This is a companion article in which we explore the viability of creating 10bit content. We researched and devised test content to determine which codec will bring us the best visual result, at the fastest render time, allow for server side compositing, and not instantly fill up very valuable media server storage space.

In this summary piece we document the test itself listing out the intention, the workshop parameters and the hardware used to conduct the test.

For the main article see the original entry here: https://dandelion-burdock.com/articles/10bit-workflow-viability

Intention

Proof of technical delivery and ensuring signal integrity from the playback device, through a signal router, fibre transmission/reception, to the projector at 4096x2160 (4k DCI), 10bit, 4.2.2, 60Hz.

Workshop

The full line test was conducted at Christie Europe, April 2019. Video playback (disguise), signaling (Lightware), projection(Christie Digital) and workflow professionals (dandelion+burdock) were present at this test.

All media displayed was created by dandelion+burdock LLC, to comply with DPX 10bit 4096x2160 pixel image sequence played back at 60 frames per second. The media reel only contains generic material. This reel was made for the purpose of 10bit signal testing.

The projector was set to about 35% output, at a throw distance of about 4m, onto a white projection canvas. We measured a consistent 4637K color temperature, 3380 Lux (direct illumination).

Hardware

Signal Proof Method

For all 10bit verification steps, a high bit depth, high resolution version of the below image, excluding the overlaid key, was employed.

For each step up the possible steps of grey increase by a power of 2. If there is banding in the top two steps, which is similar to the 8Bit row, only 8bit signal depth has been achieved. If the two top steps become finer progressively, the signal is transferred and displayed at 10Bit.

Individual Tests

The following tests were conducted in the order presented. The intention was to verify the signal flow and performance with each more complex step. We outline the flow and involved hardware, followed by any issues and the appropriate resolutions.

A

10bit signal proof and DPX sequence playback from media server to 4k UHD monitor, via direct HDMI cable (2m maximum cable length)

B

10bit signal proof and DPX sequence playback from media server to 4k UHD monitor and 4k DCI projector, via 2 discreet direct HDMI cables (2m maximum cable length)

Please note that the final projection will hit colored and textured surfaces. Therefore the projected image’s color changes, and will not comply with colorimetry as per official standards (see section 5.1/ Material Testing).

C

10bit signal proof and DPX sequence playback from media server to 4k UHD monitor directly, and 4k DCI projector, via 4k HDMI Video Matrix, with input and output HDMI cables (2m maximum cable length), 2x DCI 4k output from disguise Server to the Matrix, 1x DCI 4k to the projector, 1x UHD 4k to the reference monitor

D

10bit signal proof and DPX sequence playback from media server to 4k UHD monitor directly, and 4k DCI projector, via 4k HDMI video matrix switcher, with input HDMI cable (2m maximum cable length), output via HDMI>Fibre transmitter, via OM3 fibre, to Fibre>HDMI receiver.

E

10bit signal proof and simultaneous DPX sequence/10bit VideoCodec/8 bit VideoCodec playback from media server to 4k UHD monitor directly, and 4k DCI projector, via 4k HDMI video matrix switcher, with input HDMI cable (2m maximum cable length), output via HDMI>Fibre transmitter, via OM3 fibre, to Fibre>HDMI receiver.

Conclusion

The proposed projection solution, with playback and switching hardware fulfills the video specification, for 60Hz/60Fps 4k DCI playback at 10bit, 4.2.2.

Delivering 10Bit is a working reality, but it does currently require some ATI setup, and images on the projectors look stunning in higher bit depth. Lightware is providing a solid 4K switching and transmission solution.

The disguise VX4 playback load on DPX is fairly high - while playing 2 layers (4096x2160, 60fps) of DPX was successful, further compositing dropped frames. Testing 10bit NotchLC however suggests a lot more throughput at 20% of the DPX size. The team present at our test was not able to pick out the difference between NotchLC and DPX in moving video.

Thank You

Felix, Malis, Joe, Pal, Joao, Peter, Richard, Thomas