moire on bird feather

[xpatUSA]

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 ??

Yes, that is why I said "ripple" and "virtually". Taking the peaks as a normalized '1' the valleys are at say 0.9 which is a change of -0.15 EV - not much in terms of exposure.

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.

Either you can go on-line and research all things HSS or I can go and then tell you what I think which you may not then understand. Could take a while to reach agreement by the looks of it ...

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?

That puts us at the level of Poisson statistics ... if one photon at 555nm wavelength whacks my green sensor pixel, there is a 35% chance that an electron will be generated and stored as a charge in the photodiode capacitance during exposure.

So maybe enough photons that the ADC output would be raised by one least significant bit with a 95% confidence level.

[BobGilbody]

Thanks for the reply Ted. I will take some more photographs and see what I get.

All the best Bob

[Stagecoach]

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

An annoying minor bird left a feather on my coffee table this morning which got me thinking. Whilst we talk about the flight speed of a bird as a mass moving roughly in one direction at a speed, what is the speed of movement of fine downy feathers fluttering in the wind? I have absolutely no idea.

Being bored I decided to play, the feather placed in a clamp, light source just one speedlight in TTL set to HSS at any shutter speed above 1/320 and a good wind source.

Static at 1/250s


Considerable movement of feather from floor fan 1/250s


Considerable movement of feather from floor fan 1/4000s (HSS)


Considerable movement of feather from floor fan 1/8000s (HSS)


Shots were also taken at 1/500s, 1/1000s, 1/2000s and results similar to all others.

With the amount of movement (fluttering) of the feather it appears that if flash is going to be the only subject light source any speed between 1/250s to 1/8000s is going to freeze all movement. I may try and repeat this with an increased DOF by upping the ISO but the lower right fine feathers give the best indication.

[xpatUSA]

Originally Posted by 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 ??

Yes, that is why I said "ripple" and "virtually". Taking the peaks as a normalized '1' the valleys are at say 0.9 which is a change of -0.15 EV - not much in terms of exposure.

Going back to the graphs:



Looking at the middle graph first, the amount of exposure is depend on flash intensity and time period of the flash - because Exposure (H) = incident lux times seconds. So, to get the exact exposure, it is necessary to integrate from the first dashed line to the second dashed line for the flash values which here are shown as that little gray peak at the top. Caveat not all of the pulse is shown, see typical pulse wave-forms here:

http://kronometric.org/phot/exp/flas...alf-eighth.jpg

The HSS graph above is more difficult because the degree of exposure becomes dependent upon both the flash intensity and the position of front and rear curtains F and R - each of which takes about 5ms to go from closed to fully open or vice versa.

And just to make things hard to visualize, the front curtain F is going from closed at the bottom to open at the top, while the rear curtain R is going from open at the bottom to closed at the top!

In the HSS graph, the x-axis T is about 9ms long. But the flash power is super-imposed on the position versus time curves. So, determining how the sensor exposure proceeds with respect to time is less than easy.

[BobGilbody]

Hello Grahame,
Interesting approach to things. It removes the problem of 'getting the focus right'.

If I saw the photograph without your explanation of how you took the photograph how would I know the 'wind was blowing'?
If I look at a photograph of a bird in flight I 'know' it is moving, even if the image 'freezes motion'.

Need to think about that more.

All the best Bob

[BobGilbody]

Hello Ted,
Thanks for trying to explain the graphs. My mind tends to wonder off at tangents so I find it difficult to 'get' some things.

Just as an aside you might be interested in two papers I found several years ago:

https://dash.harvard.edu/bitstream/h...=1&isAllowed=y

https://www.jstor.org/stable/687225

I no longer have jstor. I'm sure the paper is somewhere on the net.

Needless to say I don't understand either paper :-(

All the best Bob

Set up for BiF:

Thought I would post photos of how I am photographing BiF. I don't get out much these days but fortunately I have a granny flat that allows me to photograph visiting birds. Species tend to be territorial around here so the birds decide which dumb human they can adopt. My regulars are Sulphur Crested Cockatoos, Rainbow Lorikeets, King parrots, and Scaly Breasted Lorikeets. Occasionally Kookaburras, Butcher birds, Galahs, Currawongs, and Pale-headed Rosellas. I was a bit worried the flash might be unethical so I did a selfie up against four speedlites set up for TTL. I didn't have problem so the birds should be ok.

Me by Bob Gilbody, on Flickr

Yesterday I set up a vertical perch and feed bowl on the fence to the deck. I positioned the feed bowls so that the Lorikeets have to sit on top of the stick to get the seed. They fly onto the top of the stick about 50% of the time. I can have the camera just inside the granny flat and still be within 2m of the nearest vertical perch. One speedlite [ master ] is on the camera, three others are on a tripod on the deck.

[/url]25 04 22-3661 by Bob Gilbody, on Flickr

The cockys had it all figured out within five minutes.

