300mm f/2.8L IS II vs. 400mm f/4 DO IS II ---> price difference

[Jack Douglas]

I have now done dozens of comparisons of the 1.4x and 2xTCIII on the 400mm DO II with the 5DS R. The 1.4x uprezzed 1.4x in Photoshop gives a slightly better image than the 2xTC. I think it boils down to a combination of the hit the 2xTC makes on IQ combined with the shift from f/5.6 to f/8 going through the diffraction limited aperture of f/6.7. The diffraction hit doesn't happen on the lower pixel full frames and you see an improvement in resolution from 400mm - 560mm - 800mm.

The 400mm DO II is spectacularly sharp at 560mm on the 5DS R. Here are two shots of a kingfisher with a fish I took yesterday afternoon. The little bird was over 12m away. There is some Moire, which I don't usually get with my less sharp lenses. I am tempted to get a new FF with fewer pixels and an AA filter. But, 560mm on the 5DS is so good. Maybe I'll use it on the 7DII.

If you uprez and view at 100% your CoC and 'diffraction limited aperture' changes because your magnification changes. The differences can then only be down to the optics of the 1.4 vs the 2x TC's.

Please elaborate and explain more clearly as your comment is not clear to me. Thanks.

DLA is defined as when the Airy disc covers one pixel.

For the sake of easy maths lets say your sensor is 6,000 x 4,000 pixels and that resolves 24,000,000 airy discs, it is a simplification but I believe holds true.

Now if you use a 2xTC and the subject covers the whole sensor (again for easy maths) you get those 24,000,000 airy discs and they each cover one pixel, as they should. Now if you swap the 2xTC for the 1.4TC your subject only covers 3,600 x 2,800 pixels giving you 10,080,000 pixels and 10,080,000 airy discs.

If you upres those 10,080,000 airy discs/pixels to the same pixel numbers as the shot from the 2xTC they each occupy an area equivalent to 24,000,000/10,080,000 or 2.4 pixels.

Effectively the airy disc is over twice the area from the upresed 1.4TC shot compared to the 2xTC shot. Now the airy disc was the same size at capture, one pixel, but you are now looking at that one pixel upresed so the corresponding airy disc is upsized. The fact that an airy disc now covers more than one pixel destroys the DLA assumption.

Same concept with the CoC and because of that the DOF all change. DOF should be around one stop difference, so you'd need to shoot at f5.6 with the 1.4TC to get the same DOF as the 2xTC at f8 when viewing output at the same subject size.

Boy, you lost me. Sounded more like a fairy tale. Alan maybe you can translate for me.

Jack

 

[AlanF]

I have now done dozens of comparisons of the 1.4x and 2xTCIII on the 400mm DO II with the 5DS R. The 1.4x uprezzed 1.4x in Photoshop gives a slightly better image than the 2xTC. I think it boils down to a combination of the hit the 2xTC makes on IQ combined with the shift from f/5.6 to f/8 going through the diffraction limited aperture of f/6.7. The diffraction hit doesn't happen on the lower pixel full frames and you see an improvement in resolution from 400mm - 560mm - 800mm.

The 400mm DO II is spectacularly sharp at 560mm on the 5DS R. Here are two shots of a kingfisher with a fish I took yesterday afternoon. The little bird was over 12m away. There is some Moire, which I don't usually get with my less sharp lenses. I am tempted to get a new FF with fewer pixels and an AA filter. But, 560mm on the 5DS is so good. Maybe I'll use it on the 7DII.

If you uprez and view at 100% your CoC and 'diffraction limited aperture' changes because your magnification changes. The differences can then only be down to the optics of the 1.4 vs the 2x TC's.

Please elaborate and explain more clearly as your comment is not clear to me. Thanks.

DLA is defined as when the Airy disc covers one pixel.

For the sake of easy maths lets say your sensor is 6,000 x 4,000 pixels and that resolves 24,000,000 airy discs, it is a simplification but I believe holds true.

Now if you use a 2xTC and the subject covers the whole sensor (again for easy maths) you get those 24,000,000 airy discs and they each cover one pixel, as they should. Now if you swap the 2xTC for the 1.4TC your subject only covers 3,600 x 2,800 pixels giving you 10,080,000 pixels and 10,080,000 airy discs.

If you upres those 10,080,000 airy discs/pixels to the same pixel numbers as the shot from the 2xTC they each occupy an area equivalent to 24,000,000/10,080,000 or 2.4 pixels.

