Using teleconverters

[jjj]

Are you sure?
The physical aperture is the same, I agree. But the 'f stop' is the ratio of physical aperture size to focal length, so if you add a teleconverter to go from 400mm to 800mm, the physical aperture is constant so the f value changes accordingly.

Indeed.
This thread is beginning to seem like the one where Ray insisted that perspective was changed by lenses.

[BJL]

Regardless of whether the movement is due to subject movement or camera shake, I would think that in order to get the maximum resolution advantage when using a converter, one should always increase the shutter speed beyond what one would use shooting the same scene without converter, when the camera is hand-held.
I'm not sure why you are comparing in your above comment, a 200mm lens at 50metres and a 400mm lens at 100 metres. If one uses a 200mm lens with 2x converter it is presumably to get an enlarged image (greater pixel count) from the same distance.

It depends on why you use the TC; note that I was trying to stake out the range of possibilities from worst-case to best-case.

If the goal is indeed to half the angular FOV, and subject motion is a factor, then a doubling of shutter speed is indicated.  But if the goal is to be able to shoot the same subject (say filling the frame with a bird, or dangerous animal) while not having to get as close, then my 200mm from 50 mm vs 400mm from 100m comparison applies: when you fill the frame with the same subject, focal length is not directly a factor in the shutter speed needed to freeze subject motion.  And if the subject is not moving significantly and the camera is suitably stabilized (tripod or whatever), then the "1/f" guidelines about adjusting shutter speed in proportion to focal length are irrelevant.

P. S. As to which comparisons are relevant: much of what you are criticizing about TCs is the effect of the combination of a longer focal length with a higher minimum f-stop, such as using a 2x TC to change from 200/2.8 to 400/5.6.  That is not specifically an issue with TC's; it is an issue with that combination of focal length and minimum f-stop, which happens equally with a "native" 400/5.6 lens as with a 200/2.8 lens + 2X TC.  And the likely reason for using 400/5.6 in either form is that:
a) A shorter, brighter lens like 200/2.8 or 300/4 has the wrong field of view for the subject, so that using it without a TC would require a substantial crop, and
b) a brighter lens of the desired FOV, like a 400/2.8 or 400/4, would be far heavier and more expensive.

So in a lot of cases, the most relevant alternative to using a TC is not a lens of greater focal length and of equally low minimum f-stop; it is either cropping or accepting the higher minimum f-stop in order to get the arrow FOV desired.

Since many photographers (those lacking infinite budgets for gear and sherpas) use long telephoto lenses f/5.6 and slower, I do not see a fatal flaw in achieving a long focal length at about f/5.6 by the relatively cheap and light addition of a TC to an existing f/2.8 or f/4 lens.

[jjj]

Ray - also of note.
If you shoot with a 200mm and crop to a 400mm FOV instead of using a TC, then you also need to up the minimum shutter speed just as if you used a teleconverter.

[BJL]

It's been explained over and over again in this thread that a 300mm lens at F5.6 will have the same DoF as that same lens used at the same aperture with a converter.

That has been _claimed_ over and over again, but it has also been explained that this is not true if you make the traditional comparisons of images displayed at equal size and viewed from equal distance.

The effect of a 2x converter is to unavoidably drop the aperture by 2 stops, ie. F5.6 becomes F11, but the aperture of the 300mm lens is still physically F5.6 despite the reading being F11.

Ray, if you are ging to enter vigorously into technical debates, you should learn to use technical terms correctly.  A value like f/5.6 or f/11 is not the "aperture", it is the aperture _ratio_; the ratio of focal length to entrance pupil diameter, a.k.a effective aperture diameter.
Adding  a TC does not change the effective aperture diameter, but it really does change the aperture ratio, so that 300mm f/5.6 lens + 2x TC is for _all_ purposes a 600mm, f/11, with no sense of it really being still f/5.6.

Details, for the 300/5.6 example:

A 300mm lens of minimum aperture ratio f/5.6 has a maximum effective aperture diameter of 300mm/5.6 = 54mm (or close enough).
Attaching a 2x TC doubles the focal length to 600mm but has no effect on the effective aperture diameter, which stays ay 54mm, so the the aperture ratio becomes 600/54 = 11: the combination has aperture ratio f/11.

