So I was taking some portraits over the weekend using electronic shutter on the a7Riv. Fast lenses with fast shutter speeds. I see some distortion in some of the files (stretched or compressed on the vertical axis) . I am thinking this was caused by camera movement as the sensor was being read. Basically rolling shutter by my not holding steady enough. At fast shutter I am less careful about holding the camera still. I know poor technique that must have some effect even on very fast shutter speeds (something to work on for better results).
No real disasters, but a mix of good and poor shots.
I have seen the numbers posted here before, but can't find them in search. Does anyone know the sensor readout speed of the a7Riv, a7iv, and A1? Also how does this compare the the speed of the mechanical shutter? I am thinking a camera with a flash sync speed of 1/250 must have a shutter that crosses the sensor in at least 1/500. So the A1 must be at least 1/800. This will help me understand how good my technique must be for electronic shutter, maybe there are very few use cases for electronic shutter with the Riv for me. Also one more consideration in my next camera either the Rv or the A1.
I don't understand how you're getting rolling shutter effects taking portraits?
Are the people moving around a lot, are two cars driving along side one another?
I use electronic shutter a lot, on all my cameras, of course, I'm shooting wildlife, but I don't believe I have any
issues on mostly static subjects.
The A1 ES is 1/237 sec, the flash sync speed is 1/200 sec
The A1 mechanical shutter has only a mechanical second curtain. The first curtain is electronic, EFCS. The flash sync speed is 1/400 sec.
The A1 ES is 1/237 sec, the flash sync speed is 1/200 sec
The A1 mechanical shutter has only a mechanical second curtain. The first curtain is electronic, EFCS. The flash sync speed is 1/400 sec.
The A1 ES is 1/237 sec, the flash sync speed is 1/200 sec
The A1 mechanical shutter has only a mechanical second curtain. The first curtain is electronic, EFCS. The flash sync speed is 1/400 sec.
The A9 ES is 1/140 sec.
The A7R4 ES is 1/9.25 sec.
See the above links for more details.
Dave
Thanks Dave, that is what I was looking for. That sure explains my problem. I was shooting everything over 1/1000 and holding with lack of care that would work at those speeds. I'm not sure how the R4 can shoot at 10fps with that readout speed, but your method for measuring looks reasonable. So other than static shots on a tripod, seems like electronic shutter on the R4 is a poor choice.
So the A1 shutter is moving about the same speed as a typical shutter, but only has the second curtain. So if shooting high speed shutter switch to electronic on the A1 to avoid bokeh issues.
davev wrote:
I don't understand how you're getting rolling shutter effects taking portraits?
Are the people moving around a lot, are two cars driving along side one another?
I use electronic shutter a lot, on all my cameras, of course, I'm shooting wildlife, but I don't believe I have any
issues on mostly static subjects.
The subject wasn't moving. I wanted the camera at a height that was too low for me standing, but too high for kneeling. So I was holding in my hand and looking at the back screen for framing instead of against my eye. So it was less stable, and I wasn't working super hard to make it stable as shutter speeds were all high.
Using proper technique I usually don't have obvious issues. But given the speeds quoted I think I will stick with mechanical unless on a tripod.
Sounds more like motion blur due to your unsteady hands and not so much rolling shutter. If you switch to mechanical and still have it, you'll know it's poor technique and motion blur. I don't doubt DClark's data, but it's quite easy to get nice shots with the a7r4 and moving subjects with the electronic shutter. I previously used it for sports all the time as a 2nd body. Here is just one example using the a7r4 and the 400 2.8. If the data is accurate as far as making shots with moderately moving subjects, there is no way I'm getting sharp images with a 400. But I did routinely and the only time I used mechanical shutter was when using strobes. For fast moving objects like a hummingbird wings, the sensor readout speeds made electronic shutter useless. But for most everything else, I never had an issue.
I'm sure I've got lots more images and probably better action ones, but this is what I could find quickly.
I was thinking it was much more likely to be motion blur also, and very little to do with the electronic shutter.
This shot was taken with an a7R4, at 1/1600, it's a panning shot from a boat.
The bird is solid because I did an ok job panning, but the trees are showing the rolling shutter effect.
