moire on bird feather

[BobGilbody]

Hello,
Got this image:

https://www.flickr.com/photos/167726...posted-public/

Sorry can't show photo here, don't know how.
Anyway forget the photo it's the moiré on the feathers I'm interested in. I'm using a Canon 5Ds with a Sigma 70-200 f2.8 lens, and speedlites.
I got this off Google:

"Moiré occurs in a photograph when a scene, an object or a fabric being photographed contains repetitive details (dots, lines, checks, stripes) that exceed the sensor resolution. The camera produces a strange-looking wavy pattern that is very distracting and not what you want from a corporate headshot."

I don't think this is relevant to my problem. See image:

https://www.flickr.com/photos/167726...posted-public/

This image is taken with the same Canon 5Ds albeit with a different lens. I can see no moiré effect in this image.

What do I need to do differently to get images of BiF that do not exhibit moiré?

The first image is cropped, maybe that is relevant.

Any help appreciated, Bob


20 04 22 rainbow-3616-DeNoiseAI-standard-SharpenAI-Focus by Bob Gilbody, on Flickr

[odds]

...Moiré occurs in a photograph when a scene, an object or a fabric being photographed contains repetitive details (dots, lines, checks, stripes) that exceed the sensor resolution.

The old LuLa site had demo image(s) to show that also the monitor can cause moiré to appear in a moiré-free image. Some caution may be required when judging on screen.
--
Odd S.

[Manfred M]

Moire was more of an issue with lower resolution sensors. Once you get into sensors like yours with 50MP, it is generally not an issue.

I have seen the monitor induced moire that Odds describes and I suspect that downsampling to 2MP images plus software interpolation issues show it on the screen.

Also - take a look at the edit I made to your post. I inserted the BBCode link that the Flickr site generates. That is the best way to display images here.

[BobGilbody]

Thanks Manfred. I found the BBCode thing but no idea what to do with it.

Bob

[billtils]

Bob, copy the BBCode and paste it into your post.

The unedited paste will include data in addition to that which is required for posting the image here - you only need the parts bounded by the leading and trailing "[img]"

[BobGilbody]

Blue-faced Honeyeater 27 12 20-5633-DeNoiseAI-denoise by Bob Gilbody, on Flickr

Thanks Bill it works.

[BobGilbody]

Thanks odds. I think it is more than a monitor thing.

All the best Bob

[xpatUSA]

The old LuLa site had demo image(s) to show that also the monitor can cause moiré to appear in a moiré-free image. Some caution may be required when judging on screen.

A very good point, Odd. The monitor should always be set to 100 percent zoom when checking for moiré. Anything less can cause aliasing artifacts especially if the image is not smoothed by the viewing software. Moiré becomes particularly likely when viewing a full-size Canon image in "fit-to-screen" mode.

In fact, checking image quality at anything less than 100 per cent is wasting our time. Equally, zooming in must be at intervals of 2x, 3x, 4x, 5x, etc and with smoothing disabled ... unless we like puzzling over the results of re-sizing interpolation!

This might be of interest to a few of us:

http://kronometric.org/phot/iq/Down%...%20methods.htm

Bob, g'day, feathers are an interesting subject: Let's say the feather pattern is 0.5mm which is two cycles per scene mm. Let's say the sensor is 0.005mm pixel pitch i.e. 200 samples per image mm. Remember that Harry Nyquist said that aliasing, e.g. moiré, occurs above half that frequency i.e. above 100 samples per mm. From there, it's all simple geometry including focal length and shooting distance. I can not comment on the posted images which are of different birds probably shot at different focal lengths and shooting distances. But, taking my hypothetical bird feather, it's easy enough to figure a theoretical distance that will cause moiré to appear at the sensor on that feather if shot dead on at 200mm focal length. I make it anything above 10 meters, for what that's worth.

Too technical for "most of us", I'm told - but here's how that works anyway:

It takes two pixels to resolve something, so that would be 10um. For a 200mm lens shooting BIF we can say that the sensor is approx 200mm from the lens. "Similar triangles" then leads us to the same relation for a pattern size and shooting distance.