So far, so good, but
The minimum aperture ratio is really changed to f/11;
it is the effective aperture diameter that is unchanged, at 54mm.


By the way, the design of many true telephoto lenses like a 400/5.6 is internally a lot like a shorter brighter lens with a permanently attached TC inside: this is seen by a predominantly concave (magnifying) rear lens group.  So a 400/5.6 is likely internally to be optical like a 280/4 in front with a rear 1.4x magnifying group or even a 200/2.8 with rear 2x magnifying group. Indeed this is more or less the definite of a true teapot lens: one that is physically shorter thinnest focal length due to this device: http://en.wikipedia.org/wiki/Telephoto_lens

[Ray]

Are you sure?
The physical aperture is the same, I agree. But the 'f stop' is the ratio of physical aperture size to focal length, so if you add a teleconverter to go from 400mm to 800mm, the physical aperture is constant so the f value changes accordingly.

That's what I wrote, the physical aperture of the 300mm lens remains the same. The light passes through a 300mm lens at F5.6 before it reaches the converter. Imagine a completely manual lens with no automatic read-out of F/stop, and no automatic adjustment of F/stop. Imagine you have to change the aperture by twisting a ring on the lens.

You manually set the f/stop on the 300mm lens to F/5.6. You add a 2x converter. The 300mm lens still shows a reading of F5.6 on the barrel. However, if you then take a shot on the assumption that the two combined lenses are F5.6, the shot will be underexposed by 2 stops. This is because the 2x converter grabs one quarter of the image, or light, that has passed through the 300mm lens at F5.6 and spreads it over the entire sensor, so that each pixel receives only one quarter of the light that it would have received without the converter.

This also, perhaps, explains more clearly why the DoF does not change. The image that has passed through the converter is an F5.6 image with the DoF of an F5.6 image (in relation to the 300mm lens). Although the automatic f/stop readout on a modern camera changes to F11 when the converter is added, to ensure correct exposure, the qualities of that F5.6 image (what's sharp, what's not sharp, what's in focus, what's out-of-focus etc), remain unchanged, ideally, if the converter is a perfect lens.

But here's the problem. There's no such thing as a perfect lens. Even the best converter available will degrade that F5.6 image to some degree. However, such degradation is offset, at least partially if not completely, by the increased pixel count of the final image.

This is the same principle that applies to all new camera models boasting a higher pixel count, such as the recently announced 50mp Canon 5DS. The immediate reaction of some people is that they don't need so many megapixels and that they are unnecessary for their purposes and just slow down all processing. I wonder if such people are aware that increasing the pixel count of a sensor, whilst maintaining the basic pixel quality of previous models, has the effect of upgrading any lens that is used with the camera, whether the lens be high quality or mediocre.

As I've mentioned before, whatever the quality of the lens used with a converter, the new lens of longer focal length that results, will be a lower quality lens, period.

[Ray]

So in a lot of cases, the most relevant alternative to using a TC is not a lens of greater focal length and of equally low minimum f-stop; it is either cropping or accepting the higher minimum f-stop in order to get the arrow FOV desired.

Since many photographers (those lacking infinite budgets for gear and sherpas) use long telephoto lenses f/5.6 and slower, I do not see a fatal flaw in achieving a long focal length at about f/5.6 by the relatively cheap and light addition of a TC to an existing f/2.8 or f/4 lens.

That is why the most important comparison is between the cropped image from the lens without converter, with the full-sensor image through the converter. If you find that the improvement in detail, under ideal conditions with camera on tripod, is quite marginal, then it's quite likely that such improvement will disappear, and might even be reversed when the camera with converter is hand-held to capture either stationary or moving subjects.

On the other hand, if you find your shots are simply more interesting when using the converter because you can see more clearly subtle changes in the subject and are better able to 'capture the moment', then that in itself justifies the use of the converter, regardless of any slight degradation in image quality such as an increase in noise.

[Ray]

Ray, if you are ging to enter vigorously into technical debates, you should learn to use technical terms correctly.  A value like f/5.6 or f/11 is not the "aperture", it is the aperture _ratio_; the ratio of focal length to entrance pupil diameter, a.k.a effective aperture diameter.
Adding  a TC does not change the effective aperture diameter, but it really does change the aperture ratio, so that 300mm f/5.6 lens + 2x TC is for _all_ purposes a 600mm, f/11, with no sense of it really being still f/5.6.