Effectively the airy disc is over twice the area from the upresed 1.4TC shot compared to the 2xTC shot. Now the airy disc was the same size at capture, one pixel, but you are now looking at that one pixel upresed so the corresponding airy disc is upsized. The fact that an airy disc now covers more than one pixel destroys the DLA assumption.

Same concept with the CoC and because of that the DOF all change. DOF should be around one stop difference, so you'd need to shoot at f5.6 with the 1.4TC to get the same DOF as the 2xTC at f8 when viewing output at the same subject size.

Thanks for your helpful reply. I now realise what you are getting at but I don't think your example is fully correct but it is partly right.
1. Your first assumption is that the Airy disk with the 2xTC exactly covers a pixel. This assumption doesn't hold for the case we are discussing of the 5DS, where DLA is 6.7 µ and the aperture of the widest aperture of the f/4 DO with the 2xTC is f/8. The Airy disk is significantly greater than the DLA.
2. Your second assumption is that Airy disk is the same size in the presence of the 1.4xTC as for the 2xTC is puzzling. The diameter of an Airy disk is directly proportional to the f-number of a lens. The Airy disk for the DO with the 1.4xTC is f/5.6, 1.4x smaller than that for f/8 with the 2xTC.

The situation with the 5DS is a little complicated because the DLA is in the middle of the values for the 1.4 and 2xTCs. Extreme situations are easy to analyse.
Case 1. At one extreme, if the lens is diffraction limited before a TC is added, then a TC won't increase resolution since the increase in focal length is exactly balanced by the concomitant increase in f-number, which increases the diameter of the Airy disk.
Case 2. At the other extreme, if the lens is way below being diffraction limited, then a TC will increase focal length without a noticeable increase in diffraction and will always increase resolution (unless it is optically cr*p).

The 1DX has a DLA of f/11, and so both the 1.4xTC and 2xTC keep the 400mm f/4 under the DLA, approaching case 1, and the 2xTC gives more resolution than the 1.4xTC.

The situation with our 400mm DO at 560mm and f/5.6, below DLA of f/6.7, and at 800mm and f/8, above DLA is more difficult to calculate. If there is no hit on IQ with the 2xTC, then it should still outresolve the 1.4xTC. But, as we both agree, the optical quality is lowered so you don't get much advantage.

Does that help Jack?

 

[Jack Douglas]

;D ;D Not much. I've never read up on this topic so it's hopeless. OTOH, I did understand the last paragraph. Which leads me to believe you're not quite satisfied with this purchase?? Or, have you got another camera in mind? If so please do your homework well, so I may benefit.

If, and it's pretty likely, I sell the 300, I'm OK with my purchase in that 560 is very good and outclasses 600, especially relative to AF snappiness. However, I was really banking on using 800 a lot.

Jack

 

[AlanF]

Jack
The take home message is that if you have a full frame like the 1DX and to a slightly less extent the 5DIV, you will notice the increased resolution of the 2xTC. On a 7DII or similar density pixel 5DS R, there is not much extra resolution at 800mm than 560mm. However, the 400 + 1.4xTC on the 7DII or 5DS R is spectacularly good, so I am not disappointed.

Canon has raised the price of the 300mm f/2.8 II and I will wait until the higher prices filter through to the used market before selling.

 

[arbitrage]

Hi guys....

So today I did a few in the field comparisons on a perched bald eagle that stayed around long enough for me to shoot some 1DX2 and 5DSR both at 800 and 560.

I haven't had a good look at the 1DX2 files yet. However, after looking closely and doing some exports at different sizes I am starting to agree with Alan on the 5DSR. I found the up-ressed 560 shot looked just a bit better than the 800 shot. They were very close but obviously that means the 800 is not helping things and having the extra stop at 560 and possibly even a slightly better image makes me wonder about the 2xTC on the 5DSR.

Downsizing the 5DSR at 560mm and 800mm to the 1DX2 at 800mm, the two 5DSR images showed a bit more detail than the 1DX2 image. Again the two 5DSR images were very similar and probably the 560mm one was a very small amount better.

Later in the day I went and tried shooting some ducks/coots on a local marsh where reach is key. The 800mm on 5DSR was terrible. I'm still not sure if it was atmospheric interference over the water or a MA thing or what. It appeared to be back focusing so I tried some shots at -10MA and then -20MA. I still need to look closer at these shots to see if it is a MA issue. Although a quick look at the -20 shots seem to look better. I'm never happy if a lens needs -20 where it is fine on other bodies. For now I'm not reading much into these results as I need to evaluate it in a more controlled environment. When I had the Sigma 150-600C and tried it with a 1.4TC I had a similar issue where it did well on the lesser MP 1DX and 1D4 but was not good with any amount of MA on the 7D2. The results on the 5DSR today remind me of that scenario.

