Focal length and shutter speed

casp3r

casp3r

Suspended / Banned
Messages
77
Name
Michael
Edit My Images
Yes
I was a very keen photographer many years ago and have just bought my first digital SLR with a standard lens kit which included a 70-300mm lens. Now with 35mm lens I was always told that I should keep the shutter speed roughly equilavent to the focal length of the lens. So in the case of this lens 1/60 - 1/320.
But with the DSLR does it not have a conversion factor, in my case (EOS400D) 1.5x? So the shutter speeds should be 1/125 - 1/500 (105-450mm)? Would I be right in thinking this?

Sorry if this is a dumb question?

Michael
 
Spot on :thumbs:

:thinking:
 
Actually the focal length of the lens stays the same no matter what body its attached to, so the shutter speed to focal length thing would stay the same as with a 35mm camera.

The "conversion" factor, is merely a cropping factor, this means that any image will fill the image as if the lens were 1.6X longer, but the lens is still the same length.

The shutter speed to focal length also only applies to cancelling out your own body movements and shake while taking the picture, it may need to be increased to freeze motion of moving objects etc to get the shot you want.

My understanding of it anyway :)
 
Ian is right the lens focal length is the same no matter what camera it's on and it lets in the same amount of light, the crop factor doesn't come into it. I remember the rule of thumb as being "The shutter speed should be faster than the focal length" to avoid camera shake when handheld. i.e. 1/60th @ 50mm.
 
What the fellers say - you don't use the crop factor conversion at all - the focal length stays the same. :)
 
I'm going to have to disagree I'm afraid. Whilst the focal length stays the same, the rule of thumb has been developed as a guideline to avoid visible shake on prints and slides.

So, although the cropped sensor is smaller, if you print to the same sized enlargement as you would from a 35mm neg. then you get the equivalent of a longer lens from a field of view to linear inches on the print point of view.

Therefore, you should include the crop factor is calculations of 1/f for the minimum shutter speed. But, remember it is a guideline and good technique can help lengthen the shutter speed for a given focal length.

Likewise, cropped pictures are more demanding on sharpness than uncropped.

Paul
 
grumpybadger said:
I'm going to have to disagree I'm afraid. Whilst the focal length stays the same, the rule of thumb has been developed as a guideline to avoid visible shake on prints and slides.

So, although the cropped sensor is smaller, if you print to the same sized enlargement as you would from a 35mm neg. then you get the equivalent of a longer lens from a field of view to linear inches on the print point of view.

Therefore, you should include the crop factor is calculations of 1/f for the minimum shutter speed. But, remember it is a guideline and good technique can help lengthen the shutter speed for a given focal length.

Likewise, cropped pictures are more demanding on sharpness than uncropped.

Paul
Click to expand...

I used to think like you but CT cleared it up for me in the following way:

Focal length remains the same whether the glass hangs in front of a full frame or smaller sensor.

The fact that the smaller sensor records a smaller fraction of the image circle makes the lens no longer physically - it only appears that way.

Therefore if you could take a piccie at 1/200 sec at 200mm focal length you could do so as easily whether you have a 70-200 on a 20D or 5D.

Only thing is the image will be different.

In any event no two of us are equally stable so it is generally not a bad idea to err on the safe side until you know exactly what you can and cannot get away with!

:thumbs:
 
Anton Roland said:
I used to think like you but CT cleared it up for me in the following way:

Focal length remains the same whether the glass hangs in front of a full frame or smaller sensor.

The fact that the smaller sensor records a smaller fraction of the image circle makes the lens no longer physically - it only appears that way.

Therefore if you could take a piccie at 1/200 sec at 200mm focal length you could do so as easily whether you have a 70-200 on a 20D or 5D.

Only thing is the image will be different.

In any event no two of us are equally stable so it is generally not a bad idea to err on the safe side until you know exactly what you can and cannot get away with!

:thumbs:
Click to expand...

