p.2 #1 · Call for help - building sensor readout speed database
Adam, I have a Sony A7Riii and the A7rV, if you can guide me I will participate. As long as we aren't shipping items to each other you know what I mean
p.2 #2 · Call for help - building sensor readout speed database
Just finished my testing with the board. A few thoughts:
1. I didn't use a tripod. As shown in the instruction, a tripod is ideal and I strongly agree. Especially during video test, it's hard to prevent the camera shaking while press the button without a tripod. I may need to redo some tests if my results don't meet the requirements.
2. Video shots have different crop factors for different modes. You may need to pull the camera back a bit to avoid the LED circle get out of the frame.
3. Somehow the first time install the IDE, my board cannot be detected. the "Select Board" box is empty. I reinstalled the IDE and then the board shows up. I'm working on a windows10 laptop.
The total time I spent is about 50 min, very close to Adam's estimate. It could be less if the board is up the first time. I also tried several times for some videos due to shaking or framing issues. Any way, it's fun to participate this project.
p.2 #3 · Call for help - building sensor readout speed database
tonychen wrote:
Just finished my testing with the board. A few thoughts:
1. I didn't use a tripod. As shown in the instruction, a tripod is ideal and I strongly agree. Especially during video test, it's hard to prevent the camera shaking while press the button without a tripod. I may need to redo some tests if my results don't meet the requirements.
2. Video shots have different crop factors for different modes. You may need to pull the camera back a bit to avoid the LED circle get out of the frame.
3. Somehow the first time install the IDE, my board cannot be detected. the "Select Board" box is empty. I reinstalled the IDE and then the board shows up. I'm working on a windows10 laptop.
The total time I spent is about 50 min, very close to Adam's estimate. It could be less if the board is up the first time. I also tried several times for some videos due to shaking or framing issues. Any way, it's fun to participate this project....Show more →
Thanks again Tony, you're my first submission! I just completed the calculations from the A7M3 images you provided and added it to the results page:
p.2 #7 · Call for help - building sensor readout speed database
snapsy wrote:
Sony A1 added. Special thanks to @dclark@ for supplying a huge data set of A1 photos+videos and detailing everything in exquisite detail.
p.2 #8 · Call for help - building sensor readout speed database
j4nu wrote:
Ahh, I just got my Arduino delivered today ...
1/256s! That's faster than I thought...
Interesting note: the A1 and Z9/Z8 have identical readout speeds when adjusted for their slight 4% difference in sensor row count. They both can read a line of pixels off the sensor in just 0.677 microseconds - a microsecond is 1/1,000,000 of a second! It's actually a bit slower than that since they read 12 rows at a time.
p.2 #9 · Call for help - building sensor readout speed database
Naive, and sincerely not sarcastic, question: assume you have all of these numbers, what would be the use of the data? I have only encountered one situation where having even a global shutter would have materially impacted my image quality, and that was photographing fighter jets that were approaching the speed of sound.
p.2 #10 · Call for help - building sensor readout speed database
Grenache wrote:
Naive, and sincerely not sarcastic, question: assume you have all of these numbers, what would be the use of the data? I have only encountered one situation where having even a global shutter would have materially impacted my image quality, and that was photographing fighter jets that were approaching the speed of sound.
Cheers,
Jim
Determining rolling shutter distortion likelihood for fast-moving subjects/cameras (both stills and video) and whether banding will occur for a given lighting cycling frequency. It also helps in reverse-engineer how many samples the AF system will see and also what refresh rates LV can be driven at (if you use the 10-bit video mode rates since that's likely how the sensor is sampled for stills LV). It also helps reverse-engineer the sampling modes used for video, including which are oversampled and which employ line-skipping/binning. which has implications for video IQ.
p.2 #12 · Call for help - building sensor readout speed database
jcolwell wrote:
Great project. I'll contribute, once I get my first RF or Sony body. I'm holding out for the high res RF unicorn, I mean camera. I already have the LED. ;
Excepting the LED..and well Sony....we are brothers!
