RF Native Lens Cropped vs 1.4x vs 2x

After I finished testing my lens, it occurred to me the method I used would probably give me a comparison from cropping to using an extender.
I had set the test up at the same distance (real world you would be using the extender or cropping from the same distance). Shutter speed remained constant at 1/500, all the shots are wide open and the ISO is set to Auto. This means that each shot has a higher ISO, with the exception of the 2x which I supplemented the light because it had become unusable. The 2x has a lower ISO than the native lens. The samples are cuts of screen shots hand scaled and zoomed in to be the exact same size (the exception is the Native shots which wouldn't scale precisely but were close), where I am now I didn't have editing software. All shots are with the R5. YMMV if you are using a lower MP body.

In the center what I am seeing is a benefit to use the 1.4x in the center. The Native crop is slightly better than the 2x, compound that with the loss of light due to the higher aperture I give the Native the edge. The 1.4x Crop is straight up better than the 2x. My opinion is that for any long work you would use the 1.4x and crop rather than use the 2x.



Attachment 2955Attachment 2957Attachment 2956

For the Corners the 2x is weak. The Native might have a little edge over the 1.4x cropped, and it might not. It is debatable.

Attachment 2958Attachment 2959

At this point I see no reason to own the 2x. I can get the same results by cropping and not lose the light. For distance the 1.4x is the way to go.

As for the new lenses released, I think you would get fair results out of the 800mm but you loose light which is a big negative. The 1200mm MTF's show the lens as sharp as the 600mm with the 2x in the center. If that is the case the 1200mm is going to be a semi soft lens and for my use very limited applications.

Neuro did a comparison of the 1.4x and 2x EF versions vs the RF versions. It is a good read. Hope he doesn't mind, but here is his link:
https://www.canonrumors.com/forum/th...l-is-ii.41215/

I wasn't able to get the 2x Corner and Native Corner comparisons in the last post.
Here they are.

Attachment 2960Attachment 2961

Here is the same test on the 500mm. It was shot at 20' from target where the 600mm was shot 24'.
The 1.4x is excluded because of the bad results, something isn't right with the extender.
The Native was better in the center. I give the 2x the edge in the corners.
It should be noted the 2x had a very high ISO.
My opinion is that if you have a select purpose to use the 2x and have time to arrange it it could be used. But on the fly you could crop and accive close to the same results.

Attachment 2962Attachment 2963Attachment 2966Attachment 2964

Well something you have to consider is the affect of overall IQ or how far you subject is.

If you are very far away, you have to have some kind of resolution on target you'd have to have the 2x on and crop from there. Crop too heavy and it will look that way.

You also can't escape the increase in noise which is proportional to bumping up ISO with crop. You are not gaining light by not usingan extender. You increase perceived noise at the same output size with cropping just as much as using the extender and bumping it up. Used to be in the old'en days when using the 1Ds III as your body of choice. You had a native ISO 1600, expanded 3200. So it wasn't always a simple matter of just bumping up the ISO or pushing it past in post. So sometimes cropping was the solution. Sensor tech was not there yet.

Today with say the 1Dx III, it handles noise like a boss and a native ISO of 100,000!!!!!!

I'm just tossing in more variables in the mix and more scenarios why you might want to use the 2X. I would suspect you only have so much room to crop that regardless of the softness just having enough sensor and resolution on target will be more important.

Also, If you make another test like this. Maybe do it outside so you are able to use ISO 100 on all your shots and not inducing detail killing noise. Also with the exposures being so dim and white balance off. It kinda plays tricks with my eyes and makes it more difficult to judge them. I can still tell what is sharper. But them being the same exposure and WB would be nice.

Also, did you manually focus these results? I would not trust the AF for critical results. Because some look like they may have been miss focused by a significant amount. It looks much more than softness on at least one of them. The 2X results.

Just my 2 cents anyway.:)

Yes manual focus, IS off, on my heavy duty Gitzo tripod and Wimberly head. Set the 10 second timer because you could see vibration from your hand when touching the camera.