I still disagree. Since shake is caused by light falling on adjacent pixels, it is actually a function of pixel density. I do agree that a crop doesn't change the focal length of the lens, but the higher pixel density of smaller sensor cameras means that it takes a smaller movement to cause light to fall on adjacent sensing pixels.

My point is that with the smaller sensor (and hence the so called 1.6 multiplier) you have to enlarge the picture by a further 1.6x to generate the same size print (I'm assuming the aim of photography is to generate prints - could be a wrong assumption for some people).

Therefore, the shake will be more noticable on the 1.6 crop camera for the same sized print.

The rule of thumb was convenient for 35mm and that was (is) pure luck. It's about the angle the final image you view subtends. If you are holding a pair 12x8 prints in your hand, each pixel will represent a smaller angle of view when taken on a 1D Mk II or a 5D. Therefore, the stability requirement will be higher.


Paul
 
grumpybadger said:
I still disagree. Since shake is caused by light falling on adjacent pixels, it is actually a function of pixel density. I do agree that a crop doesn't change the focal length of the lens, but the higher pixel density of smaller sensor cameras means that it takes a smaller movement to cause light to fall on adjacent sensing pixels.

My point is that with the smaller sensor (and hence the so called 1.6 multiplier) you have to enlarge the picture by a further 1.6x to generate the same size print (I'm assuming the aim of photography is to generate prints - could be a wrong assumption for some people).

Therefore, the shake will be more noticable on the 1.6 crop camera for the same sized print.

The rule of thumb was convenient for 35mm and that was (is) pure luck. It's about the angle the final image you view subtends. If you are holding a pair 12x8 prints in your hand, each pixel will represent a smaller angle of view when taken on a 1D Mk II or a 5D. Therefore, the stability requirement will be higher.


Paul
Click to expand...

If this holds true then all Nikons must be mounted to tripods permanently.

Their pixel densities outweigh the Canon's by far - do the math - and so they should be prone as heck to shake. But they are not...

:thinking: :suspect: :shrug: :suspect:
 
Not sure where your maths comes from. As you say "Do the math..."

Nikon D2X pixel - 181 pixel/mm
EOS 400D pixel - 175 pixel/mm

Sounds about the same to me. Yes, the 5D has a lower pixel density of 121 pixels per mm (or to put it another way, Nikon's have about 1.5 times the pixel density, about a 1.5 crop and so the rule of thumb makes most sense at about 1.5x the 35mm version)
 
grumpybadger said:
Not sure where your maths comes from. As you say "Do the math..."

Nikon D2X pixel - 181 pixel/mm
EOS 400D pixel - 175 pixel/mm

Sounds about the same to me. Yes, the 5D has a lower pixel density of 121 pixels per mm (or to put it another way, Nikon's have about 1.5 times the pixel density, about a 1.5 crop and so the rule of thumb makes most sense at about 1.5x the 35mm version)
Click to expand...

Blimey, my maths may be dated, will get back to you shortly!
 
O.K. Here goes:

D2Xs : 23.7mm x 15.7mm = 372.09 square mm on sensor

12 800 000 / 372,09 = 34 400 pixels per mm sq.




1Ds Mk II : 24.0mm x 36.0mm = 864 square mm on sensor

17 200 000 / 864 = 19 907.407 pixels per mm sq.




5D : 24.0mm x 36.0mm = 864 square mm on sensor

13 300 000 / 864 = 15 393.518 pixels per mm square



400D : 22.2mm x 14.8mm = 328.56 square mm on sensor

10 500 000 / 328.56 = 31 957.633 pixels per mm square


So while the 400D does, indeed, have nearly double the pixel density of the 5D and a 50% increase over the 1Ds MkII it is still a good bit behind the D2Xs - OK not that much...:bonk:

Still, as I understand your line of thinking the 400D (ALONG WITH THE D2Xs )should always be shot on tripod, the 5D never and a tripod is optional for the 1Ds MkII...:shrug:
 
Forget per square millimetre, it's per linear mm that matters as shake is linear in effect - which is where my numbers come from.

