Do Teleconverters Halve Resolution of Lens?

<p><em>Disclaimer: This is more of an academic question for people with optics knowledge with, of course, practical implications :)</em></p>

<p>I read <a href="../medium-format-photography-forum/001rY7">here</a> in a post by Huw Evans that a 2x teleconverter, at best, halves the resolution of your lens. The exact quote by Huw is:</p>

<blockquote>

<p>Essentially the converter just magnifies the central portion of the field. Even a theoretically perfect converter will thus lose you at least half your 'resolution'. In MTF terms the 20 lpmm curves will become the upper limit for 10 lpmm curves</p>

</blockquote>

<p>Is this true? I understand that magnifying a portion of your lens can enhance optical defects of the lens, but a 2x magnifier cutting the resolution in half sounds like you're magnifying some digital sensor grid 2x & then imaging that... & I'm dubious about a lens/light rays working like that. </p>

<p>But then again if you're using two components in an optical system having the same resolving power, the resulting resolving power is half of either component, since:</p>

<p>1/Rtotal = 1/R1 + 1/R2 = 1/R1 + 1/R1 = 2/R1<br>

Rtotal = R1/2</p>

<p>Of course, that's why beginning with a lens of higher resolving power gives you a final higher resolution with the teleconverter in place.</p>

<p>But back to the original question: is Huw's statement correct?</p>

<p>Thanks in advance!<br>

Rishi</p>

<p>The more correct statement is that a 2x teleconverter will halve the aerial image resolution of a theoretically perfect lens.</p>

<p>Since the maximum theoretical resolution of a lens can be approximated by 1600/(f number) lp/mm and since a 2xTC will double the f number of any lens it is attached to, it follows that it will also halve the maximum possible theoretical resolution in the aerial image plane.</p>

<p>Of course that does not mean that the resolution as measure on film or a sensor will drop by a factor of two.</p>

<p>The reciprocal formula for adding resolutions is a very rough empirical rule for film based systems. It does not necessarily hold for digital, though again you can probably use it as a very rough "guestimate" function. To do it right you have to multiply the MTF of the lens and sensor/film and then define what you mean by "resolution".</p>

<p>Take a look at <a href="http://www.bobatkins.com/photography/technical/multipliers.html">http://www.bobatkins.com/photography/technical/multipliers.html</a> for more details</p>

<p> </p>

<p>A very good 1.4 teleconverter on a matched lens has no real degradation of image quality. 1.7 & 2 will push the limits on many lenses.<br>

Pixel Peepers will always challenge this.<br>

On paper, a Nikkor 1.8 is a sharper lens than a 1.4 but 1.4 users seem to get by.</p>

 

<p>Bob's comment brings up some fascinating possibilities.</p>

<p>Imagine you have a 100mm f4 lens. The theoretical max resolution is 1600/4 = 400lpm. On a theoretically perfect 2x converter, the resolution would drop to 1600/8 = 200lpm.</p>

<p>Now, imagine that the lens is very imperfect and spherical aberration has knocked the actual resolution down to 1/10 of it's theoretical limits, from 400lpm to just 40lmp.</p>

<p>And imagine that the converter also has oodles of spherical aberration, but the converter's SA just happens to be in the opposite direction as the lens's SA. So the converter cancels out some of the lens's SA, and instead of giving you 1/10 the theoretically calculated performance, the combination actually delivers 1/4 of its theoretical performance. So, you get 1/4 of 200lpm, and you end up with 50lpm, actually higher than the lens without the converter.</p>

<p>This is not just a "thought game". I don't know any 2x that matches a lens that well, but I know a documented case for a 1.4x. The Nikon 300mm f2.0 has a measurably higher resolution with the TC-14C 1.4x converter than it does alone. So, if there are 1.4x converters in the world that really increase the resolution beyond what you get without the converter, it's no stretch of the imagination to imagine that there are lots of 1.4x converter and lens combinations that don't have more than Mr. Evan's 0.7x the resolution of the lens without the converter. I'm betting that there are lots of 2x converter and lens combinations that have a lot more than his 0.5x resolution, too.</p>

<p>In lens tests that I have done over the years with different 300/2.8 and 400/2.8 lenses, and a few others, I thought I could see slight improvement when adding a matching 1.4x converter. I was always shocked by this and often wondered if the 1 stop lowering of the effective maximum aperture of the combo was providing this increase in resolution.</p>

<p>For my own peace of mind when I add the 1.4x I try to stop the lens down one more stop to make up for any degradation. I have noticed a 1 to 2 stop difference in sharpness with a 2x though. I am referring to the fact that a lens, like my Nikon 400/2.8, needs to be set at f5.6 with a 2x converter attached (effectively f11), to achieve the sharpness with the lens itself set at f2.8 </p>

<p>You may as well argue that each optical element (and air surface) negatively affect - halve, whatever... - the resolution of a lens because a TC is nothing but just another optical group in a rather complex optical system. Adding complexity to a system does not necessarily lead to loss of resolution...</p>

<p><em> a TC is nothing but just another optical group</em></p>

<p>Actually it's <strong>five </strong>additional groups in a typical case...</p>

<p><em>> Actually it's <strong>five </strong>additional groups in a typical case...</em><br>

FWIW: Although there are single-element optical groups, most optical groups consist of more than one optical element. Say, the new Canon Extender EF 1.4x III has 7 elements in 3 groups and the new 2x TC - 9 elements in 5 groups. Kenko Teleplus PRO 300 DGX 1.4x has 5 elements in 4 groups (hence 3 single-element groups), etc. So it does vary.</p>

<p>Groups and elements are irrelevant. It's the fact fact that you lose two stops which leads to halving the resolution (due to diffraction) in the case of a perfect lens coupled to a perfect TC</p>

<p>In the case of an imperfect lens a 2xTC may indeed increase resolution if it exactly compensates for aberrations in the original lens (" a matched TC"), though the result will still have half the resolution (or less) the original lens would have had if it had been perfect. Though this could, in principle, happen, it rarely if ever does.</p>

<p>So if you put a 1.4x TC in the same housing with a f/2.8 lens and get a f/4 lens of the same focal length, you'd somehow magically halve its resolution..? Perhaps, but not necessarily so. The "perhaps" part refers to the fact that a TC (in a non-dedicated, non-matched setup) is a third leg which extends the lens' focal length from the "wrong end" and that's why modular lenses (say Leitz Apo Telyt 280/400/560) have the optical groups rearranged comparing to the original, single focus, designs of Telyts of the same focal lengths.</p>

<p>As an aside, I can't wait to see the new EF 200-400 zoom with a - presumably - well matched internal TC.</p>