Can You Figure Out Z9 Reading Pattern from This Sample?

If you would like to post, you'll need to register. Note that if you have a BCG store account, you'll need a new, separate account here (we keep the two sites separate for security purposes).

Tibor Hrnko

Tibor Hrnko

Well-known member
Supporting Member
Marketplace
Member's gallery
It is clear, that it is the mirrorless equivalent of rolling shutter effect, I just cannot figure out the rule. Do you?
20 FPS 1/125 s, unrotated RAW file
Chinook front rotor is moving counterclockwise and aft rotor clockwise as seen from above
Rotors detail.jpg
You can only see EXIF info for this image if you are logged in.

On net I have found that Z9 reads 12 horizontal lines in parallel, but the chunk of front propeller seems to me thicker than that. I have counted at least 2x more from vanishing edge of the upper tip to the starting edge of lower tip. Here is a detail:
Rotor pixels.jpg
You can only see EXIF info for this image if you are logged in.

The position of the front blade tip on the upper rows of pixels seems to be later in time compared to lower rows. That is consistent with inverted image on the sensor and camera scanning from the top of the sensor to the bottom, but I am vexed with the shape of the front blade of the aft propeller.

I cannot guess either how it compares to the mechanical shutter of DSLR. It seems slower to me as I have not noticed it on any picture by D850, but that is no proof :).

Can you please comment?
 
Last edited:
Member's gallery
There does appear to be a slight rolling shutter affect in what might be described as an extreme example.
Does anyone know how fast were the blades were rotating at the point where the effect can be seen?

Back to a Nikon DSLR, at faster shutter speeds on pro bodies the image seems to have been taken by a narrow slit travelling vertically at about 1/350 of a second - with some possibility of an effect similar to a rolling shutter with a subject moving very fast from left to right.
 
Member's gallery
Len Shepherd said:
There does appear to be a slight rolling shutter affect in what might be described as an extreme example.
Does anyone know how fast were the blades were rotating at the point where the effect can be seen?

Back to a Nikon DSLR, at faster shutter speeds on pro bodies the image seems to have been taken by a narrow slit travelling vertically at about 1/350 of a second - with some possibility of an effect similar to a rolling shutter with a subject moving very fast from left to right.
Click to expand...
It seems to be 225 rpm as stated here
 
Member's gallery
boviscopophobic said:
I think the rear rotor blade that is passing in front is effectively acting as a second shutter. Notice how neither of the "broken" tips are as motion blurred as the rest of their blade. So in effect the rear rotor is adding its own rolling shutter to the mix.
Click to expand...
Can you be more specific, please? I was thinking along those lines, but i could not guess the mechanics of it.
 
Member's gallery
Tibor Hrnko said:
Can you be more specific, please? I was thinking along those lines, but i could not guess the mechanics of it.
Click to expand...

The rotor blade is pointing a bit to the right at the beginning of the exposure and a bit to the left at the end of it. Since the blade blocks anything behind it from being seen, it means that objects behind it and on the left are seen in the earlier half of the exposure, while objects behind it and on the right are seen in the later half.

If we divide the whole exposure time into thirds in order from 1 to 3, then going from left to right along the area swept out by the rotor blade, the time of imaging goes roughly like

2-1-3-2

(2 appears on both ends because when the blade is pointing parallel to the vertical axis of the image, the scene behind it is visible on either side. You can see that the corresponding parts of the "broken tips" line up.)

Similarly the amount of motion blur, in order from left to right, goes something like

more-less-less-more

due to the varying amounts of time that the parts of the scene are blocked by the passage of the blade.
 
Member's gallery
If you refer to the region where two rotorblades overlap, I'd say it may be rather an optical effect, such as diffraction or whatever, than rolling shutter.
Rolling shutter effect should be seen across the entire blade.
 
Member's gallery
boviscopophobic said:
The rotor blade is pointing a bit to the right at the beginning of the exposure and a bit to the left at the end of it. Since the blade blocks anything behind it from being seen, it means that objects behind it and on the left are seen in the earlier half of the exposure, while objects behind it and on the right are seen in the later half.

If we divide the whole exposure time into thirds in order from 1 to 3, then going from left to right along the area swept out by the rotor blade, the time of imaging goes roughly like

2-1-3-2

(2 appears on both ends because when the blade is pointing parallel to the vertical axis of the image, the scene behind it is visible on either side. You can see that the corresponding parts of the "broken tips" line up.)