You could test multiple ways, truth is this is probably pretty close to what I would be doing in the real world. For instance it is a rare day I would be able to have an ISO of 100 at the prime shooting time. It is common in midday when your doing something like Road Hunting in Yellowstone. The higher ISO would come closer to showing a failure than if you had great light and ISO 100.

I thought it was a fair comparison and here is why, a shot usually presents itself at a given location and distance at the exact same time. So a real life comparison would be the results you would get by cropping vs adding in the extender at the same distance in the exact same light. I could have decreased the ISO by going to a slower shutter speed, but in the real world I would seldom be under 1/500 and almost always shooting wide open.

I found it curios that the RF 2x handled the higher ISO so poorly. The EF 2x did much better. I have thought / or at least felt for years that putting the EF 2x on was not that much different the cropping. This goes back to my first big white a dozen years ago.

It might be overly critical to be shooting a target with a square that is .8 of an inch square and comparing it on a 32" monitor blown up to 8"x 8"

Lol, right!

I do get the logic, and it is fair assumption. But sometimes trying to mimic real world scenarios don't make it easy to narrow in on a particular thing we are trying to test. In this case sharpness.

To be fair, I wonder about the 2X extender results. Maybe sample variation or sometimes even 1/500 is not fast enough, maybe just a tiny bit OOF. I have been known to flub manual focus too once in a while.

I just have a hard time accepting that the 2X is worse than cropping, lol. Because it would have to very soft to throw away all the sensor resolution and still come out on top. Not necessarily doubting your trials though. There could be other reasons for it.

I'm trying to edit the samples you gave, but for some reason DPP won't edit the files. I can't do any general adjustments for some reason.

I was gonna try my hand at getting all the white balances good, correct for all lens defects, use lens optimizer to apply CA and correct for lens abreactions/corner sharpness, any extremely minor noise reduction if necessary and sharpening to taste. As if it was a real photo. Just your basic run of the mill postprocessing.

And see with real world post processing how it turns out. And much easier to see the sharpness difference without the WB shift and exposure difference.

Actually it be nice to play with the RAW files altogether. LOL.

Quote:

---

Originally Posted by HDNitehawk

My opinion is that if you have a select purpose to use the 2x and have time to arrange it it could be used. But on the fly you could crop and accive close to the same results.

---

Agreed. I only occasionally travel with the 2xTC anymore and the couple times I used it were for tests. Whereas I use the 500 II native or with a 1.4x TC all the time. Both those results are stellar, IMO. I do see a noticeable drop off in IQ, but it is still capable. Rather, the two main issues I have are AF speed and once need a 2x TC on a 500 mm lens, your subjects are so far out there that even atmospheric effects come into play.

I do wonder if I owned a 300 f/2.8, for example, if I would use it more. But, as it stands, it is a piece of glass I am not sure if I would buy again.

Not surprised, but I am glad to see the 500 II hold up. Granted, I think the sharpest image is the 600+1.4xTC. Again, not surprised, but looks like you have a great lens in your hands.

Thanks for the tests. It is interesting.

Quote:

---

Originally Posted by Fast Glass

Lol, right!

I do get the logic, and it is fair assumption. But sometimes trying to mimic real world scenarios don't make it easy to narrow in on a particular thing we are trying to test. In this case sharpness.

To be fair, I wonder about the 2X extender results. Maybe sample variation or sometimes even 1/500 is not fast enough, maybe just a tiny bit OOF. I have been known to flub manual focus too once in a while.

I just have a hard time accepting that the 2X is worse than cropping, lol. Because it would have to very soft to throw away all the sensor resolution and still come out on top. Not necessarily doubting your trials though. There could be other reasons for it.

I'm trying to edit the samples you gave, but for some reason DPP won't edit the files. I can't do any general adjustments for some reason.

I was gonna try my hand at getting all the white balances good, correct for all lens defects, use lens optimizer to apply CA and correct for lens abreactions/corner sharpness, any extremely minor noise reduction if necessary and sharpening to taste. As if it was a real photo. Just your basic run of the mill postprocessing.

