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 :)

Epic Gradient

From what I understand about BSC, please careful, very non-scientific.

HOW SENSORS CAPTURE DATA...

The pixels on the sensor need a small amount of current, in order to function. When photons hit the sensor pixels, they create a charge in the pixel, per unit of time. This is the exposure. The amount of this current per time unit is then written as a digital value. In case of the 16bit EPIC it is 2 to the power of 16, so between 0 and 65 536 ...

If few photons hit the sensor, they create only a small charge, that is very close to the "pre-charge". The pre-charge might confuse the electronics, as how is it supposed to differentiate between the current that is leading into the pixel, and the one that the 3 photons generated when "someone" did something amazing and called it 4K timelapses ... :-D

This is where the BSC comes in. It measures the pre-charge value of each pixel. That's why the lenscap must be on, as it creates a matrix, on what the sensor sees (which is analogue noise) and what was in theory supposed to be black (that's why it is wise to put the lenscap on the camera AND put it into a dark place). Presumably, during recording, the BSC matrix is then subtracted from what the measured sensor values to arrive at the "ideal" theoretical "perfect black"...


There are 2 important values, that influence the quality of the BSC: exposure time and as we now see temperature.


EXPOSURE TIME...
With long exposure times, the pre-charge current is firing into the pixels. If you BSC at shutter 1/48 and then go to 1 sec exposure, the camera is expecting a pre-charge per pixel for only 1/48 of a second. This means, that if the BSC matrix is subtracted from the pixel, it is not sufficient over the length of the exposure to understand what might be happening to the individual pixel.
This is the reason why you get hot pixels when you do not BSC for long exposures.
It seems, that as on the photon level, at the given pixel size, all sorts of leak currents are raising up the digital noise, making it harder for the electronics to understand what is noise and what is the picture.
It also seems that some pixels react differently to different pre-charge current times.
Ergo, one has to BSC for a long exposure, because this measures how the pixel behaves over the length of the exposure during the shot later. -> the BSC matrix is equaling over the proper period of exposure, and if then subtracted from the frame, it bring back the perfect blacks.


TEMPERATURE ...
Now as we talk about those minute levels of current, we are on the pixel level in the analogue realm. Of course, temperature, that e.g. expands the sensor, and influences the conductivity of the conductors between the pixels has an impact on the BSC measurements.
If temperature X, the pixel pre-charge is y, and at temperature x+10C it is y+5, because the temperature raised the conductivity of the circuit by that amount, the BSC matrix cannot be accurate...


It is the exposure time, not the fps that determine the BSC. One can actually shoot at different fps, and keep the exposure time to have a dead-on BSC (for geeks only).
Ok, these are so small nuances, we are talking about, so to put it back into perspective:


BEST PRACTICES (disclaimer)

1. Perform BSC when camera on temperature, and on the temp of the shooting environment (approx. 30 mins preptime). Calculate this into your shooting shedule. Do it every day before shoot.
2. When going to less than 24fps, with longer exposure times, BSC IS necessary, to avoid artifacts.

For the nerds:
3. Try to avoid powering the cameras down, to avoid the sensor temperature to drop.
4. When going from 24fps to 300fps, it appears the BSC-recalibration is not need.
5. I would SUGGEST that everytime your exposure time doubles (from 1/48 upwards) from the previous with BSC, you need to re-calibrate.
6. Write down the sensor temperature at which you made the BSC.
7. If your environmental temperature changes, and you see the sensor temp differ more than 10 degrees C, consider a re-calibration.
8. If you are in a harsh environment, do not worry about the BSC too much


I am performing those tests on a glacier, with sun coming in and out, with cold winds and snow and temperature changes between 20C.
I perform BSC once camera is on temp once a day. I shoot at 800 ISO.

I sleep well.

:-D
 
@Jonas: So take this scenario and please give me a best practices list:

Let say I . . .

1. Shoot exotic Travel, Lifestyle and Wildlife Stock which is always in the field away from constant power
2. Use only REDvolts which last 20-30minutes. The Epic is switched off between shots to preserve limited battery life
3. Hike, climb or fly to get some shots so kit weight and size is critical - big heavy batteries are not an option
4. Get shots by moving along until I see a fleeting something that catches my eye - then bust out the cold, warm or hot Epic to get the shot
5. Commonly switch Frame Rates which in turn changes Shutter Speeds
6. Shoot in cold, warm and hot conditions . . . within the same day

From what I understand (and have experienced) this type of acquisition is not going to be compliant with the current build Epic BSC needs. If if takes 30-60 minutes to properly BSC in order to snag a 15 sec clip of fast moving wildlife . . .

