How fast of a shutter speed to eliminate the need for IS in a lens?

[BadtzMaru]

Please reread what I wrote, "Please expand."

Please reread what I wrote, "A fast shutter speed does not counter a
static scene in low light"

Again, read the comment, you are focusing on one part of a multipart statement. I see no need to expand on the obvious.

 

[Blue II]

That does make sense, thanks.

I know what you said, but I have no idea what you meant. There is no
movement in a static scene to counter, so I have no idea what you're
trying to say.

Please reread what I wrote, "Please expand."

What he is saying is: In low light, even if the scene is static,
you need a long
shutter speed (to expose properly). Long shutter = shake. IS can
help with that,
but a fast shutter won't (because you'll under expose). A
tripod/monopod would help to. IS doesn't help when you need a fast
shutter to freeze motion.

Ken

Please reread what I wrote, "A fast shutter speed does not counter a
static scene in low light"

--
Steve

--
Steve

 

[Blue II]

Maybe you haven't heard, but communication is a 2-way process. I don't know what the big deal is about making yourself understood more clearly, but if you want to be an a$$, so be it. You've succeeded.

Please reread what I wrote, "Please expand."

Please reread what I wrote, "A fast shutter speed does not counter a
static scene in low light"

Again, read the comment, you are focusing on one part of a multipart
statement. I see no need to expand on the obvious.

--
Steve

 

[quadra950]

See the tips & tricks column on the left side for suggested shutter speed at given focal length: http://www.pebbleplace.com/Personal/Canon_200L.html

 

[BadtzMaru]

I couldn't have said what I said any more clear. If you didn't understand it then so be it. Other people understood what was written so it obviously made sense. I don't see how you could read the entire sentence and not understand it, if you wish to get defensive and spew vitriol go ahead, its only the internet, wont bother me none.

 

[PicOne]

I would imagine reproduction size is also a major factor in determining Tv appropriate to avoid visible shake blur. Are the 1/fl or 1/(2*Fl) rules based, as for DOF uses, on an assumed 8x10 print viewed from X distance?

It sounds like, in general, 1/500 or faster is the "take home"
message here.

If that is the message you are getting, then you are not reading the
answers closely enough. It depends on focal length, and the options
you presented only apply to long focal lengths (like 300mm). On your
17-55, if you are shooting at 17, you should be able to use 1/30 very
easily and not need IS. Now, you won't freeze motion at that speed,
but you should not see camera shake either. Even on the long end of
that lens at 55mm you should easily be able to use 1/100 without IS.
I routinely shoot at various lengths using 1/60 without IS and get
sharp results.

Another thing to keep in mind is subject distance. Part of the reason
that 1 / (2*f*crop) is just a "Rule of Thumb" is that subject
distance plays a large role as well. Taking a distant landscape at
50mm, you could probably get away with 1/30 or so, but when taking a
close up at only a couple feet, even 1/100 might suffer from camera
shake if you're not steady.
--
----------
Andrew Melvin Helmboldt

'A wise man talks because he has something to say; a fool talks
because he has to say something.' -Plato

--
'Everything in photography boils down to what's sharp and what's fuzzy.'
-Gaylord Herron

 

[Blue II]

Obviously not true since others explained it more clearly. I don't see what the problem was with trying to state it differently unless you were trying to be difficult or just think you're better than everyone else. Sorry if that offends, but that's the impression you gave me. Now, back to your regularly scheduled thread...

I couldn't have said what I said any more clear.

--
Steve

 

[PicPocket]

A simple way to look at it is to pick a pixel and move the sensor
linearly. The pixel will move the same distance whether it contains
the image of a close or distant object.

This proof is simply wrong, because you are assuming wrong. Every pixel will "not" move the same distance. Only the pixels at the center will move the same distance, provided the distance is small enough, because as soon as the pixel is off center due to movement, it would start moving by a scale determined by the angle of spread and distance of image plane. This is the same reason that in the cone of view the cross section covered at each distance will increase in area as you move farther away

Another proof is none of the
in-body IS will work if images from different distances move
differently on the sensor.

