The 1200mm Sharpness Test

[AlanF]

Jon
My technique is:
1. Use DXO prime to wipe out noise, and don't spend ages on each one with paint brushes etc.
2. If you have the bird against a light background, don't use more than 1 pixel radius on sharpening otherwise you will get a halo, which is annoying to those who notice such things.

I think your final image looks unnaturally bright.

Unnaturally bright? I'm not sure I understand that...the processed image is actually darker overall than the original. Contrast was increased slightly, globally and locally, and saturation was boosted slightly. Other than that...I think it got darker overall, not brighter.

To be more precise, the colours look a bit too bright or vibrant.

 

[scyrene]

Jon
My technique is:
1. Use DXO prime to wipe out noise, and don't spend ages on each one with paint brushes etc.
2. If you have the bird against a light background, don't use more than 1 pixel radius on sharpening otherwise you will get a halo, which is annoying to those who notice such things.

I think your final image looks unnaturally bright.

I tend to tweak processing by shot, but I prefer to over sharpen at full size, then reduce the size, than reduce and then sharpen. Not sure if anyone else does it this way.

For final web versions, I usually use this technique. The oversharpen/reduce technique is fairly old, and often applied to landscapes in conjunction with one of the various soft contrast techniques. I think it gives better results. You do have to get a handle on background noise first, though...otherwise the sharpening enhances the noise.

Lol, I don't claim to have invented it, but I did arrive at it by myself I would usually use a high threshold and wider radius, so it applies above the level of noise grain, if that makes sense. I prefer to avoid denoise wherever possible, as it reduces image quality too much oftentimes.

 

[jrista]

Jon
My technique is:
1. Use DXO prime to wipe out noise, and don't spend ages on each one with paint brushes etc.
2. If you have the bird against a light background, don't use more than 1 pixel radius on sharpening otherwise you will get a halo, which is annoying to those who notice such things.

I think your final image looks unnaturally bright.

Unnaturally bright? I'm not sure I understand that...the processed image is actually darker overall than the original. Contrast was increased slightly, globally and locally, and saturation was boosted slightly. Other than that...I think it got darker overall, not brighter.

To be more precise, the colours look a bit too bright or vibrant.

LOL. Remember, it was a TEST! I wasn't interested in processing this thing to perfection. It's a junk photo that I simply snapped to see how well the 2x TC really performed. I think I spent about 2 minutes processing it. The color is immaterial...it's the sharpness that I was interested in.

 

[jrista]

Jon
My technique is:
1. Use DXO prime to wipe out noise, and don't spend ages on each one with paint brushes etc.
2. If you have the bird against a light background, don't use more than 1 pixel radius on sharpening otherwise you will get a halo, which is annoying to those who notice such things.

I think your final image looks unnaturally bright.

I tend to tweak processing by shot, but I prefer to over sharpen at full size, then reduce the size, than reduce and then sharpen. Not sure if anyone else does it this way.

For final web versions, I usually use this technique. The oversharpen/reduce technique is fairly old, and often applied to landscapes in conjunction with one of the various soft contrast techniques. I think it gives better results. You do have to get a handle on background noise first, though...otherwise the sharpening enhances the noise.

Lol, I don't claim to have invented it, but I did arrive at it by myself

Sure. I know that... ???

I would usually use a high threshold and wider radius, so it applies above the level of noise grain, if that makes sense. I prefer to avoid denoise wherever possible, as it reduces image quality too much oftentimes.

Sharpening noise reduces image quality as well. That's the problem with noise....the more you start out with, the tougher it is to process without experiencing a degradation in IQ. Denoising is valuable, even if it is a light application. At the very least, lightly denoise then process, and you might end up with the original noise you started with, rather than more.