And see with real world post processing how it turns out. And much easier to see the sharpness difference without the WB shift and exposure difference.

Actually it be nice to play with the RAW files altogether. LOL.

---

Neuro had posted his findings on EF vs RF and he felt the RF extenders were softer.
If his findings are correct then that might be some of it.
It might be that I had a bad copy.
One of the things I do also is compare from reviews, things that Bryan has said, and look at the charts and see if the differences I am seeing from one similar lens to another. In this instance I looked at the 500mm vs the 600mm, granted not identical but it gives you an idea.
https://www.the-digital-picture.com/...mp=2&APIComp=0
After all that I do believe I was seeing an accurate representation.
I also had 2 sets of RF 2x pictures, only the best picture out of the best set was used. WB is off on the RF 2x because of the type of lighting to get the ISO down.

The pictures were screen shots done on the fly at the office.
I have the JPG here at the office and if you want I could put them in a drop box and shoot you a link.
Or I could even forward them.
If you want them shoot me an message with your email address.

Yes! I would like that, see how they hold up to post. Maybe their is some details than can be resurrected.

But yes they look pretty similar, but still with critical sharpness any amount of error, and it doesn't take much, could introduce softness.

I was just reading up on how Bryan does his crops because I'd like to so something similar for my own uses. As I'm sure you know it is pretty involved and he goes through a lot of effort to get some really clean results. Just one of the things he does is shoots remotely via computer monitor and focuses and refocused as much as 15 or 20 times! And then picks out the very best. I can see where just using the back of our cameras not being scientifically accurate or consistent enough to trust it 100%.

Or even the chart itself, it's a very finely printed one at 4000 DPI. Smooth texture. Large. Mounted on a 1/2" glass. Laser aligned ect. Definitely makes me apriciate the work that went into these crops.

Maybe your shots are representative, I can't really say. And you don't necessarily need to go to quite that far to know if your lens is sharp. But it sure does rule out a lot of variables.

Just food for thought. Your 2X ones are the only ones that seen quite soft to the point I do question them. I feel they are just soft enough where they are on the level or very slightly worse than the cropped image. And it would take very little to make it that way. Even just 10% sharper would make them slightly better than cropping. And then I could sleep at night.:p

My comparison really was for me to settle it in my own mind. It is no where the level that Bryan uses.
I wasn't surprised on the 2x. I thought it was similar.
The RF 1.4x exceeded my expectations.
Years ago I remember doing the same comparison in the field on my first 500mm and I used stuffed animals.
Plus I know the EF 1.4x results are bad, I need to see if something is wrong with the extender or my method failed.
I sent you the link, if you do not get it let me know.
I could do the RAW but it would take a substantial amount of work.

I think Jpeg serves the purpose. And yes, I just looked and I got the shots. I'll check them out on my PC later tonight. And thank you for satisfying my curiosity. Lol.

I think maybe doing it with something like stuffed animals might be beneficial since it would be more forgiving if the focus was off just a smidgen. And do some basic post processing to mimic what you would do in real life. Primarily the 2X, since the 1.4 looks very good I doubt there was any kind of user error.

But I hear ya, it very well could be that those are indeed the results and I'm gonna have to lose some sleep. lol.:D

I think I'm gonna have to do the same thing with my 300mm f/2.8. Realistically even if it is a wash I'll be using the 2x anyway just because I like framing my shots with the correct focal length and I still get to use more of my sensor.

If it is softer, then I would have to rethink it. But the 300mm f/2.8 handles extenders pretty well so I think it will be fine.

Thank you again HD for the full sized samples!

For some reason I couldn't get them to edit. But I do get a better feel for the sharpness

So after some pixel peeping they are exceedingly close in sharpness. But I cropped a portion of the chart which helped me see the differences a little bit better. These are mid frame shots, extreme corner sharpness is not of prime importance to my style of shooting but mid frame is as that is where I like to put my subjects a lot of the time.