Still lots more testing to do to figure this all out. If I had a backyard I'd try to set up something controlled but living in an apartment my outdoor space on my deck is limited!!

 

[Jack Douglas]

Thanks Geoff.

Jack

 

[AlanF]

Arbitrage
Thanks for your comparisons. I am now feeling more comfortable with the quality of my own gear now you are getting similar results. My own feeling is that 1.4xTC on the f/4 and 7DIi or 5DS R is similar in IQ to the 2xTC with FF frame. The disadvantage of smaller pixels is balanced by the IQ hit of the 2xTC. It's similar to the 100-400mm II. On the 5DIII I was happy to use it with the 1.4xTC at f/8. But, on the 7DII I use it native.

Jack, before laying out on a FF, you might want to compare your gear at 560mm with a crop vs 800mm on your 6D.

I wish I could go out and test on a bald eagle. I had to make do with a grey heron at about 100m.

 

[Jack Douglas]

Arbitrage
Thanks for your comparisons. I am now feeling more comfortable with the quality of my own gear now you are getting similar results. My own feeling is that 1.4xTC on the f/4 and 7DIi or 5DS R is similar in IQ to the 2xTC with FF frame. The disadvantage of smaller pixels is balanced by the IQ hit of the 2xTC. It's similar to the 100-400mm II. On the 5DIII I was happy to use it with the 1.4xTC at f/8. But, on the 7DII I use it native.

Jack, before laying out on a FF, you might want to compare your gear at 560mm with a crop vs 800mm on your 6D.

I wish I could go out and test on a bald eagle. I had to make do with a grey heron at about 100m.

Good idea except that it's been kind of poor weather and winter is coming and my granite isn't finished - cursed project - why do I get into such things!

So if I set up on tripod and put a detailed subject like a dry leaf on my stucco wire fence at, what distance would you recommend? ISO 100, wide open, maybe 1/1000 s, anything else. I'd have to manually focus 800, magnified live-view. I've never used live-view as a rule and the other day I wasn't able to set a faster shutter - any idea what I was doing wrong?

So you will be happy and sell the 300?

Jack

 

[AlanF]

Arbitrage
Thanks for your comparisons. I am now feeling more comfortable with the quality of my own gear now you are getting similar results. My own feeling is that 1.4xTC on the f/4 and 7DIi or 5DS R is similar in IQ to the 2xTC with FF frame. The disadvantage of smaller pixels is balanced by the IQ hit of the 2xTC. It's similar to the 100-400mm II. On the 5DIII I was happy to use it with the 1.4xTC at f/8. But, on the 7DII I use it native.

Jack, before laying out on a FF, you might want to compare your gear at 560mm with a crop vs 800mm on your 6D.

I wish I could go out and test on a bald eagle. I had to make do with a grey heron at about 100m.

Good idea except that it's been kind of poor weather and winter is coming and my granite isn't finished - cursed project - why do I get into such things!

So if I set up on tripod and put a detailed subject like a dry leaf on my stucco wire fence at, what distance would you recommend? ISO 100, wide open, maybe 1/1000 s, anything else. I'd have to manually focus 800, magnified live-view. I've never used live-view as a rule and the other day I wasn't able to set a faster shutter - any idea what I was doing wrong?

So you will be happy and sell the 300?

Jack

Jack
I an a numbers man, and really like this chart prepared by Bob Atkins - version 2 from http://bobatkins.com/photography/technical/lens_sharpness.html

Post it where you can just about resolve the 4 lines per mm and the 2.8 circle, somewhere between 12 and 20m.

Happy to sell? I always have sellers remorse and only rarely buyers remorse. I really regret selling my 5DIII. But, if I had kept it I would hardly ever use it. I virtually never use 300mm f/2.8 but use it with the 1.4x or 2xTC. I'll probably only ever use it rarely in the future. At 420mm with 2xTC, it is a full pound (500g) heavier than the 400 DO and not quite as good in terms of IS, AF and IQ, but not as makes much difference. But, at my age, every pound less helps. The 300 is the more versatile lens, especially on crop or the 5DS R because you can use it at f/2.8, f/4 and f/5.6 whereas I would recommend the 400 only at f/4 and f.5.6, but not f/8. On FF, the 400 is better for birding. Nevertheless, I will regret selling the 300 as sooner or later I will want to shoot at dawn or dusk and will need the extra stop (or stops over my 100-440).

