Welcome to our community

Be a part of something great, join today!

Register 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 :)

Flare or High contrast causes sensor split glitch

V

Vicky Marie Marks

Well-known member
Joined
Feb 24, 2017
Messages
85
Reaction score
39
Points
0
Location
California, USA
I saw some social media posts about people seeing a line in the middle of the image and one side having a slight color/brightness shift under certain circumstances.

I didn't realize the raptor sensor is two sensors glued together, like alexa LF.

I did some tests and I can see the image problem they were describing.

If you have a bright flare or strong direct light only on one side of the frame, you get a darkened vertical line directly in the center of the image, from top to bottom, and there will be a slight brightness/hue shift at this seam in the middle on one half.
its like the image blend they are doing between the two sensor halfs breaks under certain lighting conditions.

easy to test, use a wide with the focus thrown and shine a light. Didn't matter for me RF or PL lens - you can even see it with no lens at all - so its not optics, its electronic.
Resolution format, frame rate, shutter speed, compression - no effect, all produce the error.
I see it on the sdi feed and in the r3d, full resolution, new debayer v 54, or old.
camera was shaded and all green

if your shooting something with flares or big shifts in brightness on different sides of the frame you might want to wait until this is fixed. Sensor stitching probably done in the camera so I think once its in the r3d, its in there.

hope its just a firmware update to fix the processing and not a hardware fix and have to send the cameras back. don't shoot the messenger
 
This is interesting. Do you see it in RCX or post when you take the card out of the camera and offload or do play back on a computer, or only in camera.???
What serial or Firmware when did you receive you camera.?? I didn't see any mention of image correction in the latest update.. Wondering if there are differences between the first ones made or later.
If its only in camera it might be something to do with the Tools. Focus or Peaking turned on.
 
Its not only in camera, its in the R3d as I said both in the current rcx and previous with the older debayer style, I tested with 1.1.1 firmware, black shading after the update, enabling/disabling peaking/zoom/etc camera tools has no effect. I'm trying to film bobcats this weekend, don't have good internet access to upload, I'll do some more tests and post some images when I get back.

In the mean time, Daniel Bender was the person who originally posted about the problem, I don't know if he is on here.
Here is the samples he posted that show the problem.

Didn't have time before I left, and I just have my laptop with me, not practical to do testing here, so I don't know if that "seam" is always there in everything, like if you push the image in grading will it appear or if its only there when there is flare/bright source on one side which makes its visible.

Jonny from Red replied to Daniel, so I'm sure Red is aware of the problem.
 
It does look like CMOS smear to me. If the light source is that out of focus bokeh thing (in the examples), it might just be a smear. It’s apparent on any previous camera and is well better controlled ever since Komodo.

curious to see more tests
 
not an engineer.
but smear would be in the darker area adjacent to a highly clipped hotspot, that is not what is happening here.
the line is always in the exact center of the screen. look at the physical sensor and you can see a visible seam in the same spot.
 
Ok a little update.
I just checked my Raptor over and over and over again with a window with a darker interior and a very strong maybe 600w light in the back of the darker room similar conditions to the clip from Vicky, and I cant duplicate the issue no matter how hard I try. 24-70mm 2.8 Canon, Otus 55 1.4, 19mm 2.8 Leica, 15mm Canon Fisheye. I took it out in the full California sun and shot with a 15mm up in to the sky in to a tree with blue sky and the sun in the middle, on the edge top right and left and in the lower corners and middle sides. No line or affect.

Im guessing you have already contacted RED? and if not contact them or go to the web page and fill out a service form to start the process.
I know not much of a solution but it seems to not be as noticeable or there at all if its not a back lit flair sort of lights situation.. I guess try to avoid those situations until you have it looked at or fixed by RED.

I have no idea and have only seen something like this one time before and it was on my Monstro very early in its life, and it ended up getting serviced and fixed at RED and never saw the issue again. But that center line issue on my Monstro was quite a bit different. not like this.
You have been able to duplicate it you said even with out a lens??
Not sure but good luck.
Im sure RED will sort you out.

One last possibility is it could be some sort of Gate flare from the side of the sensor gate or port. Look up Gate Flare on the web and it has a similar look and that light on the right showing up and then it kind of disappeared.
 
I have tested on my Raptor:

Night time in a room, using a iPhone torch directly on the lens and move the lens to find if there is a line in the centre,
1. Turn on room light, almost I cannot find and line in the centre.
2. Turn off room light, I cannot notice the line when I am using a Sigma 35 T1.5 lens. When I am using a Tokina 11-20 lens and move the zoom to 11mm, I can see a line when there is some angle of the torch.

Day time outdoor, shooting an object against the sun.
I cannot find any line here.

It looks the issue is here but very rare to represent in reality. Will see how RED will deal with this issue.
 
Thanks for testing Russ and Jie. I had to look up what gate flare is. It wouldn't be gate flare as the line is always in the center, but I guess it could be something similar, light bounce inside causing the seam on the sensor to cast a shadow or something? The area surrounding the sensor is a little shiny. I'm heading back to the city and will do some more testing tomorrow.

Here is another example photo someone posted on facebook.
 

Attachments

  • photo112559.jpg
    photo112559.jpg
    77.4 KB · Views: 3
Find an easy way to show the line but it is not common shooting condition. Do not mount the lens, make the iso to 250, slowly use your hand cover the lens mount, then you will see the line easily. Very easy.
 
It's a known issue/artifact. Same as the Arri LF/Mini LF. No anything gate flare or CMOS smear related, Raptor actually shouldn't exhibit CMOS smear as you've all come to know it btw.

It's optics related. The more interesting thing regarding this is it's not always present and even some lenses seem to not produce the seam.

Also only occurs at very specific angles. Example, sometimes at common distances a car pulling up to the camera won't exhibit the artifact, but 3 feet away with a flare off axis can.

Pretty interesting discussing this for the last several years among various DPs.

It's one of those occasionally an issue things that gets a little stickier with the variable of light source, light distance, and lenses at play.
 
Definitely not CMOS smear, that's generally a horizontal line from pixel highlight overload. Almost all cameras have that or have contained incredibly well like the V-Raptor.

This is something different to me in the image captured.

Mini LF / LF uses the same sensor stitch to get to Large Format as V-Raptor. Monstro, and Dragon VV, I don't believe did at all. Sony Venice, FX line, I don't know if that's how they achieve Large Format sensors. I've generally only used Venice in low light and never noticed it.

I've been a dual Monstro owner and never saw this. Currently, own two Mini LFs and cannot reproduce this issue like the V-Raptor ( haven't seen it in two years of use ), even though you can see the two sensors in the MiniLF as well.

It may just be a sensor calibration issue at the factory or a sensor batch mismatch. Below is my test clip on the V-Raptor. I could upload a MiniLF clip as well, but it doesn't have the vertical line on either one and I'm not out to make it Red v Arri. I've also owned many more Red cameras than Arri over the last 10 or so years.

Link to V-Raptor Sensor Issue : https://www.dropbox.com/t/EPEXGTDKUGRNZbsT
 
I took a picture of the V-Raptor sensor today. It definitely has a line in the middle, but is very difficult to photograph. I was defined very surprised…
Then I’ve checked my beloved Monstro and found the same line on its sensor. Not as strong as on Raptor but it’s there! Never had issues on Monstro, though.
 

Attachments

  • 70DA5B0B-7481-40E7-B3D4-4AECDAA6204B.jpeg
    70DA5B0B-7481-40E7-B3D4-4AECDAA6204B.jpeg
    120.6 KB · Views: 1
You must log in or register to reply here.
Back
Top