Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

NAB 2013..

Gavin Greenwalt said:
Medium Format + Lightfield + Huge color gamut + HDR (native) @ high FPS so that you can change your motion blur/shutter as meta data by blending frames.

Then exposure, focus, depth of field, framerate and motion blur are all meta data. That means that everything would be meta-data. A true RAW workflow. No more NDs. Just rely on the HDR. No more special effect filters, just simulate pro-mist etc based on HDR data. Want Anamorphic bokeh? Add it in post. Want anamorphic lens flares? Simulate it in post from the HDR data. Want to look like Kodak Vision 3? Use the huge color gamut to perfectly emulate it. Want 48fps? Pull out the sub-frames which add up to each 48fps frame. Want 60fps? Pull out every nth frame for 60fps.

Eventually add extreme resolution and multiple-camera interpolation and you can reposition and zoom.

Just strap on a wide angle lens and set your zoom in post. Need to get closer? Reconstruct 3D Geometry from the lightfield and dolly the camera in and out in post as well.

Next step after that: Volume capture. Super dense 3D point cloud of the whole set.
Click to expand...

So pretty much just ruin every aspect of the filmmaking process and turn movies into fancy video games? Yeah, no thanks. Autofocus was bad enough.
 
...
p
Josh Beadle said:
Heck, I'll be very happy when RED delivers on NAB 2012 announcements like the Leica mount and a Dragon sensor. The new sensor would rock if it has much, much lower noise, 3+ stops better low light performance, real-time black shading and look around at 5K. Throw in a chip set that does not require the current cooling system and related turbo fan and we gots some happy customers.
Click to expand...
 
Jannard said:
NAB 2013 will be historic... mark my words.

Jim
Click to expand...

Historic would be if you buy back Oakley and make 5,6,7 ( or whatever K's you fancy) eyewear...thus destroying the GoPro market and everything else ever thought or created up to now...
 
john jeffreys said:
So pretty much just ruin every aspect of the filmmaking process and turn movies into fancy video games? Yeah, no thanks. Autofocus was bad enough.
Click to expand...
Some of what Gavin has written is either BS or not yet possible. However, if lightfield sensors can look good as conventional sensors, I can't see why you wouldn't use lightfield sensors! Think about the peripheral advantages, such as having lenses with only one aperture and optimized for that.

I'm not sure about reconstructing 3D geometry from the lightfield, though. The assumption is that resolution will not be affected even if you can do that.
 
Free Dragon upgrade to the first 100 guys to show up with their Scarlet.

I remember running first thing in the morning and being in the second group for the viewing back in 2011.

Good times!!
 
James McLellan said:
Looks like I'm going to be at NAB this year. First one for me. Hope it's a good one!
Click to expand...

See you there James :)

Still wondering if there are any REDEducation options during the week.

Would like to fill my days (heck even nights) with nothing but RED :)
 
I am gonna be flying straight from venezuela to las vegas baby....wanna meet the great team and the loyal member of the family "us the red users" :)
 
Antony Newman said:
"You have no idea...

Jim"

What about:

+) (guess that) current EPIC ASICs are based on 45nm detail size. In 2010, 32nm's were out, now 22nm are possible - so 4 times the density as the current EPICs boards. This could mean 1/4 the number of circuit boards, 1/4 the power and even less weight.
-> Space could also be used to install a mini fridge on the back of then sensor (eg to keep it at 5 degrees C).
-> Or for 64 Gb or RAM so super high frame rates can be attained ( 450fps in 6k, 2000 fps 2k)
-> or a 3 min puffer battery so you can switch Bricks
-> maybe space to put an ARM core / GPU to reduce black calibration down to under 1 minute?
-> maybe space to put an internal electrically selectable CTB / IR filter?


+) Option to have the internal brain completely waterproofed - eg a Liquipel coating as this would completely weather seal it.
-> further option to remove the fans and for liquid cooling to be used.
-> Option for a red riser, base plate, top plate, and outside surface of camera to be Graphene coated (amazing thermal conductivity) and used as a heat spreader / heatsink for the heat pipe on the back of the sensor.