p.2 #13 · Call for help - building sensor readout speed database
snapsy wrote:
Thanks again Tony, you're my first submission! I just completed the calculations from the A7M3 images you provided and added it to the results page:
I notice that the A7iii results don't include 12-bit RAW - ie for this model Continuous compressed, or APSC 14/12bit. IMHO one of the most useful practical aspects of this project is to provide users with the speeds of different modes so we can decide which mode to use for a particular subject with regard to the tradeoff of roller shutter vs quality. Knowing the fastest speed available is super useful. I didn't know at the time that the R models could half the scan time by selecting continuous mode and so sold my A7Riii and then A7Riv and moved to A7C - had I known this then I might well have stuck with an R!
p.2 #14 · Call for help - building sensor readout speed database
newdom wrote:
I notice that the A7iii results don't include 12-bit RAW - ie for this model Continuous compressed, or APSC 14/12bit. IMHO one of the most useful practical aspects of this project is to provide users with the speeds of different modes so we can decide which mode to use for a particular subject with regard to the tradeoff of roller shutter vs quality. Knowing the fastest speed available is super useful. I didn't know at the time that the R models could half the scan time by selecting continuous mode and so sold my A7Riii and then A7Riv and moved to A7C - had I known this then I might well have stuck with an R!...Show more →
I agree and have already asked the contributor if he could spot re-test continuous compressed so that I can add it to the results. In fact I plan to split the existing "Photo" column on the consolidated results to "Photo (Best IQ)" and "Photo (Fastest)"
p.2 #16 · Call for help - building sensor readout speed database
snapsy wrote:
Determining rolling shutter distortion likelihood for fast-moving subjects/cameras (both stills and video) and whether banding will occur for a given lighting cycling frequency. It also helps in reverse-engineer how many samples the AF system will see and also what refresh rates LV can be driven at (if you use the 10-bit video mode rates since that's likely how the sensor is sampled for stills LV). It also helps reverse-engineer the sampling modes used for video, including which are oversampled and which employ line-skipping/binning. which has implications for video IQ.
Great project and very useful and interesting to me. My concern with sensor readout speed has to do with avoiding banding from LED or other flickering lights while shooting with any form of electronic shutter, which has become a major problem as these lights have become more ubiquitous.
I personally have never had a problem with flickering/banding with either the A1 or the A9, the two cameras on your list that I own. I have never shot billboard lighting or theatrical lighting, which may be more difficult cases.
Since the A9 reads out at 1/152 of a sec and the A1 at 1/256 sec, that suggests that readouts well short of a global shutter (with its compromises in dynamic range and perhaps colors) are sufficient to avoid banding with electronic shutter under flickering lights.
This makes me wonder if the A1II, whenever it finally gets here, will simply speed up the sensor readout rather than try for a global shutter. This would keep the cost down, the image quality and versatility up, and make the camera more attractive for many of us.
p.2 #17 · Call for help - building sensor readout speed database
chiron wrote:
Great project and very useful and interesting to me. My concern with sensor readout speed has to do with avoiding banding from LED or other flickering lights while shooting with any form of electronic shutter, which has become a major problem as these lights have become more ubiquitous.
I personally have never had a problem with flickering/banding with either the A1 or the A9, the two cameras on your list that I own. I have never shot billboard lighting or theatrical lighting, which may be more difficult cases.
Since the A9 reads out at 1/152 of a sec and the A1 at 1/256 sec, that suggests that readouts well short of a global shutter (with its compromises in dynamic range and perhaps colors) are sufficient to avoid banding with electronic shutter under flickering lights.
This makes me wonder if the A1II, whenever it finally gets here, will simply speed up the sensor readout rather than try for a global shutter. This would keep the cost down, the image quality and versatility up, and make the camera more attractive for many of us....Show more →
Thanks. Unfortunately there are some LED lights that cycle very quickly, particularly if they're dimmed below 100% due to PWM. For example here's a photo from the Z9 on a Kasa Smartbulb I own that cycles at 1920Hz:
p.2 #18 · Call for help - building sensor readout speed database
snapsy wrote:
Thanks. Unfortunately there are some LED lights that cycle very quickly, particularly if they're dimmed below 100% due to PWM. For example here's a photo from the Z9 on a Kasa Smartbulb I own that cycles at 1920Hz:
Alas. How fast would the sensor readout on the Z9 or an A1 have to be to avoid the banding and exposure variance on the Kasa KL125 Smartbulb LED at 50% dimming? Is it a speed short of a global shutter?
p.2 #19 · Call for help - building sensor readout speed database
Grenache wrote:
Naive, and sincerely not sarcastic, question: assume you have all of these numbers, what would be the use of the data? I have only encountered one situation where having even a global shutter would have materially impacted my image quality, and that was photographing fighter jets that were approaching the speed of sound.
