p.1 #1 · Does the R5 and R6 have issues with bokeh rendering using EFCS?
Hello!
So I did a video explaining why it was EFCS and not silent shooting with Sony that had the bokeh quality issues.
&t=6s
From my understanding this was for any camera using EFCS but someone commented that he does not see that on his R6.
I was wondering if you guys can provide some samples at large apertures and high shutter speeds like 1//4000 1.4 for example, with EFSC and without it (like full electronic shutter).
p.1 #4 · Does the R5 and R6 have issues with bokeh rendering using EFCS?
It's not a dramatic issue unless you are using a super-duper wide aperture at a super-duper fast shutter speed, and even then it would be hard to call it "dramatic". I could easily understand how someone would think that they don't see the issue for X, Y, Z, reasons.
In fact, I can barely claim to have noticed it, using f/2.0 or f/1.4 primes and fast shutter speeds.
p.1 #5 · Does the R5 and R6 have issues with bokeh rendering using EFCS?
amacal1 wrote:
It's not a dramatic issue unless you are using a super-duper wide aperture at a super-duper fast shutter speed, and even then it would be hard to call it "dramatic". I could easily understand how someone would think that they don't see the issue for X, Y, Z, reasons.
In fact, I can barely claim to have noticed it, using f/2.0 or f/1.4 primes and fast shutter speeds.
From my tests (which you can see on the samples on my video) the difference is very dramatic to my eyes specially when I buy expensive lenses that have smooth rendering but of course, to each its own.
Thanks
p.1 #6 · Does the R5 and R6 have issues with bokeh rendering using EFCS?
GabrielPhoto wrote:
[
From my tests (which you can see on the samples on my video) the difference is very dramatic to my eyes specially when I buy expensive lenses that have smooth rendering but of course, to each its own.
Thanks
To each his own. But, you are not wrong. The effect is real and it depends on the gap between the first shutter (electronic shutter = sensor plane) and the second shutter (physical shutter) and the shutter speed.
p.1 #7 · Does the R5 and R6 have issues with bokeh rendering using EFCS?
amacal1 wrote:
It's not a dramatic issue unless you are using a super-duper wide aperture at a super-duper fast shutter speed, and even then it would be hard to call it "dramatic". I could easily understand how someone would think that they don't see the issue for X, Y, Z, reasons.
In fact, I can barely claim to have noticed it, using f/2.0 or f/1.4 primes and fast shutter speeds.
I've tried to find it wide open on my 50/1.4 (my only <f/2.8 lens) in broad daylight and could not. Maybe I'm not that picky or maybe it isn't a photo-ruining issue.
Notice how in the EFCS shot, how rough the bokeh is around the wheels of the car in the background, and the faded highlights in the yard above that, where the top half of the bokeh balls is faded out towards the top, while in the full mechanical shot, the background is smooth, and the bokeh balls are even and full size.
Notice how in the EFCS shot, how rough the bokeh is around the wheels of the car in the background, and the faded highlights in the yard above that, where the top half of the bokeh balls is faded out towards the top, while in the full mechanical shot, the background is smooth, and the bokeh balls are even and full size.
Thanks for the sample. As expected it does happen.
Appreciate the help!
p.1 #10 · Does the R5 and R6 have issues with bokeh rendering using EFCS?
And still neither Canon nor Sony have sorted it out via firmware update.. (i.e. an option to automatic switch to mechanical when a certain predefined aperture/shutter speed combination is reached).
It's pretty annoying because I usually use EFCS and forget almost all the times to turn it off when I should..
p.1 #11 · Does the R5 and R6 have issues with bokeh rendering using EFCS?
Tirpitz666 wrote:
And still neither Canon nor Sony have sorted it out via firmware update.. (i.e. an option to automatic switch to mechanical when a certain predefined aperture/shutter speed combination is reached).
It's pretty annoying because I usually use EFCS and forget almost all the times to turn it off when I should..
I lost count of how many times I nagged everyone I could talked to from Sony about it. Seems like such a simple thing to fix :/
Notice how in the EFCS shot, how rough the bokeh is around the wheels of the car in the background, and the faded highlights in the yard above that, where the top half of the bokeh balls is faded out towards the top, while in the full mechanical shot, the background is smooth, and the bokeh balls are even and full size.
I am still trying to understand this behavior with EFCS specifically the harsh bokeh. I fully Understand the shading and bokeh balls getting cut-off. What I don't understand is on mechanical it's soft bokeh while on EFCS it is harsh like nervous?
What they didn't explain is what causes the OOF harshness on EFCS? Since the mechanical is doing the same thing on 2nd curtain it's still cutting off light rays only difference is the 1st curtain is on the same height as the 2nd.
p.1 #13 · Does the R5 and R6 have issues with bokeh rendering using EFCS?
This is not a camera specific issue. It's physics. In optics up is down and down is up. The first and second curtains of a mechanical shutter both move in the same direction so all shadows cast by the shutter curtains onto the sensor move in the same direction. So each pixel gets the same total exposure.