Again, you are proving something wrong here. The shake is determined at the sensor and the sensor gets compensated by the determined amount. This has nothing to do with what is at 20m away or 40m away. They still fall within the same cone of view. The lateral shift of a point at 40m away might be less than that of a point 20m away, but at the same time the resolution of the point 40m away is also less that the nearer point on the sensor plane. You will need a proportionally bigger point farther away to experience a similar shake on sensor plane. The output hence is mostly unaffected as long as sensor is correctly compensated

BTW with a PH.D in physics I hope I can still
understand HS optics.

If this is indeed true, either it's been a long time and you have not been revising; or they didn't teach you very well; or perhaps you specialized in some other field altogether. If you can think of any other reason why you can be wrong and still brag about the qualification, just let us know. Oh yes, another reason may be that people sometimes simply make mistakes (I may have made a few here, and frankly, I don't have a degree in optics)

In any case, I think if you are into applied Physics rather than theoretical science, the experiment suggested by Gina should help to clear it

--
PicPocket
http://pictures.ashish-pragya.com/GalleryIndex.html

 

[jackma]

Could IS harm the sharpness of pictures taken at high shutter speeds?

I use my Canon 70-200 f4 IS lens for sports. I typically keep my shutter at 1/1500, and I often crop the picture insanely e.g. to a 50% or 100% enlargement on screen. I have not been using the IS.

At shutter speeds of 1/1500, is it possible that IS could help image sharpness? Is it possible that IS could harm image sharpness?

The reason why I have not been using IS is that I imagine that the IS mechanism applies very high-frequency adjustments to the image to keep it stabilized. I can imagine that these adjustments might actually exacerbate image blur if the image is greatly enlarged (as I do). Does anyone have actual testing experience with this?

 

[boels069]

Maybe our Rule of Thumb shouldn't be a function of lens focal length
at all, but rather a function of the width of the scene at the focal
distance (which would be a measure of the relative detail expected)?

Sharpness can be expressed as a number equal to the ratio of the actual AOV (angle of view) of the lens and the angle introduced by camera shake.

 

[boels069]

http://www.enginova.com/Minimum%20Shutter%20Speed.htm

 

[PicPocket]

I would imagine reproduction size is also a major factor in
determining Tv appropriate to avoid visible shake blur. Are the
1/fl or 1/(2*Fl) rules based, as for DOF uses, on an assumed 8x10
print viewed from X distance?

I think there would be some number for sure

But this reminds me of another rule of thumb of sorts. That if you take an image that fills your vision, you can enlarge it as much as you want without any loss of percieved sharpness as long as your field of vision doesn't change

In fact thats what is already happening with increasing mega pixels. Higher sensor resolution means that at pixel level shake will be visible much sooner, but the real question is - does it matter if our output sizes (monitor/print/whatever) doesnt increase propertionally, or if we start looking from farther away

--
PicPocket
http://pictures.ashish-pragya.com/GalleryIndex.html

 

[carlk]

No problems just that I wasn’t thinking. I agree linear camera movement does affect close objects more than far away ones. However I stick to my guns this part of camera shake contributes very little to the total shake as I originally stated. The effect of the total shake probably is not much distance dependent. If you don’t agree you need to answer the question how could in-body IS could work if this is not the case? How does the sensor move to compensate distant and close objects with the same algorism or at the same time? I have not heard the answer yet.

A simple way to look at it is to pick a pixel and move the sensor
linearly. The pixel will move the same distance whether it contains
the image of a close or distant object.

This proof is simply wrong, because you are assuming wrong. Every
pixel will "not" move the same distance. Only the pixels at the
center will move the same distance, provided the distance is small
enough, because as soon as the pixel is off center due to movement,
it would start moving by a scale determined by the angle of spread
and distance of image plane. This is the same reason that in the cone
of view the cross section covered at each distance will increase in
area as you move farther away

Another proof is none of the
in-body IS will work if images from different distances move
differently on the sensor.

