Wednesday, 22 July 2015

Do lens extenders ruin image quality? Not always!

Should wildlife photographers use 1.4x and 2x lens converters to extend the reach of their lenses and get better pictures, or do they just ruin image quality? 

I have been wanting to write a Blog to address this question for some considerable time but have been a bit unsure about how to approach the topic. It is a very important issue for nature photographers and I really want to get it right. So here goes.... and I am going to take my time to explain things properly, so please settle yourself comfortably. The outcome may surprise you.

Here for a start is a shot of a Grey Heron taken with a 2x teleconverter attached to my lens. I think it is quite good, especially looking at the original! But is it a fluke? What can one generally expect?




I have found a lot of confusion amongst wildlife photographers as to whether adding a teleconverter to their lens would actually improve or worsen the final image quality. The reason that people worry is that they have probably read in numerous internet articles that "teleconverters degrade image quality".

To be clear, I am talking about those additional lens attachments, sometimes called converters, teleconverters or extenders, that will increase the focal length of your lens by a factor between 1.4 and 2. I will try to call them "extenders" from now on in this Blog but, if I slip up, you will know what I mean. Canon have two models at 1.4x and 2x, whilst I believe Nikon have one at 1.7x too.

To be fair, it does seem too good to be true doesn't it? That you can spend a few hundred pounds on an extender and transform your 400mm lens (say) to either a 560mm lens using a 1.4x converter, or to a whopping 800mm using a 2x extender. What about the people who paid thousands more to get lenses of that focal length? Surely you cannot expect very much from these cheap add-ons. Isn't it cheating?

Given all that, we are really not surprised to find that many authoritative websites show photos of lens test charts, and other objects, that indicate poorer images are obtained when using an extender. But wait a minute! What are they actually saying and what does it mean for us as wildlife photographers in particular?

If we take the following excellent lens comparison site for example:

http://www.the-digital-picture.com/Reviews/ISO-12233-Sample-Crops.aspx?Lens=962&Camera=963&Sample=0&FLI=0&API=0&LensComp=962&CameraComp=963&SampleComp=0&FLIComp=2&APIComp=2


It is a long URL because it shows a specific comparison I set up to compare the performance of my lens with and without a 2x extender. It is quite safe to click on the link. It will show you a test chart (like the one below) produced in their tests on my lens and camera. I am interested in the Canon 7D MkII camera with a Canon EF 400mm f4 DO II lens, plus the 1.4x and 2x extenders of course. They have studied these combinations.



Rolling your mouse over the chart (on the website, not here) allows you to see the same portion of the chart photographed with the addition of a 2x extender. It is really great. If you click on the link and roll your mouse over the chart you will see that  degradation in image quality that we expected! Go on, do it.

So, end of story then? An extender will degrade your image. Best save up for a 800mm lens. No! Lets start from the basics.

The important thing to note when using The Digital Picture website is that the image size does not change even when you add a 2x extender. Close reading of the information provided on the site reveals why this is so. 


To obtain the test results I am interested in they firstly position my camera and bare lens in front of a large, well illuminated, test chart. (Click on the ? mark at the top of their page to read all the precautions they take.) They place the camera at such a distance that it will exactly frame the portion of the test chart they wish to use. They take numerous photos with it like that and choose the sharpest one to use in the published results.

Then, as far as I can understand, they add the 2x extender and reposition the camera so that the new lens combination frames exactly the same area of the test chart as before. That is, they move the camera twice as far away as it was for the bare lens test. That is how the test images with and without the extender come out to be the same size. It is significant to note that both images therefore have the same number of pixels in these tests. The issue of number of pixels will be discussed again a bit later, it is important.

So the test results on that site compare the bare-lens image with the extended-lens image obtained at a greater distance.

Think about this for a moment. You are a bird photographer desperate to get a more magnified image of a distant eagle sitting on a dead tree a hundred metres away. If you move it will fly away so the only options you have are:

1. Take a photo with the bare lens and hope to resize it up when you get home to your computer. (You know that your bare lens performs well in the tests.)