Cheers,
Jim
snapsy wrote:
Determining rolling shutter distortion likelihood for fast-moving subjects/cameras (both stills and video) and whether banding will occur for a given lighting cycling frequency. It also helps in reverse-engineer how many samples the AF system will see and also what refresh rates LV can be driven at (if you use the 10-bit video mode rates since that's likely how the sensor is sampled for stills LV). It also helps reverse-engineer the sampling modes used for video, including which are oversampled and which employ line-skipping/binning. which has implications for video IQ.
chiron wrote:
Great project and very useful and interesting to me. My concern with sensor readout speed has to do with avoiding banding from LED or other flickering lights while shooting with any form of electronic shutter, which has become a major problem as these lights have become more ubiquitous.
I personally have never had a problem with flickering/banding with either the A1 or the A9, the two cameras on your list that I own. I have never shot billboard lighting or theatrical lighting, which may be more difficult cases.
Since the A9 reads out at 1/152 of a sec and the A1 at 1/256 sec, that suggests that readouts well short of a global shutter (with its compromises in dynamic range and perhaps colors) are sufficient to avoid banding with electronic shutter under flickering lights.
This makes me wonder if the A1II, whenever it finally gets here, will simply speed up the sensor readout rather than try for a global shutter. This would keep the cost down, the image quality and versatility up, and make the camera more attractive for many of us. ...Show more → snapsy wrote:
Thanks. Unfortunately there are some LED lights that cycle very quickly, particularly if they're dimmed below 100% due to PWM. For example here's a photo from the Z9 on a Kasa Smartbulb I own that cycles at 1920Hz:
While I don't shoot with stacked sensor cameras, rather the Canon R6 and R6II, I've found Canon's high speed flicker compensation in the R6II quite useful for some stage lighting situations. In my case, corporate events. The spots used at these are LED and often dimmed quite a lot because they're used in smaller venues and the corporate execs doing the on-stage presentations don't like bright stage lights in their eyes... At the most recent event I did, the flicker compensation feature indicated these spots operated at 1200Hz. The e-shutter banding in these situations with the R6II was fairly faint, but still noticeable if you looked for it because there were numerous bands. My guess is the a1/Z9 would still see some banding but it would be quite broad and probably disguised by the scene unless there are a lot of blank, even-toned surfaces. Based on the calculations in the above example of the LED at 1920Hz, the scenario I was in, if shooting with an a1 or Z9 would likely have resulted ~9 bands across the frame.
FWIW the high speed flicker compensation feature works very well, but is an additional step that has to be done (camera analyses the light and sets/recommends the correct shutter speed).
p.2 #20 · Call for help - building sensor readout speed database
rscheffler wrote:
While I don't shoot with stacked sensor cameras, rather the Canon R6 and R6II, I've found Canon's high speed flicker compensation in the R6II quite useful for some stage lighting situations. In my case, corporate events. The spots used at these are LED and often dimmed quite a lot because they're used in smaller venues and the corporate execs doing the on-stage presentations don't like bright stage lights in their eyes... At the most recent event I did, the flicker compensation feature indicated these spots operated at 1200Hz. The e-shutter banding in these situations with the R6II was fairly faint, but still noticeable if you looked for it because there were numerous bands. My guess is the a1/Z9 would still see some banding but it would be quite broad and probably disguised by the scene unless there are a lot of blank, even-toned surfaces. Based on the calculations in the above example of the LED at 1920Hz, the scenario I was in, if shooting with an a1 or Z9 would likely have resulted ~9 bands across the frame.
FWIW the high speed flicker compensation feature works very well, but is an additional step that has to be done (camera analyses the light and sets/recommends the correct shutter speed). ...Show more →
At what sensor readout speed would dimmed LED lights like the Kasa KL 125 cited above no longer cause banding or exposure variations?
you know what I mean
...