Alexa imaging and workflow questions

Daniel

First of all thanks for your keen interest. The process of CMOS imaging with a Bayer mask is discussed a lot with lots of mis-information. Its important to note that a Bayer mask is just a method for color imaging from a single plane sensor and deriving color information using software formulas to decode the information into RGB colors. Each company that uses this imaging technique has their unique way of processing the information to derive what they consider the optimal image.

We do it by carefully selecting the color dyes used in the color filters in front of each pixel, along with a very sophisticated optical filter in front of the sensor and then carefully decoding the color information derived from the Bayer mask.

The science and physics that go in to ARRI’s color science are part of what makes the ARRI solution produce such a great image, remember we have a very high end scanner (ARRISCAN) that scans motion picture film and provides very high resolution very precise RGB color images for film DI. The same people and science are part of the development for Alexa’s ALEV III sensor. In short, we have A LOT of experience in very high quality imaging with CMOS.

Also keep in mind that realtime playback of ARRIRAW (uncompressed RAW data from the CMOS imager) does not require hardware acceleration like others, chiefly because we use no compression (thus no decompression hardware needed) and there are several solutions based on NVidia Coda which use GPU (Graphics specific Processors) which are very inexpensive and reside on retail available video cards.

Not to stray too far into the technical, lets just say it this way.

Really smart and experienced scientists and engineers at ARRI developed a sensor called ALEV of which Alexa uses the 3rd generation. This sensor makes very high quality color accurate naturalistic organic film like images. We take the raw data from the sensor and derive (through processing in the camera) uncompressed 1080P HD, at 422 and 444 color spaces and output these streams from the back of the camera. We also encode this information as Quicktime files with ProRes encoding at 220Mb ProRes 422HQ and 330 Mb ProRes 444. We then record these images on SxS cards,

Look at the images we produce and record and decide if they meet your technical and aesthetic requirements. On the high end we offer ARRIRAW for the most discerning applications and we move down from there to uncompressed dual 3G (quad link) or dual link for 444 RGB applications as well as uncompressed 422 HD then ProRes encoded 444 and 422, you choose. Choose based on the application and image quality, not the on paper specs. Everyone I have ever showed images from Alexa to, has commented on their quality and natural film like quality, but you must decide for yourself.

Many (almost everyone) Digital Cinema movies have been acquired at 1080P RGB so there is plenty of image quality and resolution quality available for the big screen if ARRIRAW at 2K is too expensive or time consuming of a workflow.

What makes ALEXA more expensive than a couple of cameras (and far less expensive than most cameras) is the build quality, image processing inside and choices it affords for workflow. I think you’ll find that the cost of the camera is FAR outweighed by the savings in post in terms of time money for whichever workflow you choose. Remember 444 ProRes is still RGB and affords a huge range for sophisticated and complicated keying and compositing along with wide range color correction.

Please see the images and workflow then decide what works for you in Chile. ProRes 444 is a non compromise workflow when it comes to image quality and flexibility.


Michael Bravin
Alexa Jedi
VP Market Development
ARRI, Inc.