2. Add an extender to your lens and obtain an adequately sized image immediately. (But, because of what you have read, you are worried about the resulting image quality.)

What you cannot do is crawl 1.4 or 2 times closer and take a shot with your bare lens from that more advantageous position, but that is the only option which the lens comparison results address. In other words, the comparison is irrelevant to most wildlife photographers.

I would absolutely expect that a better shot would be obtained with the bare lens moved closer compared with a distant shot with a 2x extender. But that is not the issue for us. What we need to know is which of options 1 or 2 above would give us the better image, i.e.:-

To get the enlarged image we seek, do we take a shot with the bare lens and resize it up on the computer, or do we slap on the extender and avoid the need to resize? Which gives the better image? This option is not covered by the tests.

This could mean that the image degradation observed in the test is not going to be a problem for us, let's see.


Using my Canon 7D MkII and my Canon EF 400mm f4 DO II I have taken several photos of a plastic doll. I have chosen the sharpest of these and cropped out the face area. The image is sized 312 x 208 pixels and is a direct cut out from the original, you can see it below. No sharpening of any sort has been applied either in the camera or in processing to any of the following images. This starting image has to be a small area because I want to resize it up later without having this Blog site impose any resizing of its own on the image.


312 x 208 pixel crop from image taken with bare 400mm lens

I then added a Canon 1.4x III extender and took another shot from the same camera position. From that larger image I have cropped out the same portion of the picture, or at least as best I could, the image has shifted slightly and I am a few pixels adrift in the size but it doesn't matter.

448 x 299 pixel crop from image taken with 400mm lens plus 1.4x III extender

Since the image is bigger, I have obviously had to crop a bigger area to get the same portion of the image,i.e. the dolls face. In fact, the linear pixel count has to be 1.4 times greater in this crop than in the original to get the same part of the doll. This means that the second image has 1.4 x 1.4 = 2 times as many pixels in it as the first. This has to be a big plus in favour of the extender.

What I am now going to do is use the Photoshop Bicubic algorithm to resize the original image taken with the bare lens to produce one which is the same size as that taken directly with the 1.4x extender. Here it is:

448 x 299 pixel image obtained by up-sizing the original image obtained with bare 400mm lens

This is the comparison that interests us as wildlife photographers stuck at a fixed distance from a bird or mammal. You will hopefully agree that the up-sized image is inferior to the image obtained directly with the extender. It is not a bad image, but it is less clear. This has to be due to the fact that the original image from which it is derived contains only half the number of pixels compared with the image obtained directly with the extender. 

So we immediately conclude that, for my camera and lens, I will get better images of a distant bird by using the 1.4x extender than without it. Up-sizing a bare-lens image on the computer will not produce the same result. This also means that, whatever imperfections the 1.4x extender may be introducing, they are not significant when weighed against the fact that the extender is allowing twice as many pixels to be dedicated to reproducing image detail.

We can, of course, go further and investigate what happens if we use the 2x extender. These 2x extenders have a particularly bad name in terms of the image quality they produce. One famous British wildlife photographer recommends that the best thing to do with them is beat them to pieces with a sledge hammer - really, can that be true?

Here is an image taken from the same position with a Canon 2x III extender attached to my lens. To get the same image portion as cropped from the bare-lens image I have now had to crop out an area which is twice as big in linear pixel count, i.e. the image is now 624 x 416 pixels compared with the original 312 x 208. There are thus four times as many pixels available to pick up detail on the dolls face.

 624 x 416 pixel crop from image taken with 400mm lens plus 2x III extender


As before,  I have now up-sized the original bare-lens image to produce one the same size as that obtained directly with the 2x converter

624 x 416 pixel image obtained by up-sizing the original image obtained with bare 400mm lens

The up-sized image is clearly inferior and once again the extender wins out.