Inside the camera, the triangle is a 10um base and a 200mm height, giving a ratio of 10um over 200mm equals 50 µrad.

Apologies to "most of us" for mentioning angular measure (radians) ...

Outside the camera we have the pattern as 0.5mm (d) which we divide by the above ratio to get the shooting distance (D). Ergo, D = 0.5mm/50µrad = 10 meters or more gives aliasing of that pattern.

HTH.

[BobGilbody]

Thanks Ted. What you have explained seems reasonable enough but does it take into account movement of the bird?
Bit more info:
The rainbow was photographed from about 1.8m 1/4000 sec. f11 70mm ISO 800 three speedlites HSS
The juvenile blue faced honey eater about 3m 1/1000 sec. f6.3 388mm ISO 1600 natural light

If the explanation you give assumes stationary subjects [ I'm guessing it does ] then the possibility is that the moiré problem with the image of the rainbow is related to movement of the subject. That would leave me with two possible solutions [ maybe ]:
1. I raise the shutter speed to 1/8000 sec
2. I set up more speedlites at reduced power and hope the duration of the 'flash' [ I am aware of the mysteries of high speed sync ] is small enough to 'freeze movement' [ no HSS just ttl ]

One further possibility would be to set up lights to illuminate the scene continuously, but still apply 1. and 2. above.

What are your thoughts?

All the best Bob

[xpatUSA]

Thanks Ted. What you have explained seems reasonable enough but does it take into account movement of the bird?

No, Bob.

Bit more info:
The rainbow was photographed from about 1.8m 1/4000 sec. f11 70mm ISO 800 three speedlites HSS
The juvenile blue faced honey eater about 3m 1/1000 sec. f6.3 388mm ISO 1600 natural light

If the explanation you give assumes stationary subjects [ I'm guessing it does ] then the possibility is that the moiré problem with the image of the rainbow is related to movement of the subject. That would leave me with two possible solutions [ maybe ]:
1. I raise the shutter speed to 1/8000 sec
2. I set up more speedlites at reduced power and hope the duration of the 'flash' [ I am aware of the mysteries of high speed sync ] is small enough to 'freeze movement' [ no HSS just ttl ]

One further possibility would be to set up lights to illuminate the scene continuously, but still apply 1. and 2. above.

What are your thoughts?

All the best Bob

As far as I know, subject or camera movement does not cause moiré, unless there is a nearby and similar overlaid pattern, see:

https://en.wikipedia.org/wiki/Moir%C...tern_formation
.

[BobGilbody]

Thanks for the post Ted. I think the moiré on the Rainbow's feathers is caused by two overlaying patterns. My understanding of HSS is that the speedlite emits a regular pulse of light [ somewhere I read the duration of the pulse is about 1/30000 of a second ]. The shutter works by a slit passing over the sensor. The slit allows a slice of the sensor to be exposed for 1/4000 of a second. During that 1/4000 of a second approximately 7 light pulses have hit the sensor. If the bird is moving then in 1/30000 of a second some distance has been travelled. I might have all this wrong but it seems to me that there is the possibility of the sensor receiving 7 different 'images'. These images are on top of each other. This fits your "unless there is a nearby and similar overlaid pattern" maybe. If this is the case then moiré could be generated by the pulses of light and movement of the bird.
This image shows no moiré in the wing feathers and tail feathers but does show moiré on the leg feathers.
[/url]17 04 22 PP2-PP3-P4-P5-P6 by Bob Gilbody, on Flickr
My guess is that the bird is rotating its legs so the speed of the leg is greater than the speed of the wing and tail feathers.
What do you think?

Bob

[xpatUSA]

Thanks for the post Ted. I think the moiré on the Rainbow's feathers is caused by two overlaying patterns. My understanding of HSS is that the speedlite emits a regular pulse of light [ somewhere I read the duration of the pulse is about 1/30000 of a second ]. The shutter works by a slit passing over the sensor. The slit allows a slice of the sensor to be exposed for 1/4000 of a second. During that 1/4000 of a second approximately 7 light pulses have hit the sensor. If the bird is moving then in 1/30000 of a second some distance has been travelled. I might have all this wrong but it seems to me that there is the possibility of the sensor receiving 7 different 'images'. These images are on top of each other. This fits your "unless there is a nearby and similar overlaid pattern" maybe. If this is the case then moiré could be generated by the pulses of light and movement of the bird.
This image shows no moiré in the wing feathers and tail feathers but does show moiré on the leg feathers.
[/url]17 04 22 PP2-PP3-P4-P5-P6 by Bob Gilbody, on Flickr
My guess is that the bird is rotating its legs so the speed of the leg is greater than the speed of the wing and tail feathers.
What do you think?

