What is the native ASA for the new MX Sensor?

4

4knewave

Guest
Hello,

I have shot with the RED One M sensor always at the recommended 320 ASA. Outdoors I have always found myself usually using a IRND .9 or 1.2, and for darker scenes having to boost the shutter or add more light. I have seen footage and read a lot about how the new MX sensor is great at ASA 800-1000. Also that the most DR comes from 800. Though what about shooting in daylight or a bright interior? I can see the 800 helping you get a nice stop on your lens inside but 800 outside on a sunny day is like a 64 reading on a light meter. Though with the added headroom with overexposure at that ASA I guess dropping in a lot of ND's is worth it. Though is it grainy for a outdoor image at that ASA? Also is there a lower one to shoot with for outside or essentially is 800 the new 320? I would like the most DR and information I could get in my highlights so I guess 800 with a lot of ND would be best. Has the new sensor fixed the IR problem?
 
If you set the camera to 400 ASA and switch between RedColor and RawView, you find that the levels don't really change when looking at the signal on a waveform. This suggests that the M-X sensor is around 400 ASA if you want to expose an 11-step chart so that on the RAW file, the middle grey square is at 50 IRE.

But because the sensor is so much lower in noise, you have extended shadow detail that is usable, allowing you to expose the RAW image lower so that you get extended highlight information. For this reason, most people are finding that 800 ASA is optimal for the M-X sensor in most situations; at 400 ASA you have to be more watchful for clipping, but at 800 ASA, overexposure to white looks more graceful and film-like.

I think the confusion is that people think that you should expose an image so that the RAW file looks normal in brightness.

If you look at a Log camera like the Genesis, a correctly exposed image has white (Zone 10) at 70 IRE, not 100 IRE. The Log image is supposed to be a bit down in the highlights. Same goes for RAW, as long as you have a clean signal with minimal noise. The old Red sensor was noisier and thus forced you to expose in the 200 to 320 ASA range to keep it clean-looking. Now that isn't necessary. In fact, the M-X sensor is so low in noise that shooting in 3200K looks OK now too because the increase in the blue channel noise is not objectionable even at high ASA ratings.
 
The Mysterium X sensor is more sensitive than the original Mysterium sensor and also has less noise. I always set exposure with my RAW histogram. It's somewhat difficult to nail down the which ASA matches the histogram most accurately due to the fact that the non raw histogram is affected my many other factors including white balance. I would say that a properly exposed image at ASA 400 at daylight color temperature would give you the greatest dynamic range because it would generally match your raw histogram fairly closely. However I would definitely check my raw histogram to check for any clipping. If you clip your highlights , it is a hard clip and there will be no detail in the clipped areas. But if you create as thick a historgram as possible then you will have the greatest dynamic range possible. These cameras gain dynamic range by increasing the signal to noise level. So it is generally a good idea to expose for the highlights, without clipping. The extra dynamic range comes from being able to reach into the shadows. You can definitely shoot this camera at ASA 800 if you are creating a thick histogram, and it is a good idea to buy a little safety in bright situations, but in low light situations where you may have a very thin histogram that is no where near clipping, you may be underexposing the chip and would do better to rate the camera at a 400, forcing you to light it more to keep your image out of the noise floor, while still keeping it looking right on the monitor for the client who might think it looks over lit at ASA 800. The danger in rating the camera at ASA 800 is that you are already underexposing the chip slightly, add to that using 3200k lights, and any underexposure that may occur, and you'll find that the image can still get noisy. This chip is cleaner than the old chip, but it's still important to get as thick a historgam as possible without clipping to get the cleanest image possible.
 
You gain more highlight information by underexposing so basically a less noisy sensor allows you to underexpose more for highlight protection.

So at 500 ASA, the highlights would be similar between the old and new sensors but the noise would be lower on the new M-X sensor.

There is also some improved highlight detail due to the FLUT controls, the new color science and transformation / transcoding -- basically more sensor information is available for use in the final gamma format chosen, like RedGamma.
 

Network Sponsors

Back
Top