Quite right, BJL. I've corrected my earlier post in case I've caused any confusion. I've been aware for many years of course that f/stop is a relationship between aperture diameter and focal length. However, in common parlance the two terms are often used interchangeably.

I was merely trying to convey the concept that the cropped image from the 300mm lens at F5.6 remains the same in all its qualities of sharpness and DoF (provided a perfect teleconverter is used which has a 100% MTF at all relevant frequencies etc), and that the role of the teleconverter is merely to enlarge that cropped image to the full area of the sensor.

[Ray]

Indeed.
This thread is beginning to seem like the one where Ray insisted that perspective was changed by lenses.

At least I gave some specific examples which even Bart couldn't sensibly refute. My claim was that if all factors involved were kept constant, such as use of the same camera, no cropping, same print size and same viewing distance to print, then different focal lengths of lens would result in a different perspective from the position of the viewer of the print.

As I recall, your argument was, if one crops the wider-angle shot to the same FoV as the shot from the longer lens, then the perspective in both images will be the same, which I never disputed because I understand very well that cropping produces a result which is effectively equivalent to a longer focal length of lens. Any two lenses of the same 'effective' focal length will of course exhibit the same perspective, just as any two lenses of the same 'actual' focal length will.

Have you still not got it, Jeremy? 

[spidermike]

That's what I wrote, the physical aperture of the 300mm lens remains the same. The light passes through a 300mm lens at F5.6 before it reaches the converter. Imagine a completely manual lens with no automatic read-out of F/stop, and no automatic adjustment of F/stop. Imagine you have to change the aperture by twisting a ring on the lens.

You manually set the f/stop on the 300mm lens to F/5.6. You add a 2x converter. The 300mm lens still shows a reading of F5.6 on the barrel. However, if you then take a shot on the assumption that the two combined lenses are F5.6, the shot will be underexposed by 2 stops. This is because the 2x converter grabs one quarter of the image, or light, that has passed through the 300mm lens at F5.6 and spreads it over the entire sensor, so that each pixel receives only one quarter of the light that it would have received without the converter.

This also, perhaps, explains more clearly why the DoF does not change. The image that has passed through the converter is an F5.6 image with the DoF of an F5.6 image (in relation to the 300mm lens). Although the automatic f/stop readout on a modern camera changes to F11 when the converter is added, to ensure correct exposure, the qualities of that F5.6 image (what's sharp, what's not sharp, what's in focus, what's out-of-focus etc), remain unchanged, ideally, if the converter is a perfect lens.

If you print an image at two different sizes, DOF changes if you view both from the same distance - all DOF calculation tables make assumptions on image size combined with viewing distance (IIRC it is viewing distance is 2x the diagonal for a 3:2 image). The reasoning here is that DOF is a construct by photographers to talk about acceptability: there is only ever one plane truly in focus and once you start to vary from that even so much as an inch the clarity of that focus starts dropping; and the biffer you blow th eimage up the more obious it is that certain things are not in focus (just like you check sharpening at 100% or 50% not at 10%).
So what you are seeing here is, I believe two counteracting issues:

-  The tc is like taking a crop of a much larger image so the DOF narrows
-  The effective aperture drops which increases DOF

As DOF drops with image size, if you are saying that the DOF passed to the teleconverter is the same as the native 300mm lens, then the large projection created by the tc will decrease DOF compared to 300mm alone.

[Ray]

If you print an image at two different sizes, DOF changes if you view both from the same distance -

Exactly! That's why it makes sense to compare DoF only on same-size images or prints viewed from the same distance, and viewed at a distance close enough for one to be able to discern the sharpest parts in the images.

-  The tc is like taking a crop of a much larger image so the DOF narrows.
-  The effective aperture drops which increases DOF

If you crop an image, the image as a whole changes, obviously. Objects are removed and the cropped image becomes smaller as a result, but the objects within the crop do not change. What was blurred in the cropped part of the image before cropping, remains blurred after cropping, and what was in focus in the cropped part of the image before cropping, remains in focus after cropping.