To my eyes the 2x beats out the native shot cropped to the same dimensions and viewed at the same dimensions. But they are exceedingly close! But if you count how much more resolution the 2X has it makes it lean more towards the 2X even more.

Native:

https://community.the-digital-pictur...7&d=1646519282

600mm with 2X:

https://community.the-digital-pictur...8&d=1646519290

But the other thing about this is the 2X has far more pixels on target, and more sensor. So if you are in a situation of excellent light the 2X would have the advantage of the whole sensor's dynamic range and colors. Even with sharpness being very close. Low light, might be worth doing some testing to validate that.

What you guys think? Which one would make more sense to use?

Main thing is I now can sleep at night knowing it is still not worse to use a 2X extender. Even if it is very close. LOL.




Edit, now looking at the crops. The left hand of the images look softer than the right hand. Particularly the 2X results. But the crop is from the right hand of the side of the chart. So if there was a noticeable fall off in sharpness it would be the other way around.

Either the chart is not parallel or there is some decentering going on with the lens.

The 2x comes far closer to making sense mid day with excellent light, but with all the advantages and disadvantages weighed I think for myself the window is to narrow for it to be useful. The center is more important to me than the mid, but both are needed.

I figured I would have some corner variation. That is nothing more than a slightly out of square target and narrow DOF.

I am curious why you couldn’t edit them. They are straight out of the R5. The files were emailed to my wok account through gmail and google drive, then loaded in Dropbox. Maybe something was lost in the process.

I'm not sure, I can do some functions. But extremely limited. Like NR and cropping. I know if I had the RAW files I'm sure I could edit them. But I think the Jpegs served their purpose. Update; For some reason it does this to other Jpegs that I didn't process from from RAW, weird.

I would say if the target is off a reshoot for the 2X results VS native would be beneficial. If the mid frame is off like it is, I can see it transitioning softer towards the center of that very tiny crop (This as you know is a VERY tiny section) it means the center is not in focus by a fair amount. Which to me looks about right because it looks more than just softness. It looks OOF.

Up to you if you want too reshoot. But I think getting it as square as you can. Getting the chart glued to something flat. Well lit at ISO 100, flash lighting. Try manual and AF. Do more than 2 sets of focusing, do at least 10 or more for every combination and aperture. If you can use a remote monitor even better. It might also be worth getting the chart printed larger for a few bucks, then you will get much clearer lines, and standing back further gives a little more DOF and margin for error.

Then you can see for sure what is going on, right know I would say the 2X results are misleading and not accurate.

I'll be trying to recreate Bryans setup. Minus the crazy expensive charts (Unless I can find one for significantly less than $1000 plus price tags), I'll get a big one printed out professionally and it will serve the purpose. Mounted on glass and laser aligned ect. (I do got access to laser levels to do this). Basically recreate it exactly how he has it. That way I can compare mine to his charts accurately.

Because I got several lenses (Many more coming) that I want to know if they are focusing absolutely perfectly or not. Or if they are a good sample or not.

But this is just me, I don't want to tell anyone what to do.:)

All those are good suggestions for being more accurate. The narrower DOF of the 2x comes in to play and as you mentioned even gluing the target down might improve the results in some areas. Same thing on making it square. I could have worked a few more hours to make at it as precise as possible. A different printer would have helped as well.

I do think the focus is on. I refocused every shot so I had three samples. The method I used I always have good results. Any and all shake was eliminated and I focused at 10x.

Look at Bryans chart. I think the results were similar to Bryans. I think it is representative of what I am seeing. Toggle back and forth between the 600mm at native vs extender. The Center is noticeably softer.

https://www.the-digital-picture.com/...mp=0&APIComp=1

Unfortunately I can not repeat the test, and if someone asks and I could I would. I sent the 2x back. With my preferences I realized that I would not be happy with it. The RF 2x would have to perform beyond its abilities. The EF 2x III I have is hardly ever used on my 500mm. It was only useful in very rare instances, and in those instances fine resolutions wasn't as important. Also if I keep my 500mm I am good to 1000mm if needed for a certain application.