It would be helpful if the very smart engineers at RED could chime in to set the record straight here.

Thank you in advance!
 
X 382, fw 2.0.5, started at room temp, no gradient up to 12800 visible in both touchscreen or Bomb.
 
Michael Mayda said:
Christian, was your testing done immediately after camera startup or after some minutes of warmup time? What firmware version? What is your camera, M or X and serial number or date you received it? Just curious about potential differences in camera builds.
Click to expand...

I did it twice. First immediately after startup and after +/- 15 minutes of warm-up time. No gradient or if is there it is not visible at all…. Happy to do more tests if someone need it.
Epix-X, Beta 2.0.5. serial number around 14 hundreds.
 
Josh,

let me try to suggest what I would do, but this might not work for you at all. I have right now a pretty similar setup and have to work around its limits, in contrary to a commercial shoot a while ago, where I implemented the "perfect" BSC procedures - we were also shooting at 1:3 ratio ... ;-D.


1. Your shots are more important than perfect a BSC (remember that you only see it, when you crank up past 2500 ISO AND the calibration is off) - this is the nature of your work. There will be times, where your BSC will not be perfect. But I rather have a non perfect BSC of the one-of-a-lifetime snow leopard shot, than none at all. Most importantly, you will almost never have this issue in real world, unless you BSC is really way off.


2. The sensor temperature, once the camera is hot, does not vary as much as the environment temperature. E.g. When I shot at -5C, the temp of the sensor was around 45C. When the sun came out, it got up to 5C, the sensor temp was still around 50C - in my book no issue, when BSC done at temp of 45C. I guess it might get problematic when going from -15C to 15C, but this is an extreme change, and I think one has to account at some point that equipment has its limits... same as with batteries and lenses at sub-zero temperatures.

3. You say, that you pull out the camera and shoot, the second you see something. I find myself however very often get into a position first, and then wait for something to happen, e.g. when observing birds, or wildlife. Very seldom, one is truly surprised by something happening - as one is constantly aware of the environment and is predicting things to happen - its not only the camera that needs to get ready, its the lens, the tripod, etc... So what I do, is that the first thing I do is to turn the camera on, then setup tripod, etc... this gives me 1-2 minutes of camera on, till I get ready. ... and when really the aliens arrive out of nowhere ... I will shoot their spacecraft, rather than worry about BSC :-D

4. Setting the camera on 25% fan manually lets it heat up quicker... :reddevil:

5. Consider getting more batteries, even when they are heavy... :-D Consider big V-lock batteries, rather than the redvolts. I worked for a whole day on 2 batteries only. I had a strict power saving regime, but I also gave the camera the extra 3-4 minutes to get into the BSC "zone". I understand that big batteries may cause problems with airlines.

6. Keep in mind, that with a cold camera, every minute you can wait makes a bit of a difference, with the first minutes being most "crucial".


At the end of the day, I would say that one can only try as much as possible with that setup for a healthy BSC, as is focus, as is exposure, as is everything else out there, when you are in the wild.
It will never be as convenient as on a studio shoot, or big location shoot. What matters though, that even when less than 100%, the epic delivers images no other camera can. I will live hundred times with a gorgeous 5K image, even when i knew the sensor temp and the BSC was a bit off.
I bet every other camera would have that same issue - but noone looked into that (how can I blackshade the alexa, btw. ?).

Its time to leave the nerd at home, and let the filmmaker take over, who can shoot "anything everywhere" (TM) ... now in 5K.


Can't wait for some of your samples.
 
Tom.Wong said:
do you absolutely need to black shade to get rid of it, or if you just leave it on for 15-20 minutes it's gone? because high end cameras need to reach optimal temperature before it hits peak performance. what you're seeing from a cold boot could be from that. all of it works in a sweet spot, temp is part of it. black shading is another part of the equation for optimization.
Click to expand...


Word up
 
Rick Darge said:
Then why didn't the Red One ever display this issue? That camera is high end...
Click to expand...


can the r1 do 120 fps at 5k?

or 5:1 compression and below?

or shoot at 5k at all? things to consider, lot more parts, lot more processing, crammed into a tinier box. same sensor, used differently, optimized differently for different tasks. they aren't the same camera, and it took years for the r1 to get to the state it's in now. so it's probably gonna take awhile for RED to fine tune the firmware to get the hardware in harmony a certain way. faster black shadings and better optimization on a cold boot for starts. but the question is still if the gradient clears up after 10-15 minutes of run time? if it does, the clear cut work around now would be to turn on your camera earlier than you were going to, and let it warm up, or just don't turn the camera off and drain batteries all day and keep it running. R1 days you pretty much kept the camera on as long and much as possible cause of the boot times so it might just be the most cost effective workaround for now until better FW comes out.
 