What might happen to some small degree, is that the sharper parts of the cropped image are degraded slightly more by the teleconverter than the blurry or out-of-focus parts are, resulting in a slight increase in DoF, or reduction in shallowness. However, this effect is probably counteracted when the enlarged crop is recorded by the full sensor.

The point is often made when people buy a new camera with a high-resolution sensor, such as the Nikon D800, that good lenses show a more significant increase in performance than poor lenses, and that the soft edges in the image are not raised in quality as much as the sharper centre. This has the subtle effect of decreasing DoF, ie making it shallower.

[jjj]

At least I gave some specific examples which even Bart couldn't sensibly refute. My claim was that if all factors involved were kept constant, such as use of the same camera, no cropping, same print size and same viewing distance to print, then different focal lengths of lens would result in a different perspective from the position of the viewer of the print.

As I recall, your argument was, if one crops the wider-angle shot to the same FoV as the shot from the longer lens, then the perspective in both images will be the same, which I never disputed because I understand very well that cropping produces a result which is effectively equivalent to a longer focal length of lens. Any two lenses of the same 'effective' focal length will of course exhibit the same perspective, just as any two lenses of the same 'actual' focal length will.

Have you still not got it, Jeremy?  

Sigh! Lenses do not change perspective, only the position where image is taken from relative to subject does. Lenses only alter the field of view captured.
And your debating here is a repeat of the bonkers perspective thread I mentioned.

[spidermike]

Exactly! That's why it makes sense to compare DoF only on same-size images or prints viewed from the same distance, and viewed at a distance close enough for one to be able to discern the sharpest parts in the images.

And if you put a tc on the lens, or if you crop and print to the same size, you are effectively looking at a larger version of the image from the same distance. So in practical terms you are changing the parameters

If you crop an image, the image as a whole changes, obviously. Objects are removed and the cropped image becomes smaller as a result, but the objects within the crop do not change. What was blurred in the cropped part of the image before cropping, remains blurred after cropping, and what was in focus in the cropped part of the image before cropping, remains in focus after cropping.

The bit in bold I agree with...but only at the plane of focus. Anything outside tha plane of focus is de facto out of focus and this is a gradual variation.  Depth of field is all about what is acceptably sharp which means we all know it is out of focus but ask ourelves the question 'does it look sharp enough'. We are all familiar with the phenomenon that an image looks woderfully sharp on your camera LCD screen but is total pants when you load it up (whether it is a focus issue or a camera shake issue), so let us expand on that: assume focussed on one person in a crowd:
the A4 print looks like everyone three rows back is in the DOF
Print to A3 and look at it from the same distance (or crop to create an A4 print - same thing) and you realise the guy in the third row back is actually a bit blurred. Your definition of 'acceptably sharp' has not changed but blowing up the image has shown gradations in the edges that were not clear on the smaller print and things that were within the DOF have now fallen outside the DOF.

[BJL]

That is why the most important comparison is between the cropped image from the lens without converter, with the full-sensor image through the converter.

OK, so you are comparing the two options of
a) using a TC
b) cropping to achieve the same FOV as with the TC.
Then we agree: for _that_ comparison, DOF is the same; TC's and cropping are very similar -- except for the increased resolution achieved by using a TC, so long as the lens without TC delivers more resolution that the sensor can record (and the TC doesn't totally suck).

However, due to the higher resolution of the subject with the TC compared to the crop (or at least, increased pixel count), advocates of comparing DOF at the maximum magnification allowed by the resolution of the image, such as displaying at equal PPI and then viewing from equal distance (which result is calculated by using a DOF calculator with the allowable CoC set in proportion to pixel size) would argue that the version with TC has a higher pixel count, and so can be viewed closer or displayed at a larger size, and then will have roughly half the DOF.  Not me though!

[telyt]

No, I said more than that.  

I said I would like to see comparisons of the lens with and without converter. Your picture of the bird is very nice and certainly acceptably sharp at the size presented, but gives me no idea of the relative quality of images with and without converter in circumstance when one cannot avoid the f/stop disadvantage with the converter.

