I apologize in advance for the long post. For those who don’t want to go through the whole thing here is the Executive Summary: My back does not apply dark frames in a way that makes any sense so I am going to talk to my dealer about submitting a case. For example, when in normal camera mode and zero latency I see hot pixels even though it takes a dark frame in that mode. I suspect this problem is lurking in the background of all these results and this whole thread, so I doubt it is just my back.
The Rest of the Story
I haven’t posted recently because of a combination of 1) travel, 2) fear this thread would go exactly where it went for a while, and 3) the effort it takes for me to do something that should be simple! I got pulled into this early with the mistaken idea I could help with something, but alas, not so much. I’m in a real conundrum because I am far in the camp of this application being not just out in left field, but beyond BFE. It’s like Yunli is complaining that the air conditioner in his 911-R goes out if it is in Track Mode for over 4 hours. On the other hand, he has done a good job demonstrating that for him, this is a very important application. I certainly understand laying out $four-figures$ and then finding out the product might not perform the way you wanted it to in an application that is important to you.
Yunli, there are a couple things you might consider when approaching these issues: When you post images in these situations, also report what you are doing in C1 to highlight the problems you see (in this case shadow and exposure settings). This would do a few things:
- It would make it clear how others could reproduce what you are seeing
- It would diffuse some of the aggressive responses because it would be clear how far you are pushing things when seeing these results. Many of us in this community are tired of the constant bashing, and it is too easy to dump this into that category. Perhaps not justified, but for a community to work we all have to be considerate of each other.
(sorry for the lecture).:lecture:
As for my testing, I am not going to post too many photos because there is really nothing new to see. Last week I sent Yunli some of my raw files, which he agrees looked fine. However, as I’ve posted previously, taking a very short initial exposure is a big part of the issue here because it forces the back to use an irrelevant dark frame. Yunli and I disagree about how cumbersome it is in the field to take an initial exposure on the order of minutes, but I bow to his experience in these types of shots, which I really never do.
I can duplicate the red cast by taking short initial exposures then a long exposure and pushing the exposure and shadows. I moved to the basement to get controlled lighting conditions and because the temperature is a consistent 67-68d F down there. Two weeks ago I did a first test and got weird results that included images shot in normal camera mode and normal latency that looked like there was no dark frame applied. I sent those images to Doug just to let him know. I tried to convince myself I screwed up my notes!
So I focused my testing on different modes and situations to see if and when it applies a dark frame. Again, my results are not at all what I expected. Because of the results in my first test, I was suspicious of what Doug mentioned above regarding sync errors, so I used the following sequence:
- 277 Normal Mode – Normal Latency
- 278 Aerial Mode – Zero Latency (no dark frame)
- 279 Normal – Zero
- Turn off the back
- 280 Aerial Mode – Zero Latency (forces a dark frame)
- 281 Short exposure
- 282 Normal – Normal
Please note as I mentioned in previous posts, it is not switching to Aerial mode that forces an initial dark frame – it is instead powering down and back up. Here are the SOOC images, purposefully underexposed.
From now on, all images are pushed 100% shadow and +1 Exposure.
I was going to continue testing and force a sync error, but my results at this point were strange enough that I stopped. Because I’m taking long dark frames (not 1/60 sec like Voidshatter), you can see the red cast when pushed but it is more subtle; you need to compare images to see it. The more interesting result is how dark frames are applied given the above sequence. If I see hot pixels I assume either a generic dark frame was applied or no dark frame was applied:
- 277 Normal – Normal: No Hot Pixels
- 278 Aerial – Zero: No Hot Pixels(!)
- 279 Normal – Zero: Hot Pixels(!)
- Turn off the back
- 280 Aerial Mode – Zero Latency (forces a dark frame): Hot Pixels(!)
- 281 Short exposure: No Hot Pixels
- 282 Normal – Normal: No Hot Pixels
Sorry, the hot pixels do not come across very well in the JPEGs, but you can see in the 100% crop the reddish shadows; those are the two with hot pixels.
I cannot explain the use of a dark frame on images 278-280. 278 did not take a dark frame but there are no hot pixels. Both 279 and 280 took a dark frame but there are hot pixels. I know what you all are thinking: I again screwed up my notes. Of course I thought the same thing, so I repeated it again this morning – same result! Here are screen shots FF and at 100%.
- 549 Normal – Normal: No Hot Pixels
- 550 Aerial – Zero: No Hot Pixels(!)
- 551 Normal – Zero: Hot Pixels(!)
- Turn off the back
- 552 Aerial Mode – Zero Latency (forces a dark frame): Hot Pixels(!)
- Forced sync error
- 553 Normal – Normal: No Hot Pixels
It appears to me the back is not applying dark frames as it should, and it is more related to zero latency than it is Aerial mode. I do not normally use zero latency. For those of you who do, you may want to investigate.
Dave