I researched it a little more after reading your response.
The WB difference is miss leading, but also telling in that I had to add more light to get a decent result.

I did a little PP on the 600mm 1.4x and 2x just to see what I could get.
Initially I applied white balance to both and I have to say that the 2x is slightly better.
But the 1.4x was shot with a higher ISO, so I went ahead and applied sharpening and and noise reduction.
With noise reduction and sharpening on both they were fairly equal.
My conclusion is this, you can get about the same results out of both methods. Maybe the 2x will give you slightly better in certain situations but not "significantly" better. It also depends on who is doing the processing and how a person flavors it.

I dropped some of the RAW files in the Drop Box folder if you still want to see them.

Yes for sure!

My conclusions were pretty much same for effective sharpness. Because resolution is not everything if the overall sharpness is not there. But we both knew that. LOL.

My thought process, and I'll be doing this as soon as I get my mark III extenders on my 300mm. Is to rule out any errors in the charts. Which looking at them I do see slight alignment issues. But if even slightly canted and still holding its own. It will do even slightly better in real world were DOF rarely manifests itself on a flat plane like a chart.

I think an interesting test would be in low light. In theory you could crop or use an extender and it give you the same noise levels at the same output size.

But putting it to the test and see if there other variables that could influence it.

So after messing around with the RAW files I was impressed with what a difference doing basic postprocessing made!

There was MUCH to much luminance noise reduction, like FAR too much! I completely mushed the details of everything. Used 0 luminance NR and chrominance and 2 color moire. There was a lot of chrominance noise.

Second thing I did was set it to what I like in the picture style as a base which is lowest contrast highest saturation (Lower the contrast the higher the saturation I need to maintain good saturation). This is my preference for post processing. To each there own no right or wrong here.

I applied lens corrections including maximum Digital Lens Optimizer. Reduces CA and improves corner sharpness ect.

Sharpening is pretty fine compared to the standard sharpening setting that was one there, it was REALLY blunt and very good at all. I set it to 1.2 strength, .5 radius, 1 threshold. This is my taste in sharpening, relatively fine and not to strong.

And of course click white balance correction.

So, with my taste in postprocessing done, I can see the texture of the paper of both the 2X samples and the native samples! Both are very sharp and very little difference between them. There is FAR more softness in the chart than the differences between the 2X results by themselves. It is more misleading than anything else because of it, we are far exceeding the resolution and sharpness of the chart with both.

In conclusion the 2X is FAR better than cropping and less than ideal test chart adds a few challenges to get a clear result.

Native
https://community.the-digital-pictur...0&d=1646665979

2X
https://community.the-digital-pictur...9&d=1646665962

As you can see the 2X after basic post processing is looking far better than previous results, I think what was happing was there was MUCH to much luminance noise reduction and eliminating the fine details of the paper. And having the WB corrected makes it much easier to view sharpness, for me anyway.

If this was done with ISO 100, good lighting (This lighting is to dark the exposure and enhancing chrominance noise, influencing the results.) and daylight or strobes. So idealy you have zero NR and you are seeing only the sharpness of the lens.

Also the chart is not very fine at all, a much larger and higher quality chart would make a HUGE difference in seeing the actual sharpness of the lens. Right now I'm using the texture of the paper as a means to help measure sharpness lol. That's pretty darned sharp at that point. But if you even just look at the chart itself and the lines in particular, it is definitely pulling much more detail compared to the native results. And using the full sensor in the process.

Another comparison to look at, do the same with the 1.4x.
It holds up much better, see what your opinion is there.
After I processed I had similar results and like I said in my last post it is better than the cropped native. I was referring to the 1.4x as being very similar.

We are seeing the same thing.

If I could get a chart locally available I might have grabbed one. In the past I used my pixma printer to make the chart on photo paper, it would have been better. I probably should just buy a chart. I found several online that were larger over $300. What I really need to do is replace my printer.