Again, you are proving something wrong here. The shake is determined
at the sensor and the sensor gets compensated by the determined
amount. This has nothing to do with what is at 20m away or 40m away.
They still fall within the same cone of view. The lateral shift of a
point at 40m away might be less than that of a point 20m away, but at
the same time the resolution of the point 40m away is also less that
the nearer point on the sensor plane. You will need a proportionally
bigger point farther away to experience a similar shake on sensor
plane. The output hence is mostly unaffected as long as sensor is
correctly compensated

BTW with a PH.D in physics I hope I can still
understand HS optics.

If this is indeed true, either it's been a long time and you have not
been revising; or they didn't teach you very well; or perhaps you
specialized in some other field altogether. If you can think of any
other reason why you can be wrong and still brag about the
qualification, just let us know. Oh yes, another reason may be that
people sometimes simply make mistakes (I may have made a few here,
and frankly, I don't have a degree in optics)
In any case, I think if you are into applied Physics rather than
theoretical science, the experiment suggested by Gina should help to
clear it

--
PicPocket
http://pictures.ashish-pragya.com/GalleryIndex.html

 

[BadtzMaru]

I couldn't have said what I said any more clear.

--

No worries, like i said, its the internet, doesn't bother me one way or the other. As for explaining more clearly, so long as you got it, no problem, I don't see what was said any different, it just used more words. I wasn't trying to be difficult and I know I'm not better than anyone else, so lets chalk it up to me not understanding what you weren't understanding and we can all have fun shooting instead of participating in unneeded forum silliness.

 

[kward]

Could IS harm the sharpness of pictures taken at high shutter speeds?

I use my Canon 70-200 f4 IS lens for sports. I typically keep my
shutter at 1/1500, and I often crop the picture insanely e.g. to a
50% or 100% enlargement on screen. I have not been using the IS.

At shutter speeds of 1/1500, is it possible that IS could help image
sharpness? Is it possible that IS could harm image sharpness?

The reason why I have not been using IS is that I imagine that the IS
mechanism applies very high-frequency adjustments to the image to
keep it stabilized. I can imagine that these adjustments might
actually exacerbate image blur if the image is greatly enlarged (as I
do). Does anyone have actual testing experience with this?

It takes IS a little time to settle when you move the lens around, so if you're
moving the lense alot, I could see it having some effect. I have a 70-300 IS
and think I've seen some IS blur when I've had fast shutter and I'm trying to

follow quick action, take a quick snap, but its far from good test (and the IS on
the 70-300 might not be as good at this as a L's IS).

Ken

 

[Gena]

A good IS system talks to the lens, it knows the aperture, the focal length, the focusing distance, and DOF. It also determines the angular vs linear shake and then it makes a decision

A simple way to look at it is to pick a pixel and move the sensor
linearly. The pixel will move the same distance whether it contains
the image of a close or distant object.

This proof is simply wrong, because you are assuming wrong. Every
pixel will "not" move the same distance. Only the pixels at the
center will move the same distance, provided the distance is small
enough, because as soon as the pixel is off center due to movement,
it would start moving by a scale determined by the angle of spread
and distance of image plane. This is the same reason that in the cone
of view the cross section covered at each distance will increase in
area as you move farther away

Another proof is none of the
in-body IS will work if images from different distances move
differently on the sensor.

Again, you are proving something wrong here. The shake is determined
at the sensor and the sensor gets compensated by the determined
amount. This has nothing to do with what is at 20m away or 40m away.
They still fall within the same cone of view. The lateral shift of a
point at 40m away might be less than that of a point 20m away, but at
the same time the resolution of the point 40m away is also less that
the nearer point on the sensor plane. You will need a proportionally
bigger point farther away to experience a similar shake on sensor
plane. The output hence is mostly unaffected as long as sensor is
correctly compensated

BTW with a PH.D in physics I hope I can still
understand HS optics.

