Monstro Speculation Thread

[Pawel Achtel]

Doesn't make us drop 40 grand in glass cuz we have to.

I think here is the rub. Many people push for high resolutions. How may of them can afford the glass to make use of those pixels?

I personally struggle to push the current sensor or the camera to their limits using Master Primes (I am the limitation most of the time :clown2: )

Few people realise that 12-13 stop DR would be massive improvement over already high DR that Red One sensor has. Remember, this is a log scale, so 1 stop is 100% (double) improvement. 12-13 stops DR would also require at least 14 bit RAW, which would require even more bandwidth and perhaps more aggressive compression. That's a big ask for a camera which is already superb in achiving those tough requirements.

 

[M Most]

I think here is the rub. Many people push for high resolutions. How may of them can afford the glass to make use of those pixels?

I personally struggle to push the current sensor or the camera to their limits using Master Primes (I am the limitation most of the time :clown2: )

One of the lessons of the Panavision videos that people here should be paying attention to (but aren't because they're often too busy dissing Panavision) is the notion that images are not the product of a sensor - they're the product of a camera system, which includes processing and optics. The resulting image can only be as detailed and sharp as the lens will allow, even with completely transparent sensors and processing, not to mention the need for low pass filtration to avoid aliasing problems, especially with single chip, color filter pattern sensors. There is a practical limit to how much resolution is useful on a sensor being used for motion picture capture, because eventually you hit a point of diminishing returns based on the limits of the optics that are in front of that sensor and the ability of human beings to accurately hold focus.

Food for thought.

 

[cinemano]

i think as the trend goes with technological progress they can make a sensor the same size with more Ks.. a bit like how a PC chip can get smaller with years.. perhaps its a 5k sensor, same size and shape as the 4k?

 

[Robin Balas]

i think as the trend goes with technological progress they can make a sensor the same size with more Ks.. a bit like how a PC chip can get smaller with years.. perhaps its a 5k sensor, same size and shape as the 4k?

Imaging chips and PC chips are not similar. PC chips have the ability to pour more juice onto it aka huge power supply and the heat loss from increased density can be sucked off with big heat sinks and fancy fan systems.

Imaging chips suffer from the fact that there is a practical limit to how much more light you can throw at it before it becomes ridiculously expensive compared to the gain in resolution. Smaller sensors on the chip equals less photons will hit it. Then you have to lower the noise floor on all components equally to compensate. There is a limit to how much this can be reduced as this is real world physics involving light capturing devices and photons pr. area, not heat generating circuits.
Look at the high end stills photo community now complaining against the latest 60Mpx cameras as they feel there is enough resolution in today's backs, they wish for other improvements like higher ISO, fps and more flexible system integration, tilt/shift capabilities, etc.
Personally I don't see the cost/benefit factor driving reolution beyond 5K as other areas needs more attention. But in that regard I am a pure amateur so I might be way off on that one.
MHO.

 

[James T Mather]

I don't think going Vistavision or Imax sized is going to work - no industry standard lenses will work with the system - tougher DOF for drama production. Seems to be too niche in general (unless you just want to shoot landscapes)

 

[Pawel Achtel]

One of the lessons of the Panavision videos that people here should be paying attention to (but aren't because they're often too busy dissing Panavision) is the notion that images are not the product of a sensor - they're the product of a camera system, which includes processing and optics. The resulting image can only be as detailed and sharp as the lens will allow, even with completely transparent sensors and processing, not to mention the need for low pass filtration to avoid aliasing problems, especially with single chip, color filter pattern sensors. There is a practical limit to how much resolution is useful on a sensor being used for motion picture capture, because eventually you hit a point of diminishing returns based on the limits of the optics that are in front of that sensor and the ability of human beings to accurately hold focus.

Food for thought.

Totally agree. Not to mention the biggest limitation of all: the display system and the viewer :watsup:

 

[David Wyatt]

I think the 65mm/Imax idea is unlikely mostly because of the lens issue (not to mention focus pulling is already pretty tough on the Red One!!)...but there's a clue in the name of Monstro...something about this sensor is going to be quite simply monstrous...frame rates, dynamic range, who knows - it's going to be a beast!

 

[Michael Lindsay]

One of the lessons of the Panavision videos that people here should be paying attention to (but aren't because they're often too busy dissing Panavision) is the notion that images are not the product of a sensor - they're the product of a camera system...

...Food for thought.

I've recommended those Panavision videos to a number of people (it is very easy to mentally subtract the anti-red bias) and it is interesting that nobody has managed to watch them all...

what does this tell us..?

 

[Daniel Browning]

Smaller [photosites] on the chip equals less photons will hit [each photosite]. Then you have to lower the noise floor on all components equally to compensate. There is a limit to how much this can be reduced.

Yes, there are theoretical limits and limits in practical applications. But we have 12 MP sensors in 1/1.7" with quantum efficiency within 50% of the maximum theoretical limit, and read noise that is proportionately *better* than than the same area of large pixels (e.g. FZ50 vs Canon 5D).

For the last eight years, as the pixel has shrunk, the read noise has shrunk with it. Image-level noise has improved or at least stayed the same. Granted, there are certainly read noise sources that don't scale down as the photosite is scaled down. For example, the Random Telegraph Signal component of the 1/f noise in kTC caused by trapped carriers in the Si-SiO2 interface gets worse as pixel size shrinks.

I have a lot of respect for the ability of sensor designers to compensate. We can expect that the effects of their ingenuity will slow down before it stops, and it hasn't showed any signs of slowing down yet, so we should have some exciting times to look forward to.

Personally I don't see the cost/benefit factor driving resolution beyond 5K as other areas needs more attention.

Those who do not need additional resolution may use compression to give them filesizes equal to their needs. For example, RED might build a 28 MB/s mode for EPIC at 5K.

 

[Charles Angus]

Yes, there are theoretical limits and limits in practical applications. But we have 12 MP sensors in 1/1.7" with quantum efficiency within 50% of the maximum theoretical limit, and read noise that is proportionately *better* than than the same area of large pixels (e.g. FZ50 vs Canon 5D).

Not enough people are aware of this. It is a common misconception that bigger pixels = less noise. This is not the case, for a variety of reasons.

One reason is that people are not comparing apples to apples. Consider the D3 vs. the 1dsMkIII. Many people tout the low noise of the D3, but are failing to consider the fact that it has half the resolution of the MkIII. Using a good downrezzing filter, the MkIII will have equal or lower noise at the same resolution.

I guess my point is that more resolution is generally better, and it does not necessarily carry the noise penalty people often think it does.

 

[Cüneyt Kaya]

i hope for
a native tungsten balanced sensor.
a faster readout to reduce the rolling shutter effects.
100 fps at 4k to Cf cards
1 more stop of DR

this would be awesome and worth every cent for the epic