For the ISO 100, for my use it is very rare that I get to shoot at that low of an ISO on that lens. Would it really been representative of real world, for some people yes, but not me.

You really need to be using low ISO for future comparisons. If you want to try with higher ones too that's great. Just to see how it affects things. But low ISO stuff to single out sharpness is important.

But the 2x is pulling far more detail overall all things considered. I didn't bother with the 1.4x results because I already new they looked fine from the last time and looking at them in RAW just confirmed it. The native and 2x are both looking nearly the same sharpness wise. Both are pulling great sharpness, if both are pulling in the texture of the paper that's pretty respectable resolution. The 1.4x is just redundant really when the 2x is looking nearly as sharp as a native lens. It just be splitting hairs. Main reason to use it is just for the correct focal length for the job and of course one extra stop if light.

I'm sure at ISO 100 we would reveal more differences. But at ISO 1000. It's looking very good.

But for kicks I can process the 1.4x results.

But for me low ISO shots, like 100 or 200 is not uncommon for me. Certainly under 400. I do my fair share of very low light. But I have found it's usually better to try and avoid those situations in the first place. Sometimes just because you can doesn't mean it's best.

Most of the shots I take with a long lens will be Morning or Evening, in addition usually 1/500 or higher. Often time it is in the shaded woods. It seems I am struggling with the higher ISO more than enjoying ISO 100 or 200. It seems I am always stuck in the low light situation.

I need to go ahead and either get a chart or a decent printer replacement. I need to figure out why the EF 1.4x looked like crap. (dirty or it has problems)

Thanks for all your thoughts on this. It is helpful.

For sure. I understand what you are saying. But I think the premise of gaining light is where you are confusing things. Cropping doesn't lose exposure but raises noise to the point where using a higher ISO would achieve the same end result noise wise.

I think it is useful to know what it does with higher ISO's. And I agree it is 100% relevant. You'll know how far you can push your settings.

I like to isolate things like sharpness in this instance and let ISO fall where it may. Whether it is ISO 12800 or 400. I'll just know for myself what ever ISO I am using, I'm still pulling the most detail I can. Because if we shoot even higher ISO's, we can render these differences moot. So while it is useful to know what it does at higher ISO's. It doesn't replace or very good at figuring out is it better to use an extender or not from a detail resolving standpoint.

Basause if it is sharper/pulling in more overall detail with a 2x, there is no reason to crop. Because if I crop by 2x, I'm gonna reveal the same amount in noise as if I used an extender and used 2 stops higher ISO. At least in theory. There could be other sensor tech stuff happening that I don't know about. I am no Daniel Browning. Haha.

But if cropping is actually sharper, then regardless of the ISO it will be the better option. But so far it seems that the 2x is still relevant.

When I get home, just for giggles I'll do the same post processing with the 1.4x and covert to Jpeg. The RAW preview doesn't tell all.

Sorry if I am sounding like a broken record player, just tell me to stop if I do. Lol.

Ok, I'll stop now.

But I'll post for giggles the 1.4x results and leave it at that so all can see.:D

Not at all.

I was in the process of explaining what I meant by loosing light. I think it is just a bad description on my part. What I refer to is when the ISO goes up, to keep it down the shutter speed must be slower. Early morning and evening this is always an issue. It "feels" like I am loosing light but really it is just pushing down the shutter speed.

I hear ya, my minimum all depends on what I'm shooting. if I am hurting for light I'll go as low as 1/160. It just means I can't stop motion. But when birds are in between movements. They are often times stock still for a brief moment. And I'll raise my shutter speed when I want to capture motion.

So I'm not cranking it to the moon the ISO at least as much.

I used to do lots of lower light stuff, but lately I have tried as much as possible to hit that sweet spot in the evening where I get a good amount of light but birds are still active. Same in the morning.

Or depending on my birding locations they will be there throughout the day. So I can have my pick of lighting. This is ideal. I can be using ISO 100 or 200 and get what ever shutter speed I need. Maybe little bit more if I need higher shutter speeds.

Of course, this is just me. And not everyone is me.