@Jonas
Man do I appreciate your efforts on this topic. Only thing better would be if your avatar name was in red . . .

I'm getting a strong gradient on start-up as reported at high ISO, even at 5600K. After a 15 minute warm-up the noise evens out but then there is a weird flickering on touch screen. After a 10 minute cool down - power up, gradient is back. In your list above you don't specify when doing the BCS either "Default" or "Current FPS/Exposure" At times I change up FPS a bunch during the day.

Epic-M, Build 2.0.2
 
@Josh
It's not like I did not apply for a Jonas Rejman ... but the selection is tough, hehehe

I am on build 2.0.5. Last posts on this topic from Deanan (before his leave from RED was announced) was that there are 2 settings for the BSC. A default one, and the "current fps" one. It does not seem to be switchable at the moment though. I always perform BSC at "current fps/Exposure" .. after I set it to my desired FPS/exposure.

I also see a flickering and noise on the screen when checking BSC this way. Remember, that this settings are only good for one thing - to judge the uniformity of the BSC. We are super-pushing the blacks to see deep, deep into the noise of the blacks.
Weird behavior of this "image" is expected.


I think that RED is working on the possibility of creating an array of BSC matrix settings. Different exposures, different temperatures, etc ... over the lifetime of the EPIC one will maybe collect about 10 BSC settings, maybe they will even switch automatically when the temperature of the sensor changes ...

... thinking further, maybe this will become part of the factory burn-in ... The cameras will be deep-frozen at -30C and then the BSC will be performed every 10 degrees ... with 10 different exposures up to 60C ...

These setups will then switch automatically during the shooting, so that if "Josh's last snow leopard on the planet" decides to jump on Josh's back from behind the treeline and Josh will be his dinner, whilest shooting it ... at least the BSC will be perfect. :ciappa: :crazy::crazy:
 
First things first -- is everyone on the same build here (2.0.5) including those who have and have not experienced this issue? Some reports say it's fixed, others not. I'd like to know which is it, as I am sure a lot of others would. :beer:
 
I'm on 2.05. The build hasn't changed the situation. Joesph Hutson, so you're telling me that you do NOT experience this issue on your current M camera? The best way to 'see' the issue is to put your camera in false color mode. The gradient will appear at 2000 iso and above. Apparent with false color turned on. I frequently shoot at high ISOs in situations where I cannot wait for the sensor to warm up. Red, please make a COLD black shade setting to offset this issue that way we can flip it over to the regular warm setting once the sensor reaches that temp. I'm just trying to gauge how many other Epics experience this. Deanan said this issue was normal of the Epic sensor but that's troubling and will have to be accepted and worked around unless Red can come up with a better solution.
 
@Rick: if RED did provide us a "cold" BSC would that even work? The sensor is only cold for 12 minutes +/- so as the camera, which was BSC at a cold temp, warms up to operating temp the problem could be in reverse for most of that warm up time. Not sure this is a fix. What about the "X" is different from the "M" to create a difference in BSC or temperature sensitivity? Anyone know who is the go to person at RED for BSC questions is now that Deanan is gone?

Edit: would also like to verify that this phenom is an actual shooting issue (as in it effects the shot footage) - not just a freaky thing that happens at ISO2000+ in preview mode. I chased a "crushed blacks" issue only to find out it was a preview bug in REDcine that has since been fixed.
 
Last edited:
Jonas Rejman said:
@Josh

I am on build 2.0.5. Last posts on this topic from Deanan (before his leave from RED was announced) was that there are 2 settings for the BSC. A default one, and the "current fps" one. It does not seem to be switchable at the moment though. I always perform BSC at "current fps/Exposure" .. after I set it to my desired FPS/exposure.
Click to expand...


You can switch between the two different BSC. I've been doing it for over a week now. I have the user setting set for 1 second exposures so I don't get any hot pixels shooting timelapse. Then I have the factory setting at normal 24fps 1/48th. Sometimes the switch needs a power cycle to take effect.
 
You must log in or register to reply here.
Back
Top