Bob

I'm pretty much lost in the world of fancy "speedlights", sorry. As to shutter action, I'm slightly more comfortable but have not read of moiré per se being caused by motion of the slit above sync speed combined with motion of the subject. All I know about that is that distortion can be caused as the slices are offset with respect to time e.g. the classic elliptical wheels on racing cars.

Couldn't make out any obvious moiré in the bird's leg, unless it is as arrowed below:



If, on the other hand, a shot taken at less than sync speed is super-imposed on another with a time lapse, moiré could occur and would not necessarily involve the sensor's pixel pitch or shooting distance.

For example, I would expect the following to show moiré if moved by a few px then blended with the original:



Voila:



Note: there is talk of motion here:

https://www.nature.com/articles/s41598-020-70427-x

"When an object moves, its visual appearance is the superposition of multiple elementary phases of movement during the exposure time of a camera; the moiré patterns also change. A dynamic picture depends, particularly, on the speed and direction of the motion. [BUT] With a few exceptions, the moiré effect is typically investigated in the static case."

It seems to imply multiple images captured with motion during the time the shutter is fully open and the use of some form of strobe lighting ... and might apply to a slit subjected to your speedlight pulsing, assuming that the slit is stepped across the sensor rather than moving at constant velocity ...

... for more about that, see here:

https://www.wrotniak.net/photo/tech/fp-shutter.html
.

[xpatUSA]

My understanding of HSS is that the speedlite emits a regular pulse of light [ somewhere I read the duration of the pulse is about 1/30000 of a second ]. The shutter works by a slit passing over the sensor. The slit allows a slice of the sensor to be exposed for 1/4000 of a second. During that 1/4000 of a second approximately 7 light pulses have hit the sensor. If the bird is moving then in 1/30000 of a second some distance has been travelled. I might have all this wrong but it seems to me that there is the possibility of the sensor receiving 7 different 'images'. These images are on top of each other. This fits your "unless there is a nearby and similar overlaid pattern" maybe. If this is the case then moiré could be generated by the pulses of light and movement of the bird.

Something is not quite right, Bob. The pulse width does not tell you the rate at which pulses are being emitted. Therefore, it can not be said that exactly 7 pulses are emitted while the slit is in a particular position on the frame.

Someone here should know for sure but I'm wondering if, in HSS, the pulses are sync'd to the shutter slit as it steps up the image plane ....

[edit] found this:



This throws a bit more light on the subject but still beyond my ken ...

https://www.scantips.com/lights/flashbasics2b.html
.

[BobGilbody]

Thanks for all the information Ted.
It seems to me that we are closing in on an answer. I was in error about the duration vs pulse thing. This highlights a problem that concerns me. The problem is that there is a great difference between very little and nothing.

"Something is not quite right, Bob. The pulse width does not tell you the rate at which pulses are being emitted. Therefore, it can not be said that exactly 7 pulses are emitted while the slit is in a particular position on the frame."

My understanding is that there is a difference between the 'slit' and an area of the sensor that is not covered by the shutter curtains. That area of the sensor is exposed to the pulsating light for a certain period of time. During that period of time there are several pulses of light. Not exactly 7 but more than one. The duration of the pulse must be short enough for there to be pulses rather than fluctuations in intensity.

"Someone here should know for sure but I'm wondering if, in HSS, the pulses are sync'd to the shutter slit as it steps up the image plane in which case there is only one pulse per slit position and in which case any patternation observed due to subject motion is not moiré ..."

Not sure the shutter slit steps up the image plane however it is possible that when using more than one speedlite the light pulses from the individual speedlites are not in sync. I use one Canon 600EX-RT [ I make this the master ] and up to three 580EXII speedlites.