 

[AlanF]

There is a whole a group of who use the 2xTCIII on the 300/2.8 II. There is not a noticeable difference between the sharpness of that combo and the native 600/4 II, aside from the loss of a stop or two.

http://www.the-digital-picture.com/Reviews/ISO-12233-Sample-Crops.aspx?Lens=748&Camera=453&Sample=0&FLI=0&API=0&LensComp=739&CameraComp=453&SampleComp=0&FLIComp=2&APIComp=4

The 600/4 II + 2xTCIII is very similar to the Tamron 150-600mm, which though frequently criticised, does sharpen up nicely as well.

http://www.the-digital-picture.com/Reviews/ISO-12233-Sample-Crops.aspx?Lens=748&Camera=453&Sample=0&FLI=2&API=2&LensComp=929&CameraComp=453&SampleComp=0&FLIComp=5&APIComp=1

 

[scyrene]

I tend to tweak processing by shot, but I prefer to over sharpen at full size, then reduce the size, than reduce and then sharpen. Not sure if anyone else does it this way.

For final web versions, I usually use this technique. The oversharpen/reduce technique is fairly old, and often applied to landscapes in conjunction with one of the various soft contrast techniques. I think it gives better results. You do have to get a handle on background noise first, though...otherwise the sharpening enhances the noise.

Lol, I don't claim to have invented it, but I did arrive at it by myself

Sure. I know that... ???

I would usually use a high threshold and wider radius, so it applies above the level of noise grain, if that makes sense. I prefer to avoid denoise wherever possible, as it reduces image quality too much oftentimes.

Sharpening noise reduces image quality as well. That's the problem with noise....the more you start out with, the tougher it is to process without experiencing a degradation in IQ. Denoising is valuable, even if it is a light application. At the very least, lightly denoise then process, and you might end up with the original noise you started with, rather than more.

You don't have to take my word for it, just saying It's pretty easy to figure out I guess. Anyway. Denoise... I did used to do it in Lightroom, and lightly as you say. Depends on a lot. Colour denoise certainly - it was good for that. Often with a touch extra saturation to make up the difference. But my preference is for more luminance noise than less - although I would apply more on out of focus backgrounds with a brush, as discussed above, if the situation warranted it. How I miss proper processing tools

 

[jrista]

What's the consesnus on the 1.4x II vs the 1.4x III? I own a 1.4xII and 500 f/4 IS USM and I leave the 1.4 in the closet. The loss in resolution is very discernible.

The 600/4 II with 1.4x TC III seems to produce very sharp results. It is not as sharp as possible right in the corners, but overall, I don't feel as though I'm losing anything with the 1.4x. The 2x produces what you see here, an apparent softening, but it is pretty easy to clean up. The TC III series were designed specifically to maximize the potential of the supertele L II series...so it I suspect the results are better than with the TC II series and original supertele L series lenses.

 

[scyrene]

What's the consesnus on the 1.4x II vs the 1.4x III? I own a 1.4xII and 500 f/4 IS USM and I leave the 1.4 in the closet. The loss in resolution is very discernible.

AFAIK the mark II v III extenders only improve AF functionality, rather than image quality. Neuro knows more, I think.

 

[jrista]

The TC III's improve corner performance over the TC II's as well, on top of the improved AF capabilities (which I believe only work with the 1D X.)

 

[jrista]

Alrighty. I've done some more testing. I am beginning to distrust FoCal...it does not seem to be calibrating my equipment as ideally as it could be. I decided to do a quick and dirty AFMA check and tweak. I just set it at 0, -10, -20, +10 and +20. Took a few shots of the same target (lens and camera were on a tripod.) The differences were fairly obvious, +10 looked best. I tried +5, that looked slightly better.


The differences with the 600 + 2x TC OOC look MUCH better now. The birds are quite sharp strait out of camera, and they get very sharp with just a little sharpening. Below is a single photo from a more challenging burst of frames, as I was at 1/100th shutter...with a Chickadee. Anyone who's photographed Chickadees knows how ludicrously insane it is to try and get a sharp result at 1/100th second.


Anyway, managed to (with the limited 6fps of the 5D III even) get one frame that was sharp, at 1200mm f/10, ISO 800. I'm attaching three full size 1:1 scale crops, of the original image, sharpened only, then with a little bit of toning to bring down those highlights...just to reveal all the detail that is there. Personally...I'm pretty impressed at how sharp the 600/4 L II+2x TC III can be strait out of camera. My earlier softness was apparently just due to an improper AFMA selection by FoCal (that my extremely rudimentary and hackish approach handily bested.)

