aCuria
Offline
• • •
Upload & Sell: Off
|
 p.10 #18 · A7V dynamic range is extremely good | |
chiron wrote:
Thank you for very much for your efforts on this. I think 4 msec is faster than the A7V? I don't want to and I won't get into a long exchange on this topic. But I actually don't need a simulation to judge banding on the the A1, the A1 II, or the A9 because I have shot all three of those cameras extensively under LED lights while using electronic shutter and I have that experience to judge from.
My shutter speeds indoors are virtually always at 1/250 or slower and, at those speeds and under the LED lights that I have encoutered, I do not get banding that is visible to me or to others looking at the images. On the other hand, when using a non-stacked sensor, I always get banding under LED lights.
As Github writes:
"Note that as the shutter speed is slowed further and approaches the readout speed the bands start to become indistinguishable. This is because the brightness difference of missing one cycle of light at slow shutter speeds has a commensurately smaller effect since one cycle represents a smaller percentage of the total light the bands are receiving."
The readout speed of the A1/A1 II is 1/256 of a second. The readout speed of the A9 is 1/152 of a second. (Both are close to the speed at which under mechanical shutter the shutter slit traverses the sensor.) At those shutter speeds or slower, I don't get visible banding. As the LED light levels go down and their flicker interval becomes longer, my shutter speeds, of course, become slower in a way that continues to prevent visible banding.
In your simulation, you maintain a 1/250 simulated shutter even while the light level falls and the length of the flicker interval increases. This will, of course, eventually produce visible banding. But it does not match what I do (or what I think most other photographers would do) under conditions of falling light--I lower my shutter speed as light falls, often down to 1/8 second or even slower. That compensates for the longer flicker interval.
The Github reference is from https://github.com/horshack-dpreview/RollingShutter...Show more →
chiron wrote:
Thank you for very much for your efforts on this. I think 4 msec is faster than the A7V?
The tool can simulate any readout speed really. This is 16ms and led flicker at 90hz
chiron wrote:
My shutter speeds indoors are virtually always at 1/250 or slower and, at those speeds and under the LED lights that I have encoutered, I do not get banding that is visible to me or to others looking at the images. On the other hand, when using a non-stacked sensor, I always get banding under LED lights.
If it works for your shooting conditions, that’s really all that matters for you. From what I can tell, if the LED flicker frequency is somewhere around 125 to 500 Hz, you should see banding on the A1 quite easily at 1/250s. You could probably prove this pretty easily with an LED and an Arduino, but I’m not great with the hardware side of things.
In real life, you have not encountered anything flickering at those frequencies, and that’s what I mean when I say it comes down to “luck.”
Other people have certainly encountered flicker on the A1 electronic shutter. Everyone's shooting conditions are different, It is what it is https://www.fredmiranda.com/forum/topic/1747977/
chiron wrote:
Note that as the shutter speed is slowed further and approaches the readout speed the bands start to become indistinguishable. This is because the brightness difference of missing one cycle of light at slow shutter speeds has a commensurately smaller effect since one cycle represents a smaller percentage of the total light the bands are receiving.
Idea is correct but this is poorly written as it has nothing to do with readout speed. I rewrote it below:
The slower your shutter speed is and the faster the LED is flickering, the less noticeable the banding becomes. That’s because when many flicker cycles are captured inside the exposure time, missing a single flicker cycle makes a much smaller difference to the total amount of light being captured. One cycle is just a tiny fraction of the overall exposure, so the brightness difference between bands is much less obvious.
chiron wrote:
As the LED light levels go down and their flicker interval becomes longer, my shutter speeds, of course, become slower in a way that continues to prevent visible banding.
Your idea here is that the LED must be dim in order to have a long flicker interval. I don't really agree with this. You can have a very bright LED billboard or non-dimmable LED that's flickering with a relatively long flicker interval.
chiron wrote:
I lower my shutter speed as light falls, often down to 1/8 second or even slower. That compensates for the longer flicker interval.
I shoot my kids at around 1/500s even if the light is poor. I will shoot at f/1.4 and use a speedlight or strobes to compensate, because shooting at say 1/250s results in majority of frames having motion blur.
Frankly "1/8s" means subject isn’t moving. And if the subject isn’t moving, there’s usually no need for high burst rates in the first place. If we don’t need high burst rates, then the obvious question is: why are we using electronic shutter at all?
With a mechanical shutter, banding is far less of an issue than with electronic shutter, even on cameras with relatively slow sensors like the A7R IV. So in those low-movement, low-shutter-speed situations, mechanical shutter is often the more robust and predictable choice.
The mechanical shutter on the A1 is remarkably fast, the 1/400s flash sync speed with mechanical shutter while electronic flash sync is only 1/200s suggests that Sony's mechanical shutter tech is at least twice as fast (2ms, or 1/500s) as A1 electronic.
|
. Snapsy uses a custom arduino LED + program to blink it I think. I bought one as well as he asked about it once