And I am not saying they should be shot from a tripod, I am saying they are 50% more susceptible to shake than a 5D when a print is englarged to the same size and so a shutter speed 50% faster should be used.

I.e. 1/300s for a 200mm lens. Don't see quite where the tripod view comes from unless to are trying to be confrontational? :shrug:
 
Will go and think about it
 
grumpybadger said:
I still disagree. Since shake is caused by light falling on adjacent pixels, it is actually a function of pixel density.
Click to expand...

If that was the case the 1.6 factor would be irrelevant as a 6MP 300D and a 10MP 400D would need totally different rules having a significantly different density for the same sensor size.

Pixel density is totally irrelevant to camera shake. The circle of least confusion is not going to change size because of the number of pixels and the camera will not shake any more because of them (assuming the same weight). If you double the number of pixels in a sensor then you would expect and hope that the light would fall on adjacent pixels. That is the point of having more!

When you make an 8x10" print at 300dpi it will be 2400x3000 dots. However many pixels captured the images it will still have to be reduced to those dimensions, negating the effect of pixel density.

You were correct the first time, that the crop factor should be applied because getting sharp prints is about the field of view of the image captured and not about the focal length of the lens used or the pixel density.

Anyone who doubts this just consider a Canon Ixus 850. At its widest the lens has a focal length of 4.6mm to capture a field of view equivalent to a 28mm lens on a 35mm image plane. Does anyone really think they will get sharp images hand holding for a 1/4 second?

Michael.
 
This is true when you have more pixels than you would print at, hence having to reduce resolution in the printing process. My point was more that to get the same size print from a 1.6 crop sensor, you have to enlarge the image by a further 1.6x...

It isn't a rule, it's a guideline. And yes, very high res images will be more demanding. I found the 1/f rule wasn't good enough on Velvia 50 for 18x12 prints that people were close to if you wanted them to look spot on.
 
Didn't realise this discussion would get so technical :) I do know that the 1/f rule was just a guideline and to get sharp images is down to technique more than anything else, I just wanted to see if it applied to the conversion factor of DSLRs. Reading though the thread I think the jury's still out on it.
 
Are we now also talking about correction of perspective of 1,6 vs full frame?

If we want two identical images from 1,6 crop and full frame cameras the maths obviously become just a wee bit more interesting:shrug: :thinking:

Hardly worth it methinks:bonk:
 
It is surely all up to final enlargement as to how much shake shows. Camera shake (like the poor) is always with us. All that simply happens with a faster shutter speed is not that the camera shakes any less but that less of this movement has time to register before the shutter closes. It then depends on the resolution of the human eye as to whether you enlarge the final image at any shutter speed enough so the human eye can detect this movement blur.

The same applies with depth of field, only one plane is in focus in any picture but if the circle of confusion of the other parts is below the threshold the eye can detect all these parts look equally sharp. Enlarge the image more and more and the DOF shrinks as the size of the circle of confusion in more and more of the image increases above the eyes resolution threshold.

Perhaps we should be considering the lenses angle of view when mounted on the smaller APS sized sensor as the lens then behaves like a 50% longer lens on 35mm in Nikon's case and slightly longer in Canon's. The question that needs answering is if it is the actual focal length of the lens that determines the degree of shake, or the effective focal length compared to the old 35mm shake formula?

If the former then you stick to the old 35mm formula, but if the latter then in Nikon's case you multiply it by 1.5 and 1.6 for Canon? Perhaps it needs somebody to do equivalent shake tests on both 35mm and the APS sensor with the same full frame 35mm lens to resolve the matter once and for all?

DaveW
 
sod it, just shoot everything at 1/2000 :woot: I don't really care about the theory :D
 
Have we got any anorak smileys?




zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz
 
You must log in or register to reply here.
Back
Top