 

[AlanF]

To those knocking the sharpness of the Tamron 150-600, here are a couple of shots I took of black-capped chickadees at 600mm f/8. The upper one face on is half the size of Jon's chickadee but, even so, sharper - look at the feet for example. The lower one sideways on is closer and much sharper.

 

[jrista]

Alan, a couple of things. First, did you click on my images for full size?


Second, regarding the feet. I was at 1200mm f/10...working with a ridiculously thin DOF. My chickadees feet aren't sharp because they are not in focus. My goal wasn't to get sharp feet, though...it was to get the head and eye sharp. To that end, I believe, despite 1/100s, I succeeded. (I just wish he hadn't hopped from the beautiful pine branch to the tripod before I did! )


As for the sharpness of your shots overall...honestly, I'm a little disappointed. Maybe it's just the disheveled nature of the birds, not sure...but, I guess I kind of expected more out of that lens. The Canon 100-400 could get MUCH sharper than that...and I mean the OLD 100-400...

 

[AlanF]

Show some much sharper ones you have taken with the 100-400mm.

 

[jrista]

Alright. Here is a set of shots taken with the 7D and 100-400 L (original). I grabbed a few, with birds of varying sizes in frame, to bring some diversity of pixels-on-target to demonstrate that isn't necessarily the issue with the 150-600. These are all original shots, original crop, no scaling, no processing, no sharpening or NR of any kind. Strait out of camera RAW exported to 75% quality JPEG from Lightroom. (Blame CR forums for links to images instead of just images...I don't know what it was doing, but it wouldn't let me post with them embedded as images.)


http://i.imgur.com/TiKthrU.jpg


http://i.imgur.com/Z2vWEQF.jpg


http://i.imgur.com/TSoCerX.jpg


http://i.imgur.com/ZBAK66E.jpg


Every image here appears to be sharper than the 150-600mm shots. To be fair, one of your shots seems to have some motion blur. For the other, I cannot say, not really sure if there is any motion blur or not, but it still seems a little soft. Not as sharp as either my 100-400 shots or my 1200mm f/10 chickadee. Now, don't get me wrong. I'm a fan of the idea of an affordable lens that reaches 600mm. For the novice or budget birder, I think having such a lens is a HUGE benefit. That said, if the sharpness from your example shots is around the best the lens can do at 600mm, then I'm rather disappointed. I'm happy to accept if the issue is technique, or too slow a shutter speed, or lack of IS use (or IS kicking in and screwing up the shot), etc. If you can demonstrate as much, then more power too you, prove me wrong! But, as it stands, I don't think the sharpness of those shots is what I would call "ideal"...I think my 100-400 does better, and my 600 with a 2x TC at 1200mm f/10 (!!) does better.

 

[jrista]

Alright. Here is a set of shots taken with the 7D and 100-400 L (original). I grabbed a few, with birds of varying sizes in frame, to bring some diversity of pixels-on-target to demonstrate that isn't necessarily the issue with the 150-600. These are all original shots, original crop, no scaling, no processing, no sharpening or NR of any kind. Strait out of camera RAW exported to 75% quality JPEG from Lightroom.

Every image here appears to be sharper than the 150-600mm shots. To be fair, one of your shots seems to have some motion blur. For the other, I cannot say, not really sure if there is any motion blur or not, but it still seems a little soft. Not as sharp as either my 100-400 shots or my 1200mm f/10 chickadee. Now, don't get me wrong. I'm a fan of the idea of an affordable lens that reaches 600mm. For the novice or budget birder, I think having such a lens is a HUGE benefit. That said, if the sharpness from your example shots is around the best the lens can do at 600mm, then I'm rather disappointed. I'm happy to accept if the issue is technique, or too slow a shutter speed, or lack of IS use (or IS kicking in and screwing up the shot), etc. If you can demonstrate as much, then more power too you, prove me wrong! But, as it stands, I don't think the sharpness of those shots is what I would call "ideal"...I think my 100-400 does better, and my 600 with a 2x TC at 1200mm f/10 (!!) does better.