This shot, with an a7III, at 1/1250, shows what a fast moving subject looks like with electronic shutter.
Note that the flower and the head of the hummer are still not blurred.
If the image is stretched/distorted, it's definitely rolling shutter. I had that happen occasionally on A7III even in rather static shots when my kids moved a bit at the wrong moment.
Seen the ES "issue" lots of times when using my R4 and it doesn't even have to be fast movement to end up distorted.
Worst case I've seen so far came when I was first testing out the camera, trying to capture a young granddaughter on both a swingset and zooming around on a little pushbike. Distorted to the point that it just looked off.
Also had it happen trying to capture a rather sedate wedding walk from the side, while most people in it looked ok others had that distorted look as they had moved at a different speed during the readout.
I still use it when I want to be quiet even when there's movement happening as all that's needed for a good shot is the subjects freezing for a moment, out of a burst all I really need is the one shot.
Not perfect framing but no distortion that I can see, and at 1/80sec also no motion blur as I caught a static moment in their wedding dance. DSC04362 by Johan Olsson, on Flickr
Rolling Shutter: I don't think it matters where the motion occurs. The hands holding the camera (or the speed of the vehicle the camera is in), the subject's movement, or a combination of both.
It's all a result of the same issue, a slow sensor 'read', whether it be Electronic shutter or a slow focal plane shutter. Even a fast focal plane shutter can result in rolling shutter effect, it's just that for the vast majority of subjects, they are fast enough.
Slow shutter curtain speeds (such as seen with many ES and a few mechanical shutters) lead to motion distortion of the image.
Slow shutter speeds (mechanical or ES) lead to motion blur of the image.
Slow shutter curtains mean the top of the image and the bottom of the image are captured at different times. The slower the curtain speed, the bigger the time difference. If either the subject moves or the camera moves during that time, the captured image is distorted. It may show no blur (if the shutter speed is short enough) but it will be distorted.
If the shutter speed is set too slow, and the subject or the camera is moving, the image will be blurred.
The way a shutter works (Mechanical or Electronic or hybrid) is that the first curtain starts the image capture and the second curtain ends the image capture. The time between the first curtain and the second curtain is the shutter speed. The time for the curtain to move from the top of the image to the bottom of the imagte is the curtain speed. If the shutter speed is shorter than the curtain speed, the first curtain and the second curtain form a slit as they move down the image. The slower the curtains the narrower the slit.
It is possible to have a much faster shutter speed than the curtain speed, even for very slow ES curtain speeds like the A7R4. The second curtain follows just a few rows behind the first curtain. This can produce images with little or no motion blur but considerable distortion.
Distortion is sometimes much less noticeable than blur, so often slightly distorted images are acceptable but a slightly blurred image is rejected.
Sometimes both are present. The hummingbird image posted by @davev is a good example. The fast moving wings show motion blur at 1/1250 sec. The slow moving hummer body does not show noticeable blur at 1/1250 sec. The ES shutter curtain speed distortion is dramatic! The closest wing has a blurred image at the top of its stroke and a second blurred image at the bottom of its stroke. In this case as the slit of the ES moved from the top of the image to the bottom it first captured the wing at the top, and then as the hummer's wing moved very quickly to the bottom it overtook the shutter curtains and was then captured again at the bottom of its stroke. The top and bottom are captured more distinctly because they are relatively motionless then, whereas in the stroke from top to bottom they are moving very fast relative to the shutter curtain slit.
Also note that the curtain slit moves from the top to the bottom of the image, but from the bottom to the top of the sensor.
My Issues were defiantly due to motion of the camera using ES. I was shooting outside in the day at f/1.4 so depending on shade or not SS was up to 1/8000, but typically over 1/2000. When I shoot at the reciprocal rule or even 2x the reciprocal rule I consciously work on holding still during the shot. At faster speeds not so much (poor technique).
So as Dave mentioned the shots are sharp because of the high shutter speed, but distorted due to the rolling shutter and my camera movement. I was shooting at med frame rate and have several shots that are similar. As I was looking through them I noticed sometimes the people looked wider than normal and sometimes thinner (camera in vertical). So these would be consistent camera movement during the shot. For some I had a vertical strip that was expanded (clearly visible in the leaves on the ground). I'm guessing maybe some small movement from my shutter press.
