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 :)

Major Announcement...

Status
Not open for further replies.
im.thatoneguy said:
I don't know what this 12 bit = 12 stop nonsense is. 12 bits can hold 4096 distinct stops or gradations.
Click to expand...
Most people here are referring to dynamic range related to exposure in lens f-stops, this is the industry standard way of expressing it. In f-stops, for a linear transfer function, 1 stop is 1 bit, and 12 stops is 12 bits (neglecting noise & quantization error).

If you use a nonlinear transfer function before the A/D, for example a log curve then yes you can have any arbitrary dynamic range with fixed # of bits. However the Red camera design is (up to now) linear, which has certain advantages.
 
Gbabymogul said:
I'm still not sure what's 90 % feature enabled, 98% feature enabled. I was thinking of waiting to get the 90% feature enabled version (sounds close enough to me). My co-producer thinks i'm bananas. :-)
Click to expand...

I can't help you with the bananas, but I'm assuming (!) that the lacking features are software features that will be updated by loading new firmware.

It's been mentioned that the first cameras to be distributed wouldn't have all the Redcode RGB functionality. So you might (speculation) not be able to record on-board RGB (2k, 1080p etc.) at first, but you'd be able to fix this with a firmware update. Similar to how Blackmagic adds new functionality to their Decklink boards with new drivers.

In the long term Red might come out with things like board replacements etc, to improve performance or whatnot.
 
Barend Onneweer said:
I'm assuming (!) that the lacking features are software features that will be updated by loading new firmware.

It's been mentioned that the first cameras to be distributed wouldn't have all the Redcode RGB functionality. So you might (speculation) not be able to record on-board RGB (2k, 1080p etc.) at first, but you'd be able to fix this with a firmware update.
Click to expand...

Yup.

I'd like to have confirmation by the RED team, though. Depending on what features are enabled will make a difference to what gear I purchase now, or gear i purchase later. It will also mean waiting a month or two depending on if i feel the features not yet included are needed for my applications. I can't decide any of that on guessing. ;)

I'm happy with the news today about better dr etc... but i've seen this asked numerous times, with no response by the RED dudes. I just wanted to know if the board had anything to do with the 'features enabled'. Maybe they were waiting for a newboard before implementing REdcode RGB for example.

Just wonderin'...
 
Maybe the upgrade will allower for higher gain settings. The new Canon 1D MarkIII can go up to 6400 and is apparently quite usuable at 3200. Anyone seen any examples of this?
 
Great news. No problem with a delay.
 
im.thatoneguy said:
I don't know what this 12 bit = 12 stop nonsense is. 12 bits can hold 4096 distinct stops or gradations. If you're willing to sacrifice every other stop you could have 8192 stops. You could have 14 stops and 8 bits. Their relationship is completely arbitrary.
Click to expand...

Just because VHS can display the 'dynamic range' of film does not mean that VHS has the same dynamic range as film. It's just not usable.

In the same way, expressing more stops of latitude in a lesser bitspace leads to nasty quantization and generally unappealing data. Having more stops than bits will lead to a flat looking image that cannot be fully recovered with a pretty transfer curve without bringing out ugly artifacting.

Consider 8-bit HDR images. Sure they appear to have massive latitude, but it's not useful latitude. If you tried to re-express the data in a more appropriate bit depth, the wheels would fall off.

This is, of course, a brief aside to this thread.
 
Brook Willard said:
Just because VHS can display the 'dynamic range' of film does not mean that VHS has the same dynamic range as film. It's just not usable.

In the same way, expressing more stops of latitude in a lesser bitspace leads to nasty quantization and generally unappealing data. Having more stops than bits will lead to a flat looking image that cannot be fully recovered with a pretty transfer curve without bringing out ugly artifacting.

Consider 8-bit HDR images. Sure they appear to have massive latitude, but it's not useful latitude. If you tried to re-express the data in a more appropriate bit depth, the wheels would fall off.

This is, of course, a brief aside to this thread.
Click to expand...

Bit depth and exposure latitude useful or not is completely arbitrary. a 12bit image is already throwing away data from the real world. You could record 12 stops to 32 bit and get really reallly smooth gradations. Eventually however it's just sort of overkill for most situations.

All A/D conversions involve quantization and loss even the RED with 12 stops. I'm hearing a bunch of "best practices" scenarios being thrown around like it's based on electrical engineering principles. To the best of my knowledge 1:1 stop to bit dynamic range has no basis engineering.
 
You make a good point - it's something very important to be familiar with. There will always be data loss whenever something "real" has to be represented with bits. My 6dB:1 stop: 1 bit thought stems from the fact that that is the general norm for capture, manipulation and playback these days.
 
Assuming a linear system, you can't get any more detail in your digital representation of an analogue signal if the digital quantisation noise is lower than the analogue noise floor. Digital quantisation noise is approximately (6db * number of bits) below full signal. So, if you have around >66db SNR, going beyond 12bits doesn't capture you any more detail, but it can capture the noise floor smoother. And hence that's where we get 1bit = 6db from. It's actually around 6.02, or precisely 20Log10(2^n).

http://en.wikipedia.org/wiki/Quantisation_noise

Graeme
 
Evin Grant said:
Graeme, sometimes my head hurts after I read your posts. I think it's my brain growing:nerd:
Click to expand...


When that happens to me Evan I just insert a Q-tip all the way in my ear and Poke that brain until it settles down.

Holy Crap Graeme I went to film school because you get to make pretty pictures and work with hot chicks. I don't understand what your saying but if it helps me make prettier pictures and work with hotter chicks I am all for it.
 
joelnet said:
That would be surprising - I'm sorta guessing this upgrade will get us to the promised 11 1/3 stops. But if I'm wrong and it's more that'll be even scarier for Arri and Dalsa.
Click to expand...

Arri and Dalsa crapped their pants at 9:30 a.m. April 16th.
 
Arri's actually pretty excited about the camera. It means they'll sell more accessories and lenses! They were super cool about letting RED use their equipment at NAB.
 
Status
Not open for further replies.
Back
Top