 

[AlanF]

The Canon 100-400 could get MUCH sharper than that...and I mean the OLD 100-400...

First of all, the two shots of the chickadees are not necessarily the best I have or even representative of the lens but they are the only two I have of that North American bird (I am in the UK), but they are similar to your shot of the chickadee.

Secondly, the quality is comparable to those of yours from the 100-400, when comparing them side-by-side and certainly not MUCH less sharp.


I was never able to get very sharp shots with the old 100-400 on a 7D. Top is a dunnock at 400 on the 7D, typical of my efforts, below is a dunnock taken on the 300/2.8 + 2xTC III on the 5D3. They are chalk and cheese.

 

[East Wind Photography]

Great thread guys. I like the real world discussions and comparisons without all of the theoretical BS we normally see and read.

I use the 2xiii with my 300 II frequently but for some reason when I put it on my 600 f4 I it back focuses so severely that afma -20 doesn't even get it close using multiple bodies. Thinking about sending it in to canon for a tweak. Manual focus is a pita for birds!

 

[jrista]

The Canon 100-400 could get MUCH sharper than that...and I mean the OLD 100-400...

First of all, the two shots of the chickadees are not necessarily the best I have or even representative of the lens but they are the only two I have of that North American bird (I am in the UK), but they are similar to your shot of the chickadee.

Secondly, the quality is comparable to those of yours from the 100-400, when comparing them side-by-side and certainly not MUCH less sharp.


I was never able to get very sharp shots with the old 100-400 on a 7D. Top is a dunnock at 400 on the 7D, typical of my efforts, below is a dunnock taken on the 300/2.8 + 2xTC III on the 5D3. They are chalk and cheese.

You know, looking at your images again...the birds do not appear to actually be in focus. Well, the first one, I'm not sure...that whole image has a softish appearance. However the second one, below the bird and a little forward everything gets very sharp. I was looking at the birds only before, and I was seeing softness. Going by your second image (of the original two you posted), I'd say the 150-600 is acceptably sharp.


Regarding the dunnock shot with the 7D, it also looks focused a little forward of the bird. That just sounds like an AFMA issue.


I've found that the great whites are pretty much dead on for the most part. My 600 is at 0 AFMA on all my bodies, except with the 2x TC, where I had to move it to +5 AFMA. My 100-400, I actually had to tweak the AFMA a lot on that...and I think in the end I ended up around -15 or -18. That was with FoCal as well, and I've learned that FoCal may not really be as accurate as it seems....which may just be due to the quality of the printed target, not sure. I bet my 100-400 could get a lot sharper if I sent it in with my body to have it adjusted...but then I'd be out my camera and a lens for a while, and it would be an extra cost. I plan to sell that lens anyway, so I'll leave it up to the buyer to have it adjusted for their body if they need (which they very well may not).

 

[East Wind Photography]

The Canon 100-400 could get MUCH sharper than that...and I mean the OLD 100-400...

First of all, the two shots of the chickadees are not necessarily the best I have or even representative of the lens but they are the only two I have of that North American bird (I am in the UK), but they are similar to your shot of the chickadee.

Secondly, the quality is comparable to those of yours from the 100-400, when comparing them side-by-side and certainly not MUCH less sharp.


I was never able to get very sharp shots with the old 100-400 on a 7D. Top is a dunnock at 400 on the 7D, typical of my efforts, below is a dunnock taken on the 300/2.8 + 2xTC III on the 5D3. They are chalk and cheese.

You know, looking at your images again...the birds do not appear to actually be in focus. Well, the first one, I'm not sure...that whole image has a softish appearance. However the second one, below the bird and a little forward everything gets very sharp. I was looking at the birds only before, and I was seeing softness. Going by your second image (of the original two you posted), I'd say the 150-600 is acceptably sharp.


Regarding the dunnock shot with the 7D, it also looks focused a little forward of the bird. That just sounds like an AFMA issue.


