Lecture 6-Digital Cinema Systems Flashcards
What are the problems with film?
Expensive to: Stock, Store, Edit
Difficult to review, time consuming to produce.
Features of analogue video:
Low dynamic range
Linear response to light
Low resolution
Generation loss in edit
Features of digital video:
Inexpensive hardware
Inexpensive storage
Instant playback
Features of digital cinema:
Comparable dynamic range to film
Instant review and playback
Inexpensive storage
Increasingly high resolution
File-Based Acquisition:
Shooting involves the recording of data directly in a digital format. Data is transferred from recording medium to the NLE by file transfer using standard IT interfaces. As long as files are not transcoded, no loss in quality in any stage of post (e.g. Editing, grading, mastering).
Common formats and recording media: Sony XDCAM (MXF)
Looked ‘burned in’ to recording-changes limited
Record in RAW format, look not ‘burned in’ to data- created in postproduction. Metadata recorded separately on camera- ISO, gamma etc.
4K is standard format, actual resolution is lower due to use of Bayer pattern (4K is about equivalent to 3.2K)
Digital Cinema Camera Systems:
Resolution comparable to 35mm film, up to 5K, dynamic range comparable to film, 13 stops natively, up to 18 stops via HDR processing.
Exposure index comparable to film, native ISO of 800 for Red M-X and Alexa.
Completely digital workflow, metadata provides exceptional flexibility in creating final look.
RAW is uncompressed 12 bit linear colour space, film-style reproduction through logarithmic colour space. Not all RAW is the same varies camera to camera.
Cinematography becomes data acquisition, RAW file allows infinite adjustment of data present. Goal is create appropriate contrast and colour levels on set then expose to maximise data acquisition, exposure is not to create the final look but to get the most data.
What are CCD’s?
Charge coupled device
Longstanding analogue system to record video, light hits a pixel position, chip converts it to a voltage, camera convert voltage into digital picture information. Typically 3 CCDs used- line each for R,G,B, prism used to split light beam to hit the three CCDs, internal optics mean chips do not mimic film camera imaging, panavision genesis and others use single CCD.
CMOS:
Complementary metal oxide semiconductor, is an active pixel sensor, light converted to voltage, amplified and converted to digital data on chip, typically only one chip used, imaging is very similar to film recording in term of optics, etc. Progressive scan from top to bottom mimicking film shutter create ‘rolling shutter’ effect.
How do Red and Alexa imagers derive colour?
Bayer pattern of sensors
What is a bayer pattern?
A mosaic of these colours, full image derived using a computationally intensive formula that ‘de-mosaics’ the colour array. Scaleable settings-higher value, better quality.
What is debayering?
Done either in-camera or in postproduction, debayering quality varies from system to system, high-end postproduction systems give best quality. Not all debayers are the same.
What is Wavelet compression?
Red and Alexa feature Wavelet-based compression, rather than reduce picture elements into macro blocks (like MPEG), signal is decomposed into frequencies, individual frequencies are sampled for amplitude. A high compression Wavelet images go soft not blocky.
Digital cinema distribution:
Digital cinemas operate at 2K or 4K resolution, common file format DCDM, digital cinema distribution master- uncompressed, unencrypted and not for distribution itself a set of files used to create DCP- Digital Cinema Packages, files that go to cinema for projection. Compressed (using JPEG2000) encrypted to prevent piracy. Maximum native resolution of 2K or 4K. Various other resolutions based on aspect ratio, 12 bit encoding in an XYZ colourspace.
File-based Workflows
Final distribution medium dictates approach the Caboose effect. Determine file format and codec based on delivery needs broadcast requirements is 8-bit, film 10-bit or 16-bit, DCP 12-bit, Red and Alexa encode at 12 bits. Downsample at mastering stage for broadcast.