When a point of light is in sharp focus in front of the sensor it is out of focus at the sensor (i.e. a bokeh ball). The real world bottom of that highlight disk is projected toward the top of the sensor and the top toward the bottom. A mechanical second curtain moving from top to bottom (like Canon) casts a shadow on the bottom of the highlight disk first because that shadow is inverted just like the light beam it occludes. The shadow of the mechanical curtain moves across the inverted image of that tiny disk on the sensor from bottom to top in the opposite direction to the shutter.
The trouble is that EFCS (the wave of progressive pixel resets) moves in the same physical top to bottom direction as the mechanical shutter. EFCS is not "casting" a shadow from above the focal sensor plane like the mechanical shutter. EFCS is "casting" a shadow at the focal sensor plane. So the EFCS "shadow" is not inverted. The pixels near the top of that tiny inverted out of focus highlight (your bokeh ball) on the sensor get exposed slightly longer than the pixels near the bottom because the shadow of the second curtain on that tiny highlight is actually moving in the opposite direction to the EFCS.
I can't imagine any firmware fix. Bringing the mechanical shutter insanely close to the sensor would reduce but never eliminate this issue. Ultimately, global shutter is the only "fix". It acts more similarly to mechanical shutter in that all areas of the image are exposed equally.
p.1 #14 · Does the R5 and R6 have issues with bokeh rendering using EFCS?
So, what is the cut off of what is acceptable? Is there a general rule of thumb that says if you stop down to say f/1.6 or f/2, or go below 1/1000th, the anomaly is gone?
p.1 #15 · Does the R5 and R6 have issues with bokeh rendering using EFCS?
jeffbuzz wrote:
The trouble is that EFCS (the wave of progressive pixel resets) moves in the same physical top to bottom direction as the mechanical shutter. EFCS is not "casting" a shadow from above the focal sensor plane like the mechanical shutter. EFCS is "casting" a shadow at the focal sensor plane. So the EFCS "shadow" is not inverted. The pixels near the top of that tiny inverted out of focus highlight (your bokeh ball) on the sensor get exposed slightly longer than the pixels near the bottom because the shadow of the second curtain on that tiny highlight is actually moving in the opposite direction to the EFCS. ...Show more →
If I am understanding this correctly the EFCS shadow on the sensor, this is the one that's causing Harshness on the bokeh right? then you have another shadow from 2nd curtain that is contributing another shadow in opposite direction?
p.1 #16 · Does the R5 and R6 have issues with bokeh rendering using EFCS?
bushwacker wrote:
If I am understanding this correctly the EFCS shadow on the sensor, this is the one that's causing Harshness on the bokeh right? then you have another shadow from 2nd curtain that is contributing another shadow in opposite direction?
EFCS does not cast an actual shadow because it isn't a physical object (hence my over zealous use of "air" quotes). The electronic shutter is just a row by row progression of data readouts that moves across the sensor in the same direction as the physical mechanical shutter. It works like a flatbed scanner as it reads one horizontal line of pixels at a time from the top to the bottom of the sensor. The intent is that each pixel gets the same total light exposure because each is illuminated in the same gap between the opening readout wave of the EFCS and the closing shadow of the mechanical curtain. Bigger gap between curtains = longer exposure.
The problem isn't the e-shutter or the mechanical shutter per se. The problem is using them together in the same exposure. Only the mechanical shutter casts a shadow onto the sensor. So only its shadow gets inverted by the shallow light beam angle focused in front of the sensor due to a wide aperture. This is not a problem when both curtains are mechanical because both cast shadows and both get inverted. So all your bokeh balls are exposed evenly.
The e-shutter does not cast a shadow to be inverted. It simply reads directly off the illumination projected onto the sensor from top to bottom. The mechanical shutter curtain's shadow being inverted means it moves in the opposite direction to both the physical curtain and the EFCS readout in the area of your bokeh ball. So you end up with a bunch of illuminated disks (the bokeh balls) all over your sensor with slightly higher exposure on the side opposite the direction of the shutter travel.
I "borrowed" a drawing from wikipedia and added some further notes to illustrate:
By !Original: GoliardicoVector: cmglee - Own work based on: Cirles of confusion lens diagram.png, CC BY-SA 4.0, https://commons.wikimedia.org/w/index.php?curid=81265400
my modified version of drawing with shutter curtains
p.1 #17 · Does the R5 and R6 have issues with bokeh rendering using EFCS?
bushwacker wrote:
I am still trying to understand this behavior with EFCS specifically the harsh bokeh. I fully Understand the shading and bokeh balls getting cut-off. What I don't understand is on mechanical it's soft bokeh while on EFCS it is harsh like nervous?
What they didn't explain is what causes the OOF harshness on EFCS? Since the mechanical is doing the same thing on 2nd curtain it's still cutting off light rays only difference is the 1st curtain is on the same height as the 2nd.
circular blur is interrupted abruptly? if you were looking at a perfect circle, and it was chopped in half, it would have a sharp edge.
p.1 #18 · Does the R5 and R6 have issues with bokeh rendering using EFCS?
Global shutter for the win? Or would there be a different issue to contend with? No mechanical shutter needed and rolling shutter issues solved at the same time, I believe.