Like others I'm certainly capable of getting good photos with ES. Many of the photos I would not notice the distortion except as I flipped thru several in quick succession the people were changing in size. Maybe a good option if you have a larger bride that is not comfortable with her size (just kidding). But, knowing the sensor readout speed helps me make informed decisions on how I'm using the camera. If I get ambitious I will try to replicate Dave's measurements.
So I decided to measure the readout speed myself. Learned quite a bit about the camera. In compressed raw the ES speed is 1 / 18.21 sec, in uncompressed raw it is 1 / 9.95 sec. Or roughly 1/20 compressed and 1/10 uncompressed. These are both slower than I would want for the FL I typically shoot handheld. My LEDs were operating at 120Hz, that seems right for a cheap rectifying circuit.
To get the shot I took off the lens and held camera near a diffused LED on low brightness (low duty cycle).
compressed: 120 cycles/sec * 961.5 px/cycle / 6336 px/frame = 18.21 frames / sec
uncompressed: 120 cycles/sec * 525.4 px/cycle / 6336 px/frame = 9.95 frames / sec
I also measured the Hi+ frame rate. For mechanical it was 10 FPS, with ES it was 7.5 FPS.
In the pictures taken with mechanical shutter, the shutter does not expose the entire frame evenly. You can see in the upper corners and to a slight degree in the lower right corner.
Also included 50% crops from the photoshoot that made me ask this question.
I guess realistically the only path forward is to pick up an A1!
good shot
ILCE-7RM4Sigma 85mm F1.4 DG DN | A (Sony E) lens85mmf/5.61/400s200 ISO0.0 EV
distorted / blurry shot from camera motion with ES
ILCE-7RM4Sigma 85mm F1.4 DG DN | A (Sony E) lens85mmf/5.61/400s200 ISO0.0 EV
compressed ES at 120Hz LED
ILCE-7RM4---- lens1/8000s320 ISO0.0 EV
uncompressed ES at 120Hz LED
ILCE-7RM4---- lens1/8000s320 ISO0.0 EV
Mechanical shutter, note shutter induced vignette in upper corners
dclark wrote:
@tschopp@ what is the orientation of the camera in the first two shots?
It was vertical. I usually hold with the "top" towards the left, but can't guarantee that. The image is heavily cropped to show the detail. So the vertical strip of distortion is consistent with the shutter curtain.
tschopp wrote:
It was vertical. I usually hold with the "top" towards the left, but can't guarantee that. The image is heavily cropped to show the detail. So the vertical strip of distortion is consistent with the shutter curtain.
Yes. The reason for my question was to make it clear that with the camera in its "normal" landscape orientation the distortion due to slow shutter curtains is in horizontal bands since the shutter is moving in the vertical direction.
tschopp wrote:
So I decided to measure the readout speed myself. Learned quite a bit about the camera. In compressed raw the ES speed is 1 / 18.21 sec, in uncompressed raw it is 1 / 9.95 sec. Or roughly 1/20 compressed and 1/10 uncompressed. These are both slower than I would want for the FL I typically shoot handheld. My LEDs were operating at 120Hz, that seems right for a cheap rectifying circuit.
To get the shot I took off the lens and held camera near a diffused LED on low brightness (low duty cycle).
compressed: 120 cycles/sec * 961.5 px/cycle / 6336 px/frame = 18.21 frames / sec
uncompressed: 120 cycles/sec * 525.4 px/cycle / 6336 px/frame = 9.95 frames / sec
I also measured the Hi+ frame rate. For mechanical it was 10 FPS, with ES it was 7.5 FPS.
In the pictures taken with mechanical shutter, the shutter does not expose the entire frame evenly. You can see in the upper corners and to a slight degree in the lower right corner.
Also included 50% crops from the photoshoot that made me ask this question.
I guess realistically the only path forward is to pick up an A1! ...Show more →
That ES blur shot is very interesting, I've never seen something like this on A7III. The effect was mostly visible as elongated features...
Also, unless you're using EFCS, I think mechanical should always result in even exposure/vignetting?