I don't test equipment, particularly if only for the sake of internet forum doubters.  I make photographs.  Having no experience with the equipment in question you have no idea how well this equipment works.  The Leica 280mm f/4 APO is nearly diffraction-limited at full aperture; the CaNikon equivalent 300mm f/4 lenses are a cruel joke in comparison.  In my real-world use of this lens along with the Leica 1.4x APO-Extender-R over the last ten years the image quality loss from using the Leica 1.4x TC is a very slight loss of contrast which in most cases is negligible (and I'm very picky about image quality).  I do see some image quality loss using the Leica 2x extender but the loss is small and the combination is much more than mediocre for large prints.

… here's the problem. There's no such thing as a perfect lens.

Since you have no experience with the Leica 280mm f/4 APO I can understand how you'd reach this conclusion.  It's not perfect but it's very nearly so.

[kers]

...the CaNikon equivalent 300mm f/4 lenses are a cruel joke in comparison.... 

Well the new 300 f4 PF nikkor is really good and has a very usefull VR... nothing cruel or funny about it...just good.

[dwswager]

Ray - also of note.
If you shoot with a 200mm and crop to a 400mm FOV instead of using a TC, then you also need to up the minimum shutter speed just as if you used a teleconverter.

And close down the aperture to get the equivalent DOF because you will be enlarging more.  Hence, the smaller CoC you would use for the greater enlargement, if calculating DOF, would drive a higher f number.

[telyt]

Well the new 300 f4 PF nikkor is really good and has a very usefull VR... nothing cruel or funny about it...just good.

As long as you use Nikon's PF-compensating software, which if the examples mean anything, is inadequate.

EDIT: The PF lens may well be the equal of the Leica 280mm f/4 but that's not why I brought up the Leica lens.  The issue I was bringing up is that Ray hasn't used the 280/4, nor the 300mm PF so it seems unlikely he's used a really good lens with a really good TC so IMHO his blanket statement that a lens plus TC will produce mediocre image quality is likely based on experience with lenses that aren't as good.

[kers]

As long as you use Nikon's PF-compensating software, which if the examples mean anything, is inadequate.

EDIT: The PF lens may well be the equal of the Leica 280mm f/4 but that's not why I brought up the Leica lens.  The issue I was bringing up is that Ray hasn't used the 280/4, nor the 300mm PF so it seems unlikely he's used a really good lens with a really good TC so IMHO his blanket statement that a lens plus TC will produce mediocre image quality is likely based on experience with lenses that aren't as good.

OK i understand...
The PF certainly has some influence in how light spots are rendered. (but much less than a mirror lens and different of course) On the whole i find it not that important and what remains is an excellent 755 gram 300mm lens.
About the usefulness of converters.. i think the success is based upon two things- the original lens must have a very sharp centerpart and the converter has to be made for that lens.

[dwswager]

.
About the usefulness of converters.. i think the success is based upon two things- the original lens must have a very sharp centerpart and the converter has to be made for that lens.

+1

I suspect this is another constraint on manufacturer lens design.  If you look through the Nikon line, there have been numerous different converters produced over the years and each only recommended with specific lenses.  I have the TC-14EIII and it performs well on the 70-200mm f/2.8G VR II on which I use it.  Glad to see it also works on the new 300mm f/4G, but is not compatible with the older 300mm f/4D!

[BJL]

About the usefulness of converters.. i think the success is based upon two things- the original lens must have a very sharp centerpart and the converter has to be made for that lens.

Optics experts might want to jump in and correct me, but from the little I have read, matching a teleconverter to lens just depends on compatibility with the exit pupil height of the lens.  This is, roughly, the distance at which the aperture diaphragm appears to be when seen from the focal plane at the back of the lens. In a telephoto lens, this is typical higher than reality; for example, in a perfectly telecentric design, the exit pupil height can be infinity!. So one TC can work well with several lens, so long as they have (roughly) the same exit pupil height.

Also, from what I have seen, Nikon has typically offered several 1.4x TC's, each recommended for one or several lenses, presumably grouped by having similar exit pupil height. Canon instead typically offers just one 1.4x and one 2x; either Canon does not care so much about TC performance, or it designs all its fast telephoto lenses with roughly the same exit pupil height -- maybe they all are very near-telecentric with a very high exit pupil, so that TC's designed for an infinite exit pupil height work well enough.