Bob

[xpatUSA]

Hello Ted,
Thanks for trying to explain the graphs. My mind tends to wonder off at tangents so I find it difficult to 'get' some things.

Hello again, Bob. I've been reading about just the shutter curtain action separately from the ever-confusing HSS.

As far as I can tell, the curtain does not move in steps as I may have implied earlier but moves as fast it can go from closed to open and vice-versa. That would be about 5ms from stop to stop - no matter what shutter speed is selected. As it passes over the sensor the curtain speed should theoretically be constant but varies a little in practice with real hardware. Also, although the HSS "slit" should have a constant time gap, that is not always so, especially at very high shutter "speeds" e.g. 1/8000 sec.

Have a look here for some more numbers and info:



https://www.dpreview.com/forums/thread/2407311

The one thing I get from the above is that, for shutter speed selections above the so-called sync speed, there is a chance of increasing image distortion due to motion, especially in the vertical. For example a wing leading edge might be rendered as more curved than it really is. That is because the actual time to expose the whole frame remains at about 1/200 sec - which is ample time for motion effects to occur ... as opposed to shooting at or below sync speed - where the same effects would render as motion blur.

More reading material for ya:

https://www.penmachine.com/2008/09/c...ashes-and-sync

I'll opine that the focal plane shutter may not be best for BIF and that maybe a leaf shutter-based camera would be "better"!

Thank you Bob, your thread has caused me to understand focal plane shutters much mo' better than before.

[xpatUSA]

The one thing I get from the above is that, for shutter speed selections above the so-called sync speed, there is a chance of increasing image distortion due to motion, especially in the vertical. For example a wing leading edge might be rendered as more curved than it really is. That is because the actual time to expose the whole frame remains at about 1/200 sec - which is ample time for motion effects to occur.

Found this which illustrates some bird wingtip speeds:

https://www.nature.com/articles/s414...47-3/figures/1

Fig 1 shows a graph at right for bird wingtip velocity in the vertical plane which they call Vz. Peak velocity say 10 meters per sec, approximately sinusoidal velocity waveform.

[BobGilbody]

Hello Ted,
Thanks for the two recent posts. I think we have both arrived at a better understanding of the shutter process and HSS. The bird wing thing is very interesting. Birds are amazing creatures.
You mention that a leaf shutter might be the thing to use for BiF. I am more inclined to predict that the clever people that build camera bodies will eventually provide us with the 'electronic' shutter that allows high speed cameras to 'freeze' bullets in flight. My understanding of those is that they have two polarizing grids at right angles one of which can get flipped by putting a voltage across it. All very Dr Who sonic screwdriver type stuff. It seems to me that the only way that camera manufacturers can stay afloat is to sell new bodies to eager amateurs like us every few years. If they gave us too much in one hit they would be out of a job. [ My take on consumer economics. ]

I will post again with images of a cocky in flight. The photos are not that interesting as photos but do record feathers that might be of interest.

All the best Bob

Photos of cockys in flight:

hs flash 25 03 22-1049-DeNoiseAI-standard-SharpenAI-Focus by Bob Gilbody, on Flickr

testing flash cocky 10 04 22 -3446-DeNoiseAI-standard by Bob Gilbody, on Flickr

[/url]14 04 22-3553-DeNoiseAI-standard-SharpenAI-Motion by Bob Gilbody, on Flickr

[/url]16 04 22-3561-DeNoiseAI-standard-SharpenAI-Motion by Bob Gilbody, on Flickr

[xpatUSA]

Hello Ted,
Thanks for the two recent posts. I think we have both arrived at a better understanding of the shutter process and HSS. The bird wing thing is very interesting. Birds are amazing creatures.

Quite so. Motion distortion is well-known in the industry:

https://www.researchgate.net/publica...s/figures?lo=1

You mention that a leaf shutter might be the thing to use for BiF.

I meant better than focal plane, not "the thing to use" ...

I am more inclined to predict that the clever people that build camera bodies will eventually provide us with the 'electronic' shutter that allows high speed cameras to 'freeze' bullets in flight. <>

Yep, Sony announced a big 128MP "global shutter" sensor over a year ago:

https://ymcinema.com/2021/03/09/sony...hutter-sensor/

Might cost you a buck or two ...

... or buy a Panasonic GH6 https://inf.news/en/photography/8b8c...726ba009d.html

... or buy a camera with a CCD sensor!

"Many, but not all, CMOS sensors use rolling shutter. In contrast to that, “global shutter” is the technical term referring to sensors that scan the entire area of the image simultaneously. The vast majority of CCD sensors employ global shutter scanning."

https://everythingwhat.com/what-does...-final-cut-pro
.

[BobGilbody]

Thanks for the info Ted. I was looking at some photos of a king parrot and realised that odds and Manfred might have had the answer a while back. But then again maybe there is more going on than meets the monitor.

28 04 22 LR1-3623 by Bob Gilbody, on Flickr

This image shows little 'moiré' [ no longer sure what moiré is in this context ]. Why it shows little is beyond me.

