Luminous Landscape Forum
Equipment & Techniques => Digital Cameras & Shooting Techniques => Topic started by: Guillermo Luijk on January 02, 2022, 05:19:31 pm
-
In a discussion on another forum, a somewhat obnoxious forum member stated that a teleconverter never adds anything over cropping an image. I made a quick test to prove the opposite: Canon FDn 85mm f/11 with a KENKO 2X CFE TELEPLUS MC7 teleconverter:
http://guillermoluijk.com/misc/dupli.jpg
Of course it doesn't mean a teleconverter is always a good idea, but when the lens used has a very high resolution over the sensor, the teleconverter can help in grasping more real detail and avoiding aliasing through the inherent oversampling achieved.
I want to repeat the test to accurately measure the amount of light lost because of the transmission path in the converter.
Regards
-
The visual difference is massive.
-
What camera was used. I've cropped many images and I get very sharp results. Light loss with the teleconverter needs to be taken into account where you might have to increase the ISO by a stop to compensate.
-
What camera was used. I've cropped many images and I get very sharp results. Light loss with the teleconverter needs to be taken into account where you might have to increase the ISO by a stop to compensate.
It's a 24Mpx Sony A7 II. The point is not to demonstrate if a crop can or cannot be sharp, but if the teleconverter can provide extra detail and antialiasing protection over the crop.
The more Mpx the sensor has and the lower resolution of the optics, there will be less advantage in the optically expanded image vs crop. In my test I used a reasonally low resolution sensor and a high resolution lens since it was addressed to show the teleconverter can be useful.
I also want to check for other two matters where the teleconverter will loose:
- Noise: using the teleconverter we capture aprox. the same total amount of light than in the crop, but the photodetectors capture 1/4 of the light using the teleconverter. This should translate as a 1 stop loss in SNR and DR once both the teleconverter and crop images are compared at the same output size.
- Sharpness, CA and distortion in the borders: did the teleconverter introduce harmful optical effects in the peripheral areas which didn't exist in the crop?
Regards
-
I would not count light loss against the teleconverter, and it should not have worse SNR.
Given the same exposure time and lens diameter, both crop and converter capture the same number of photons from the relevant section of your scene. A 2x converter provides 1/4 as many photons per pixel, but they are all useful photons, spread out across the entire sensor. The crop wastes 3/4 of the incoming photons, concentrating the useful photons on 1/4 as many pixels. After enlargement (for example upsampling x4 to provide the same pixel count), the crop provides the same number of photons per pixel as the converter.
If sensor noise is not negligible compared to shot noise, then the crop will have some advantage, since it reads out fewer sensor pixels. Except in really low light situations, this should be a small effect. Considering shot noise alone, the converter and the crop have identical SNR.
The converter can actually provide higher SNR, if you decide to increase the exposure, providing more photons per pixel, without exceeding the pixel well depth. This could be a significant advantage for the converter.
-
Considering shot noise alone, the converter and the crop have identical SNR.
Agreed on the shot noise part, but why sould shot noise be the only relevant? in fact it's read noise the DR limiting factor, so in the deep shadows the crop will have a better SNR.
SNR doesn't linearly correlate with the total amount of light collected, it also depends on how many photocaptors gathered that light.
In a 2X Teleconverter each of all photodetectors of the sensor captures 1/4 the amount of light that is captured by each of the photodetectors of the crop.
Since with the teleconverter we have 4 times more photodetectors the total amount collected is the same on both scenarios.
When we dive into the deep shadows (low exposure, read noise), the SNR of the teleconverter photodetectors becomes 4 times worse. Once reescaled down to fit the size of the crop, the image from the teleconverter only improves SNR by a factor of 2 (explanation (https://homes.psd.uchicago.edu/~ejmartin/pix/20d/tests/noise/noise-p3.html#pixelsize)). This means that once both images are equal in size, the crop will still have a 1 stop advantage (SNR will be double) than the teleconverter in read noise dominant areas (deep shadows).
---
I measured the light attenuation when using the teleconverter vs crop, using the same aperture/shutter. Just for the magnification the loss should be around 2EV, but I realized my teleconverter is closer to a X2.1 than a X2 teleconverter, so light attenuation should be higher (log2(2,1^2)=2,14EV). Comparing RAW values, 2,14EV seems to be exactly the attenuation for the G channel, so it seems light loss in the teleconverter is negligible. The other two channels have slightly higher attenuations which can be the prove of some colouring in the optics:
http://guillermoluijk.com/misc/dupliloss.png
The distributions are so gaussian like (photon noise) that makes me think any additional vignetting introduced by the teleconverter is negligible. However I checkd for the spatial distribution of the light loss (RAW G channel) and was surprised to find a 'ring' near the borders with lower light attenuation than in the centre itself:
http://guillermoluijk.com/misc/duplivignetting.png
The differences are nearly negligible (<0.1EV), but I sill wonder what can be the reason for that. Maybe some form of geometrical distortion that compresses more light per surface unit in that ring than in the center.
The converter can actually provide higher SNR, if you decide to increase the exposure, providing more photons per pixel, without exceeding the pixel well depth. This could be a significant advantage for the converter.
Totally agree. In a situation where achieving exposure is not an issue, we could increase exposure vs the crop in 2EV.