Similarly the amount of motion blur, in order from left to right, goes something like

more-less-less-more

due to the varying amounts of time that the parts of the scene are blocked by the passage of the blade.
Click to expand...
Thought a second about it and totally agree with this explanation.
 
Member's gallery
boviscopophobic said:
The rotor blade is pointing a bit to the right at the beginning of the exposure and a bit to the left at the end of it. Since the blade blocks anything behind it from being seen, it means that objects behind it and on the left are seen in the earlier half of the exposure, while objects behind it and on the right are seen in the later half...
Click to expand...
Ok I think I have got your idea.
But does your explanation fit to the timing of the camera shutter, rotor spin and Z9 readout speed?
My own analysis was based on these facts
  • Rotor spin 225 RPM = 3.75 RPS -> aft rotor "shutter speed" is 1/3.75s
  • Rotor moved during 1/125 exposure 3.75/125*360 = 10.8 degrees of rotation between the start of exposure and its end
  • Z9 readout speed is 1/270 according this article -> at 1/125 there is no rolling shutter effect due to readout speed as it is 2x the shutter speed. It should be same as with a mechanical shutter.
  • We are discussing a part of the picture which is vertically approx 480px out of 5500, so the reading time should be less than 1/1250s = 1.08 degrees of rotor angle movement
 
Member's gallery
Tibor Hrnko said:
Ok I think I have got your idea.
But does your explanation fit to the timing of the camera shutter, rotor spin and Z9 readout speed?
My own analysis was based on these facts
  • Rotor spin 225 RPM = 3.75 RPS -> aft rotor "shutter speed" is 1/3.75s
  • Rotor moved during 1/125 exposure 3.75/125*360 = 10.8 degrees of rotation between the start of exposure and its end
  • Z9 readout speed is 1/270 according this article -> at 1/125 there is no rolling shutter effect due to readout speed as it is 2x the shutter speed. It should be same as with a mechanical shutter.
  • We are discussing a part of the picture which is vertically approx 480px out of 5500, so the reading time should be less than 1/1250s = 1.08 degrees of rotor angle movement
Click to expand...

As you note, there is little to no rolling shutter contributed by the sensor readout. Here the relevant parameter is the exposure time, namely 1/125 of a second. The fact that part of the scene is not uniformly exposed during this time, due to the action of the rotor, is the only practical contributor to the effect.
 
Member's gallery
aolander said:
You're not alone, and I'm failing to see the value of this discussion.
Click to expand...
The value is in understanding how things work IMO.
It does not have direct impact on taking photos, apart of the realization that no matter how short the exposition is, shutter speed of Z9 is always maxed at 1/270s. (unless there is some computational trick applied that is so far known only to Nikon engineers.}
 
Member's gallery
Tibor Hrnko said:
The value is in understanding how things work IMO.
It does not have direct impact on taking photos, apart of the realization that no matter how short the exposition is, shutter speed of Z9 is always maxed at 1/270s. (unless there is some computational trick applied that is so far known only to Nikon engineers.}
Click to expand...
In a similar way unless SLR's could take a flash image at 1/320 (with some slight loss of exposure - D5 from memory) the vertical travel speed of the shutter slit even at 1/4000 was no more than about 1/300.
 
Member's gallery
Chaz Harrison said:
I am equally dumb, but just a thought - could it be that the "detached" image of the front rotor is a reflection in the rear? in any case I can't see that it matters. can we see the whole image please?
Click to expand...
I do not think that it is a reflection.
Here is the uncropped RAW exported into JPEG with longer size 3840px. I can post the original RAW too but I am not sure about the file size for this forums.
Chinook Full.jpg
You can only see EXIF info for this image if you are logged in.

Note: The picture was never meant for publishing online (clipped rotor).
 
Member's gallery
Thanks for that. I don't think I would be worried over much by the clipped rotor - though it does look like a photographic artefact. As a way of illustrating how the camera's processor is making an image from the output of the sensor it is of some interest but we have to accept that any camera will always produce a "strange" result with fast rotating parts - strange because it will never match what the human eye sees. - be it blurred or frozen by a very fast shutter speed. I'll shut up now.
 
You must log in or register to reply here.
Back
Top