 

[Jack Douglas]

Thanks Alan,

I'm pretty much in the same situation as you and appreciate what the 400 brings to those who are not going to get stronger in the future, like forget weight training!

I will try out the chart and see what my performance is like and post when I can - still haven't done my 2015 income tax! Not to mention a large garden with a bumper crop this year that my dear wife is struggling with.

Yes, there will be regrets not having the 300. I shoot at 300 about as much as you do, hardly ever at 420 and a whole lot at 600 where the AF is less than stellar so 560 is big, but 800 would have been the really sweet item for me. Wishful thinking since the lens comparison resolution tests clearly show 800 somewhat behind 300 X2 and many (contrary to what Alan and Jack have preached) tended to write off that combo. Must be amazing to shoot with 600 F4.

Jack

 

[arbitrage]

Here are a few shots from Monday's shooting. I plan to continue to test the TCs on different bodies. I tried to MFA my 5DSR and 800mm last evening and it seemed to be fine at zero using the LensAlign target and angled ruler. I will try it again with Focal tomorrow.

The first 3 and the 5th are with 1DX2 and 400/2x
The 4th is 400mm bare lens
The last 2 are 800mm on the 5DSR

I will say that this lens is quickly becoming a favourite. I'm really considering selling the 200-400 and will sell the 300 2.8. If Canon makes that 600DO they showed the prototype for then I would happily sell off the 600II and have a three lens combo of 100-400II, 400DOII and 600DO....heaven!!

 

[Click]

Here are a few shots from Monday's shooting. I plan to continue to test the TCs on different bodies. I tried to MFA my 5DSR and 800mm last evening and it seemed to be fine at zero using the LensAlign target and angled ruler. I will try it again with Focal tomorrow.

The first 3 and the 5th are with 1DX2 and 400/2x
The 4th is 400mm bare lens
The last 2 are 800mm on the 5DSR

Beautiful shots. I especially like the first and 3rd picture. Well done.

 

[Jack Douglas]

Thanks and keep posting shots Geoff. I didn't even get out with mine today. I heard the Pileated woodpecker that has been very absent for a while so if he/she shows up that'll spur me on to see how the 400 compares with the 300. I'd be slightly more motivated if my 6D had F8 AF. Very nice photos.

Jack

 

[Jack Douglas]

Alan, I posted in the other thread because I noticed you were there?? Is there an official 400 DO thread from some time back? I'm so tired I don't have the energy to look.

Jack

 

[AlanF]

Here are some shots of my charts. They are taken at 12.6m. All at a similar time on the 5DS R, widest aperture, iso640. All have noise reduction but no sharpening. The lines on the chart are in lines/mm. The circles are also in linewidths/mm.
First the 300mm f/2.8 II at 300mm, 560mm and 600mm. There was some inconsistency of focussing - these are the best.

 

[AlanF]

The 400mm DO under the same conditions. At 400mm, it outresolves my 300mm + 1.4xTC III. Resolution increases significantly at 560mm, and is similar to the 300mm + 2xTC, but much more consistent. 800mm is similar to 560mm.

 

[AlanF]

For comparison, the 100-400mm II. It is surprisingly good at 400mm, and approaches the 400mm DO II. At 560mm and f/8 with 1.4xTC, the resolution increases. But, the DO at 560mm is clearly better.

 

[Jack Douglas]

Alan, I have a little difficulty keying in on what to look for - any comments? Comparing 560 to 800 (of most interest to me) - these are all sized to be full frame and shot from the same location so how does the magnification or reach factor in. Are you saying that you can just shoot at 560 and produce roughly the same output for viewing or printing (equal quality)? The 300 images did not open in a new window.

Jack

 

[AlanF]

Jack
They are all taken 12.6 m away from the target. All are the 100% crops from the centre of the image, 1 pixel in the crop = 1 pixel of the original. The 300mm is so small it doesn't need a separate window to open larger. The 800mm is quite large. The best way to view is to download them.

Look for distinct resolution of the individual lines in the L-shaped clusters in the top left quarter. Also look to see the resolution of the 2.8 circles in the bottom right. For example, the 560mm DO is close to resolving the 4.5 lines per mm. The 300mm can't resolve the 2.8 circles. The 100-400 and the 400 DO can. The 560s and 800mm resolve the 2.8s very clearly.
Alan

 

[Jack Douglas]

I took a couple shots of the moon tonight and 560 is a little better than my previous 300 X 2 and almost as good as 800. Manual viewfinder focus at 800 was iffy so maybe with perfect focus there would be a greater difference.

Jack