If this is indeed true, either it's been a long time and you have not
been revising; or they didn't teach you very well; or perhaps you
specialized in some other field altogether. If you can think of any
other reason why you can be wrong and still brag about the
qualification, just let us know. Oh yes, another reason may be that
people sometimes simply make mistakes (I may have made a few here,
and frankly, I don't have a degree in optics)
In any case, I think if you are into applied Physics rather than
theoretical science, the experiment suggested by Gina should help to
clear it

--
PicPocket
http://pictures.ashish-pragya.com/GalleryIndex.html

 

[Leon Wittwer]

for those who shoot for a more or less standard print size, what is necessary can fall short of what is possible. While higher pixel density allows larger prints at a given viewing distance and requires more careful control of motion, if you don't ever print larger prints, the effort is wasted.

^^ agreed on similar print.

however we live in a day and age where sharpness is measured on a
monitor at 100% and unrealistic expectations of that are through the
roof.

Yes. Many seek a level of resolution and sharpness that they almost never use in practice. I occasionally do shows where I display my photography. I have never seen a person at a show take out a 8x loupe and comment on the lack of sharpness of my photographs. If a photograph is "good," somehow these "technical" aspects that some obsess over, don't seem so important. I do try to do as well as I can in all the aspects of shooting, processing and printing to get the best shots but I end up virtually always compromising something.
--
Leon
http://homepage.mac.com/leonwittwer/landscapes.htm

 

[boels069]

Maybe our Rule of Thumb shouldn't be a function of lens focal length
at all, but rather a function of the width of the scene at the focal
distance (which would be a measure of the relative detail expected)?

Sharpness can be expressed as a number equal to the ratio of the
actual AOV (angle of view) of the lens and the angle introduced by
camera shake.

A = diagonal angle of view in degrees
P = number of pixels on diagonal

Number of pixels per degree:

N = P / A

S = camerashake in degrees per second

Blurr due to camera shake
in pixels
:

B = N * S pixel per second

B = 1 for a "unreal sharp" image:

T = 1/(N * S) second

Example EOS 5D and 100 mm lens

A = 24 degrees
P = 5250 pixels

N = 5250/24 = 218.75

S = 2.5 degrees/second

T = 1/(219 * 2.5) = 1/547 second

S = 2.5 is a more or less arbitrary value I found somewhere (lost the link)
B = 1 means a very thin image dot moves 1 pixel

Below a tabel for the 40D with above parameters:

Focal length - 1/shutter speed (rounded to multiples of ten) for 40D
10 - 110
20 - 180
30 - 250
40 - 320
50 - 400
60 - 470
70 - 550
80 - 630
90 - 700
100 - 780
110 - 860
120 - 930
130 - 1010
140 - 1090
150 - 1170
160 - 1240
170 - 1320
180 - 1400
190 - 1470
200 - 1550

 

[Hawkman]

The rule of thumb is a shutter speed faster than 1/ (crop factor * focal length). That is an average, some do better, some worse. It is about right for me.

Since you have a 1.6 crop factor body, for a 400mm lens this would be 1/640s or faster, for a 135mm lens about 1/250s or faster...

The things you can do to get these shutter speeds are raise the ISO, open up the lens aperature, or both

• Gene

--
Gene (aka hawkman) - Walk softly and carry a big lens
21
Please visit my wildlife galleries at:
http://www.pbase.com/gaocus
http://hawkman.smugmug.com/gallery/1414279

 

[genotypewriter]

As the other person said, if you're moving the lens around a lot and taking shots, the IS might affect in a negative way for any shutter speed.

With higher shutter speeds the motion's frozen... so there's definitely less blur. If such a blurring is introduced by the IS, then it should also equally soften the image at slower shutters, if not worse.

You said that you use your lens for sports and unless you're panning, people don't use IS for moving subjects. So this might be your problem when you have used IS.

GTW
--
http://www.flickr.com/genotypewriter