+) option to replace the fan with lower with asymmetric blades (quieter)
-> with a golfballed (dimpled) funnel on the intake and outlet (to better match the air impedance, make the fan more efficient, and quieter)
-> optional top fans that suck the air (so front fans don't have to spins so fast)
-> option of external fridge/silencer than cools the incoming air down to 5 degrees C ( lower sensor noise), and fridge could connect to heatpipe on sensor to add thermal mass, so fan would not be needed (for extended periods) ... could be called the 'Red Charger'!


+) Create a new cable that plugs in the DSMC and allow relocation of this handle to the right grip (eg a clutch)
-> create a new designed handle to put on the end without the Redvolt for a more ergonomic right grip (on a clutch)


+) Upgrade the Redmote to use 802.11ac for greater range lower dropouts.


+) Create a compact wearable (stereo) video & (stereo) audio receiver compatible with the Meizler module and clips to your belt. Work with John Atkinson (JH 16) to replay 'perfect' audio. Work with MicroOLED to create goggles with their 0.61 inch screens on each eye (5.4 mpix / 100,000:1 contrast / 0.2 watts) -> this would be a step up on 3d monitoring
-> versions available with only one eye active
-> wired version available in place of EVF
-> option of z distance of superimposed camera controls so eyes focus does not have to change
-> option for Low power light to shine into one of the eyes pupils to determine where the eye is looking. This information could to be included as metadata, or used for GUI enhanced options (eg selection or focus).


+) Replace the carbon anodes in the Redvolt with Tin for a (circa) 3 times power density. Replace (at a premium) with Silicon for a increase of 8-9 times.
-> Also possibility to replace Lithium ion with Lithium Ion Phosphate (4 times greater capacity)


+) Introduce a 'Red Eye' concept in which at most two sensors can been connected by thunderbolt connectors to the Brain (1.25 GB a sec down).
These eye modules would only compose of a sensor, lens activation circuit, cooling circuit and transceiver.
-> Red brain could go in back pack - and connect to the two 'Red Eyes' by a two TB umbilical chord.
-> weight of eye could be as little as 750 g - excluding lens.
-> optional power cable if not bus power-able over TB
-> Lens directed info could include FIZ data
-> New Red lenses with inbuilt digital motors would be the perfect compliment.
-> Small width of Red Eye could allow two lens next to each other (3d rigs)
FIZ data for rig to maintain 2 x lenses at required distance apart.
-> First Red Eye to use 32Mpix x 5 micron FF sensor (Dragon tech)


+) Create large pixels border around the outside of the sensor that is patterned with 32 pixel colours, each one tuned for 1/32 of the visual spectrum.
-> use this data to (where possible) accurately guess the white balance.


+) Create an alternative Bayer pattern that is R G B and X. X is not in the visible range of the spectrum, and is a wavelength that can be efficiently strobed, and at lowish levels in daylight. This pixel would trigger after reaching a treshhold. The time until the treshhold was reached would be then value stored for each X pixel.
-> Strobe used on camera that flashed light in sync with frame rate - outputting energy at pixels Xs non visible frequency.
-> X data could now be used to record Radar like 3d data for a single lens.
-> position of strobe could be placed of a regular flash for retrospectively adjustable flash (one up on Lytro ?)


+) Meizler FIZ receive options to control to Kessler Fusion (remote control Crane or Dolly via Red camera).

AJ
Click to expand...

Geez Jim, you should hire this guy and stick him R&D! ;)

M
 
Antony Newman said:
"You have no idea...

Jim"

What about:

+) (guess that) current EPIC ASICs are based on 45nm detail size. In 2010, 32nm's were out, now 22nm are possible - so 4 times the density as the current EPICs boards. This could mean 1/4 the number of circuit boards, 1/4 the power and even less weight.
*snip*
Click to expand...

Fun post to read. Good stuff. Can't wait for NAB.
 
You must log in or register to reply here.
Back
Top