarlet. A super-16-sized sensor is way better most current video cameras (many features have been shot on super-16,) but it’s not comparable with the size and look of a DSLR, even those with APS-C-sized sensors like the Canon 30D.

I was hoping for “3K for $3K”: a camera body that would do 2K or more at at least 75 fps with a still camera lens mount and shallow depth-of-field-look, and at least same stills resolution that my old Canon 30D had. I guess I’ll have to wait another couple of years…

- Jonas

Page 1 of 1612345»10...Last »