I must re-stress however that I can only definitely say this for my lens combinations. It is perfectly possible that a 2x converter attached to a lens of poorer quality than the 400mm f4 DO II could introduce degradations that would swamp the advantage brought by the increased number of pixels. However, when you consider that there are four times as many pixels available to describe detail in the image, these degradations would have to be pretty severe! Having said that, I must admit that my previous lens, the original version of the Canon 400mm f4 DO, did not perform well with a 2x extender. It was doubtful that it was any better than up-sizing the bare-lens image - so it can easily happen. You need to check your own combination. I believe that the original Canon 100-400mm was a disappointment in this regard too.

It has often been said elsewhere that you need to have a very sharp lens to start with. Anything else is likely to lead to disappointment when using extenders, especially the 2x.

But the 2x extender may not be in the clear yet, even with me! If you have a 1.4x extender the sensible thing to do might be to up-size images taken with that by a further factor of just 1.4 to produce images the same size as those obtained directly with a 2x converter. This would obviously be better than up-sizing bare-lens images by a factor of 2 and might avoid the need to buy a 2x.

Here is an image that has undergone that treatment:


624 x 416 pixel image obtained by up-sizing the image obtained with 400mm lens plus 1.4x extender

and here is the image from before obtained directly with the 2x extender shown again for easy comparison with the image above:


624 x 416 pixel crop from image taken with 400mm lens plus 2x III extender (shown again)

It is a closer run thing but I think that the image immediately above obtained directly with the 2x extender still has the edge over the previous one obtained by up-sizing that obtained with the 1.4x extender. What do you think?

I have already said that none of the images used above was sharpened, but I thought I would just show you the photo taken directly with the 2x converter after it had been slightly sharpened in my normal processing procedure:


Final processed image obtained with lens and 2x extender, just for fun.

You can see that applying some sharpening during processing can make an enormous difference in image quality. There is not much wrong with the image above and gives me further confidence that my 2x converter is a valuable accessory.

Summary

1. Published lens comparison tests with and without extenders can be misleading as they do not always reflect the situation that the wildlife photographer finds him/herself in. We do not normally have the opportunity to move closer to the subject to get a large image using just our high-quality bare lens, and would like to know the effect of adding an extender compared with just up-sizing the bare-lens image, both images being taken from the same position.

2. Adding an extender always means that, photographed from the same position, a subject will benefit from an increased number of pixels to describe its detail. If, at the same time, the extender introduces degradations these would have to be quite severe to negate the pixel advantage. Normally, use of an extender results in higher image quality. Even so, buying a more expensive, longer focal length lens might give you the best image quality of all!

3. If you visit a different section of the lens comparison website quoted above, you will see other images which precisely address the issues I have discussed above. Scroll down to the final image on the page. It too shows that using an extender always beats up-sizing a smaller image!

http://www.the-digital-picture.com/Reviews/Canon-Extender-EF-2x-III-Review.aspx

4. These same arguments apply also to the choice between full frame and cropped frame cameras. Images taken with a full frame camera have fewer pixels to describe the subject which covers a smaller fraction of the sensor surface. If a cropped sensor introduces degradations these would have to be quite severe to negate the pixel advantage they bring. Unfortunately, I do not have a full frame camera to conduct the necessary tests.


This Blog site also contains other Blogs I have written on photographic issues, see for example:




How to microadjust the Autofocus system on your Canon 7D, 7D MkII or 1DX camera:- http://johncrabbwildlifeimages.blogspot.co.uk/2015/06/fine-tuning-7d-and-7d-mk2-autofocus.html

and

How to best set up your 7D MkII to take shots of birds in flight:-
http://johncrabbwildlifeimages.blogspot.co.uk/2015/05/birds-in-flight-with-canon-7d-mkii.html

You can also see some of my photos at www.johncrabb.co.uk   and follow me on Facebook at  www.facebook.com/johncrabbwildlifeimages