I've found that the great whites are pretty much dead on for the most part. My 600 is at 0 AFMA on all my bodies, except with the 2x TC, where I had to move it to +5 AFMA. My 100-400, I actually had to tweak the AFMA a lot on that...and I think in the end I ended up around -15 or -18. That was with FoCal as well, and I've learned that FoCal may not really be as accurate as it seems....which may just be due to the quality of the printed target, not sure. I bet my 100-400 could get a lot sharper if I sent it in with my body to have it adjusted...but then I'd be out my camera and a lens for a while, and it would be an extra cost. I plan to sell that lens anyway, so I'll leave it up to the buyer to have it adjusted for their body if they need (which they very well may not).

My 600 by itself AFMAs around +4 which is reasonable which leads me to think my 2Xiii may be out of kilter...though it's not WAY out on my other lenses. I wonder if they can adjust the extenders electronically rather than needing to do it physically?

It took me a long time as well to realize FoCal is not more accurate than doing it manually using a scale. There are way too many factors that can affect the results, not to mention the constant tinkering with their algorithms. I actually had better success in their earlier versions and then things got complicated and went downhill. IMHO.

 

[jrista]

The Canon 100-400 could get MUCH sharper than that...and I mean the OLD 100-400...

First of all, the two shots of the chickadees are not necessarily the best I have or even representative of the lens but they are the only two I have of that North American bird (I am in the UK), but they are similar to your shot of the chickadee.

Secondly, the quality is comparable to those of yours from the 100-400, when comparing them side-by-side and certainly not MUCH less sharp.


I was never able to get very sharp shots with the old 100-400 on a 7D. Top is a dunnock at 400 on the 7D, typical of my efforts, below is a dunnock taken on the 300/2.8 + 2xTC III on the 5D3. They are chalk and cheese.

You know, looking at your images again...the birds do not appear to actually be in focus. Well, the first one, I'm not sure...that whole image has a softish appearance. However the second one, below the bird and a little forward everything gets very sharp. I was looking at the birds only before, and I was seeing softness. Going by your second image (of the original two you posted), I'd say the 150-600 is acceptably sharp.


Regarding the dunnock shot with the 7D, it also looks focused a little forward of the bird. That just sounds like an AFMA issue.


I've found that the great whites are pretty much dead on for the most part. My 600 is at 0 AFMA on all my bodies, except with the 2x TC, where I had to move it to +5 AFMA. My 100-400, I actually had to tweak the AFMA a lot on that...and I think in the end I ended up around -15 or -18. That was with FoCal as well, and I've learned that FoCal may not really be as accurate as it seems....which may just be due to the quality of the printed target, not sure. I bet my 100-400 could get a lot sharper if I sent it in with my body to have it adjusted...but then I'd be out my camera and a lens for a while, and it would be an extra cost. I plan to sell that lens anyway, so I'll leave it up to the buyer to have it adjusted for their body if they need (which they very well may not).

My 600 by itself AFMAs around +4 which is reasonable which leads me to think my 2Xiii may be out of kilter...though it's not WAY out on my other lenses. I wonder if they can adjust the extenders electronically rather than needing to do it physically?

It took me a long time as well to realize FoCal is not more accurate than doing it manually using a scale. There are way too many factors that can affect the results, not to mention the constant tinkering with their algorithms. I actually had better success in their earlier versions and then things got complicated and went downhill. IMHO.

As far as I can tell, when I AFMA with the 600+TC, it actually shows "600 f/4 L + 1.4x TC" or "600 f/4 L + 2x TC" in the AFMA screen. So, I believe the camera considers the lens with a TC to be different than the lens without, meaning each combination can be independently AFMAed.


BTW, setting AFMA does not actually change the lenses or TCs. It just registers the offset in camera memory, and the offset is applied to AF commands. So, when the camera tells the lens to move the focus group, it tells them to move WITH the offset already applied. The lenses never change, only the body changes. That's why you can AFMA a lens on multiple bodies concurrently, an the adjustment will stick for each body...only the body is actually being modified.