28 04 22 LR1-3621 by Bob Gilbody, on Flickr

This image is interesting. If you zoom in on the image you will see how the 'moiré' changes and finally disappears. This may confirm odds and Manfreds take on things. My question now becomes: 'why does the first image have different characteristics compared to the second?'

All the best Bob

[xpatUSA]

Thanks for the info Ted. I was looking at some photos of a king parrot and realised that odds and Manfred might have had the answer a while back. But then again maybe there is more going on than meets the monitor.

28 04 22 LR1-3623 by Bob Gilbody, on Flickr

This image shows little 'moiré' [ no longer sure what moiré is in this context ]. Why it shows little is beyond me.

28 04 22 LR1-3621 by Bob Gilbody, on Flickr

This image is interesting. If you zoom in on the image you will see how the 'moiré' changes and finally disappears. This may confirm odds and Manfreds take on things. My question now becomes: 'why does the first image have different characteristics compared to the second?'

All the best Bob

It would seem that Odd and Manfred do indeed have the answers you need. Sorry to have wasted your time.

[BobGilbody]

Hello Ted,
I don't think either of us have wasted our time. We have a better understanding of things thanks to our exchange of posts. You have raised many interesting points. I'm now interested in the difference between a 5Ds and a 5Dsr. If the 'answer' is 'the monitor did it' then why have two models of the same camera? Is there aliasing that is not fixable by simply zooming in?

All the best Bob

[xpatUSA]

Hello Ted,
I don't think either of us have wasted our time. We have a better understanding of things thanks to our exchange of posts. You have raised many interesting points. I'm now interested in the difference between a 5Ds and a 5Dsr.

If the 'answer' is 'the monitor did it' then why have two models of the same camera?

Hello again, Bob. I am not familiar with those cameras. What is the difference?

Is there aliasing that is not fixable by simply zooming in?

The fundamental basis of aliasing is failure to render correctly detail that is above half the sampling rate (Shannon, Nyquist et al).

There are two or more sampling rates involved.

1. The sensor.

The sampling rate is 1 over the pixel pitch. Anything in the image plane over half that rate will get aliased. My Sigma: anything over 54 lp/mm.

2. The monitor.

The sampling rate is 1 over the pixel pitch. My monitor: anything over 48 lppi.

Therefore, an image straight out of the camera viewed at less than 100% zoom can get aliased beyond any sensor aliasing that is already in the captured image.

3. Resampling.

Now let's talk about zooming out and in - with smoothing turned off, meaning that the operating system re-samples for the screen in Nearest Neighbor mode.

Download this full-size (right-click > Save link As ... ) and put it up on your screen:

http://kronometric.org/phot/processi...gs_lg_orig.png

At 100 per cent zoom there is no aliasing.

Now zoom out, say to 60 per cent ... voila!!

But at 200 per cent with that image still open - there is stlli no aliasing apart from that each original pixel is now 2x2 pixels.

Now re-size (not zoom), say 60 per cent, save the image and then view the saved image at 100 per cent. There will be aliasing. Then zoom in to 167 per cent. The aliasing will still be there!!

This proves that image moiré (i.e. sensor-caused) can NOT be fixed by zooming in IF it is already in the image. In other words, there is indeed aliasing that is not fixable by simply zooming in.

Not an easy subject which can be quite confusing, Bob.

[xpatUSA]

Bob.

Can't help but notice that your images are named e.g. "20 04 22 rainbow-3616-DeNoiseAI-standard-SharpenAI-Focus by Bob Gilbody, on Flickr ". It seems to imply that all your images are de-noised then later sharpened and also have something called "standard" applied.

My point being that if one is examining the capture quality of an image, the best amount of post-processing is none.

[BobGilbody]

Thanks for the latest two posts Ted. I'm feeling much much better about the problem now. You are right to point out that any processing of the image should be done after analysis of the sensor image to avoid confusions arising. One interesting point for me was that I realised the slight wet macular degeneration I have in my right eye affects how I see things. [ I get a needle in the eye this morning, it happens every four to seven weeks ]
Thanks for your patience explaining things to me, much appreciated.

All the best Bob

[xpatUSA]

... One interesting point for me was that I realised the slight wet macular degeneration I have in my right eye affects how I see things. [ I get a needle in the eye this morning, it happens every four to seven weeks ]

I know the horrible feeling of a needle in the eye all too well, Bob. A couple years ago I had swollen maculae requiring injection of stuff. No fun at all ...

[BobGilbody]

Hello Ted,
The thought of it is a bit iffy but the guy that looks after me is a gifted specialist type surgeon. I'm lucky to live where I do and when I do. Forty years ago things would be very different.
It is of wider interest in that as far as I know we see with the brain not the eye. The eye is not much of an optical instrument but the brain manipulates the image to make it better. In that sense we do need to learn to see.

All the best Bob