I would like to think others could help us out here. I find this interesting but my hope is to get a better image of the birds that visit the granny flat.

All the best Bob

[BobGilbody]

Hello Ted,
I just saw the extra bit on your post. This is strange.

" HSS flash is triggered repeatedly and very rapidly, so that it mimics being "on" like continuous light (like sunlight or a desk lamp is on continuously)."

"The purpose of the "flash" being made to be continuous is because there is no concept of sync with any continuous light. The continuous light is always there, so any shutter speed works (which is the reason it is called High Speed Sync — just meaning there is no sync issue at any shutter speed). ANY continuous light does HSS, so to speak. Like sunlight, HSS simply works at any shutter speed. But like sunlight, a very fast shutter speed reduces the light, which then requires a very wide aperture in compensation (Not true of speedlight mode, with a very short flash exposure that is independent of shutter speed).
However make no mistake, HSS is NOT high speed flash. It is merely called High Speed Sync, since continuous light has no sync requirements. HSS no longer works same as "flash", and is no longer a speedlight — continuous light has no motion stopping capability at all (HSS does still have the shutter speed, but the speedlight is typically faster than any possible shutter speed)."

The writer goes from 'mimics continuous light' to discuss 'continuous light'.

Obviously others have problems with this.

Bob

PS

The diameter of a feather barb is of the order of 50 micron. [ 1/20 of a mm ]
If I photograph a bird traveling at 20 kph [ 5.5 m/s or 5500 mm/sec or 5 500 000 micron/s ] I will need a very high speed something to freeze motion, maybe a lot higher speed to prevent 'moiré' type problems.

My old brain is beginning to hurt.

Bob

[xpatUSA]

My understanding is that there is a difference between the 'slit' and an area of the sensor that is not covered by the shutter curtains.

Oh dear ...

That area of the sensor is exposed to the pulsating light for a certain period of time. During that period of time there are several pulses of light. Not exactly 7 but more than one. The duration of the pulse must be short enough for there to be pulses rather than fluctuations in intensity.

Not sure the shutter slit steps up the image plane however it is possible that when using more than one speedlite the light pulses from the individual speedlites are not in sync.

Focal plane shutter action is clearly shown about half-way down in this page:

https://www.wrotniak.net/photo/tech/fp-shutter.html

It might be stepping down rather than up but it does seem to stepping rather than continuously moving.

I think I'm done; I don't want to confuse anybody with my guesswork, especially in the flash department.

Take care,

[xpatUSA]

PS

The diameter of a feather barb is of the order of 50 micron. [ 1/20 of a mm ]
If I photograph a bird traveling at 20 kph [ 5.5 m/s or 5500 mm/sec or 5 500 000 micron/s ] I will need a very high speed something to freeze motion, maybe a lot higher speed to prevent 'moiré' type problems.

Maybe so but the barbs on my Texas crow feather remain spaced 1/2mm apart.

As we know, everything is bigger in Texas, yee-hah ...

[BobGilbody]

Thanks for the chat Ted. We gave it a good try.

Bob

[xpatUSA]

Thanks for the chat Ted. We gave it a good try.

Bob

More about HSS: pulses versus "continuous light". Here are some time graphs, with HSS at left:



The main thing to note is that the light output is virtually continuous with a slight ripple on it. So the concept of multiple exposures at a slit position is moot and the likelyhood of moiré being caused by motion seems quite small to me. In other words, something that looks like moiré to your good self might be just the previously-mentioned effect of shutter action at higher than sync speed causing distortion in a moving object.

[BobGilbody]

Hello Ted,

Thanks for finding the graphs. I'm confused, normal state of affairs. Couple of questions:

1. There is only one light source [ assume all this is taking place in a dark room ]. The first graph shows that there is not virtually continuous light output, but that there is continuous light output of varying intensity ??

2. The curtains appear to be heading to infinite speed.

I'm more than likely totally lost. Either that or the graphs are not what I think they are. Same thing I guess.

All the best Bob

PS Just to throw an extra clog in the gear box: What level of light intensity is required to produce an effect on the sensor?