FROM STEVEN SODERBERGHS TWITTER
If you mean trimming an 8K Monstro die that was rejected because of edge defects... most likely not. You'd have to account for re-routing all sorts of things like power, control, and I/O, so just trimming a sensor isn't likely to work well.
Red's probably re-using the pixel design from Helium, Monstro and Gemini... not to pretend that they're identical, but they probably all have a lot in common, because why not re-use something that works really well and costs a lot of time, effort, and money to develop?
B. They already have a mechanical platform to support it in production for Gemini, minimal development needed to fit the smaller housing.
That's also a massive oversimplification, since Gemini uses the far more power hungry DSMC2 electronics... the "minimal" part is rather unikely.
.
Not none, but certainly less -- another reason that I'd expect Red to re-use the pixel design that it developed as much as possible.
Realistically, anything over 12 stops is is plenty of dynamic range for anyone competent, but I'd be surprised if Red were satisfied with a sensor that had less than 16.
Tend to agree with you for the most part. Good points. Phil Holland mentioned that Monstro is a hot, power hungry beast.
Komodo needs a much lower power/thermal envelope. Nano component sized SOC fab has progressed considerably in just the last couple of years. That makes it possible.
My D16 is an honest 12 stop camera. I don’t have many situations where more is needed, but a couple more stops would improve the noise floor.
If you mean trimming an 8K Monstro die that was rejected because of edge defects... most likely not. You'd have to account for re-routing all sorts of things like power, control, and I/O, so just trimming a sensor isn't likely to work well.
Red's probably re-using the pixel design from Helium, Monstro and Gemini... not to pretend that they're identical, but they probably all have a lot in common, because why not re-use something that works really well and costs a lot of time, effort, and money to develop?
Hardly a surprise... Helium and Monstro generate 35.5 megapixel images with 16-bit color... that's a HUGE amount of data. That amount of computing power is a lot of computational work...
Probably also some custom ASICs and a clever active cooling system.
I was quite happy with the 13 stops I had on my BMCC. The 12 stops on my Cion were only pain because the Cion had a linear encoding, so the highlight cutoff was harsh as all hell. Getting a Helium was like suddenly having an infinite dynamic range available
Out of sheer curiosity, what is that focus wheel looking thing on the left side?
Me too I really really hope for interchangeable OLPFs !! I don't know if it's a realistic wish but if it's feasible for them to do it and they know we want it perhaps they've done it or they'll add it in
Out of sheer curiosity, what is that focus wheel looking thing on the left side?
Sorry this post wasn't about Komodo
The DSMC2 cameras can record up to 4K ProRes in camera (422, not 444). I don't do that often because 4K ProRes is as big as 8K Redcode.