Regards
-
"
The distributions are so gaussian like (photon noise) that makes me think any additional vignetting introduced by the teleconverter is negligible. However I checkd for the spatial distribution of the light loss (RAW G channel) and was surprised to find a 'ring' near the borders with lower light attenuation than in the centre itself:
http://guillermoluijk.com/misc/duplivignetting.png
The differences are nearly negligible (<0.1EV), but I sill wonder what can be the reason for that. Maybe some form of geometrical distortion that compresses more light per surface unit in that ring than in the center.
"
maybe reflection from the barrel ? or maybe an optical path to minimize apparent vignetting ?
-
Teleconverters are optimized for only some lenses ;
i used a 2x teleconverter on my nikon 70-200 and found the result bad; almost as a crop.
With my 300PF lens however it works much better and makes it a very usable 600mm lens ( at 1100 gr)
-
We also must include the effects a TC has on the AF abilities of the camera. Less light for the AF to work with usually results in poorer AF abilities. In fact, there are cameras that revert away from a cross point AF to a single vertical line when the aperture closes too much.
Rather than just looking at images zoomed up...we need to look at the entire package to see the effects teleconverters have on the final output.
-
We also must include the effects a TC has on the AF abilities
Of course I agree with that, by I didn't intend to do an overall "are extenders a good idea?" test. I was interested in how much extra detail and extender can provide vs a crop, as the thread title suggests. I have no AF lenses so I couldn't check for AF even if I wanted to.
Regards
-
Of course I agree with that, by I didn't intend to do an overall "are extenders a good idea?" test. I was interested in how much extra detail and extender can provide vs a crop, as the thread title suggests. I have no AF lenses so I couldn't check for AF even if I wanted to.
Regards
Do you take into account 2 stops of light loss when comparing the images? IE: keep the shutter and aperture the same and adjust the ISO to get equivalent exposure?
-
Do you take into account 2 stops of light loss when comparing the images? IE: keep the shutter and aperture the same and adjust the ISO to get equivalent exposure?
I used the same aperture, shutter and ISO100 on both shots because I wanted to accurately measure the light loss. Perhaps you missed that a couple of posts above:
http://guillermoluijk.com/misc/dupliloss.png
http://guillermoluijk.com/misc/duplivignetting.png
The loss is a bit higher than 2 stops: the median loss is 2,14EV for G channel as measured on RAW data, but this can be also understood because my extender is closer to a X2.1 extender than to a X2.
Regards
-
the lightloss does not concern me; i use the 2x converter to get more detail and that works if the combination of the converter and the lens are made for each other.
if you need more light: buy a dedicated 600mm;
if you have enough light and want to have a portable 600mm ; this combination is OK- even over the whole 24x 36mm image at f13.
example nikon 300 PF + 2x converter
left is 300mm blow up
right is 2x converter 100%
-
I added a final option to the test: brute force sensor oversampling, i.e. doing a crop on a high pixel count sensor (in this case I simulated a 64Mpx FF sensor using my 16Mpx Olympus PEN):
http://guillermoluijk.com/misc/duplivssobremuestreo.jpg
While still a native focal length of 100mm on a 24Mpx sensor is the clear winner, I can see a bit more of detail and better contrast when cropping the 64Mpx sensor (right) vs using the 2X extender on the 24Mpx sensor (left):
http://guillermoluijk.com/misc/duplivssobremuestreo2.jpg
Being the complete inexistence of aliasing the only advantage of using the 2X extender here, which in addition generates heavier files. So high pixel count wins over my extender.
Regards
-
The converter can actually provide higher SNR, if you decide to increase the exposure, providing more photons per pixel, without exceeding the pixel well depth. This could be a significant advantage for the converter.
The last test I wanted to do is to compare DR (deep shadows noise) of:
1. ETTR cropped shot
2. 2X shot with the same aperture/shutter as in 1
3. ETTR 2X shot
According to statistics DR should be (best to worst): 3 -> 1 -> 2
I prepared this high contrast scene (background wall is the highlights reference):
http://guillermoluijk.com/misc/dupliglare1.jpg
I shot it without the 2X and perfect ETTR:
http://guillermoluijk.com/misc/dupliglaresin.gif
(RAW histogram)
100% ETTR crop without the 2X:
http://guillermoluijk.com/misc/dupliglare2.jpg
Next shots 2 (2X with same exposure parameters as in 1) and 3 (ETTR 2X), both rescaled 50% down to fairly compare noise:
http://guillermoluijk.com/misc/dupliglare3.jpg
Expected noise behaviour, but a terrible veiling glare in the deep shadows caused by the extender ruins the DR. The three images have been very carefully linearly processed so that only exposure has been corrected to match the three.
Regards
-
Probably beating a dead horse, but the teleconverter will always yeild far more detail than cropping into a frame.
-
An optical magnification will always be better than a digital one, but does it matter?
I shoot leather goods for a customer and we do several hundred shots in a day. The camera never moves.
Regardless of whether it a suitcase or a leather wallet they all get shot from the same point with the same lens.
The big cases take up nearly the whole frame 8000 pixels wide. The wallets maybe 2000 pixels wide
So I crop the big cases 8000 pixels wide and the wallets 2400 pixels wide and they all get exported at 2400 pixels wide.
On the web they are the same quality.
If you have enough good megapixels on a good camera it does not matter in that case at least.
-
Probably beating a dead horse, but the teleconverter will always yeild far more detail than cropping into a frame.
Exactly. You do lose light, of course, but with modern sensors and denoise software that is rarely an issue.
-
Probably beating a dead horse...
That's a purpose for which online forums seem to be well suited and often used.