Fast apertures on the GH5 = Full frame

[Dude_ger]

I don t get the point, you always can tell which picture is fullframe and which is apsc or mft or whatever.  It s not about the light, it s the look, and this look you can t replicate using faster lenses.

[Deadcode]

41 minutes ago, Dude_ger said:

I don t get the point, you always can tell which picture is fullframe and which is apsc or mft or whatever.  It s not about the light, it s the look, and this look you can t replicate using faster lenses.

i dont agree, think about speedboosters. A Canon 50mm F1.2 L will give the same look on A7S2 and on a6500 + Metabones Speedbooster. Or even on GH5s + BMPCC speedbooster.

 

[Timotheus]

2 hours ago, Dude_ger said:

I don t get the point

Agreed ;-)

You're saying you can tell the camera's apart that took the two pictures from Andrew's original post? Nah man.

[fuzzynormal]

9 hours ago, Dude_ger said:

I don t get the point, you always can tell which picture is fullframe and which is apsc or mft or whatever.  It s not about the light, it s the look, and this look you can t replicate using faster lenses.

I get the "false-equivalency" argument. 

I shoot M43 all the time.  And, if you look reeeeeely hard enough, I do think you can tell a difference, in aggregate, between M43 and FF.  The telltales of FF shooting can be evident if you know where to look.  For instance,  usually a smoother bokeh and sharper focus with shallow DOF.  A f2 through modern glass on a FF sensor, for example, can look "cleaner" in a way that M43@f.95 does not.  

These advantages, I think, are nit-picky though.  The differences are subtle.  It's often really hard to tell!  And if a skillful someone is shooting m43 with fast glass, you're probably not going to know the difference or pay attention enough to care in the first place.

I'm shooting a doc right now with the Voightlanders on the GH5 and I'd put it up against anything I've shot with my FF Canon equipment.  If you're shooting M43 or FF, your viewer doesn't really care.  Are you a good shooter?  That's what matters.

All that said, and like I've mentioned before, I still prefer to shoot FF on a fast 50mm prime for interviews.  It's just easier to set up a shot 'kuz of space and lighting --and you can really throw the background out of focus with a f1.2.  That forgives a lot of sins in corporate locations/environments.  Hey, I'm a practical guy.  OTOH, when shooting interviews with my M43 gear I tend to use my 42.5mm lens @f1.2 --and the DOF is wildly shallow and intense too.  The flatness of the "portrait" FOV is a nice look unto itself.  One just needs more physical space in the room to make that happen.  Also, you end up farther away from the interview subject, which can impair any intimacy if you're trying to create such a thing. 

My decisions tend to be less about opticals and much more about other considerations.  Freeing your self from the dogmas of "this vs. that" with gear is a big step to make.  I encourage everyone to make it ASAP. 

[Castorp]

On 22/05/2018 at 9:19 PM, Don Kotlos said:

I don't find it that great.

First of all because especially cinematographers need to know and adjust the DoF, FoV and perspective. That means then need to know the equivalent focal length for all different cameras with different sensor sizes. If you had a m4/3 camera with the 25mm @f/1.4 and you wanted to match the image with a FF camera then you would have to use a 50mm @f/2.8. You then use ISO/NDs/Lights to control exposure. 

 Second is because he was ad nauseam saying you cannot get the same look from different sensor sizes no matter what, even when people provided him with proof. He then he made a youtube video twisting the argument and saying exposure is everything. 

Lets not go back to that bitterness. 

Hopefully this is the last thread about it but lets say it one more time: Perspective only depends on the distance from the lens. You can then adjust the focal length/aperture for a given sensor size to get a specific FoV and DoF. 

Firstly, I don't care what kind of social games are happening here. I think it is a great video. He is wrong in the thread you link but so what? Perhaps he learned something from that thread? I know I am wrong all the time, and I often make mistakes. In the video I posted he comes to the same important point that you re-iterate - that perspective depends on camera position to subject, nothing else. The more important point is that people shouldn't worry so much about it (if they're not a DP or something highly unusual like that).

The reason I posted the video is that even here, in a geeky forum (forgive me), some are confused about perspective. Why is that? Partly, I believe, there's an overly complicated, yet too simplistic, narrative around lenses and formats. I found the environment I learned way more pedagogic. "150 is normal on large format, 80 is normal on medium format and 50 is normal on small format". 

Of course there was plenty of elitist BS around format back then as there is now, but the whole "full frame" (what a ridiculous term) hype is pretty funny when you remember a couple of decades ago when small format was laughed at in some circles. That didn't stop artists from making amazing work with 135. Just as it wouldn't stop anyone now making amazing work with m43 (and the difference between 135 and, say, 6x7, was order of magnitude larger than the difference between popular digital formats which is truly negligible). 

Lenses and formats aren't that formulaic. The quality of a de-focused area in a picture isn't just about focal length and aperture. I understand most here are well aware of that. Yet we constantly plow through things like "that lens equals that f stop on this format, therefore that lens needs this f stop for that bokeh" blah blah . It's a shallow spec sheet oriented discussion. I encounter many who are worried and confused about this when there is no need to be confused at all. With experience you see that some lenses look way nicer than other lenses, no matter what the f-stop is. Some f2.8 lenses have way nicer de-focus than some f1.8 lenses. Again, I know you all know this but I think it's important to remember that image quality can't be summed up in some formula. There's more than that to it. 

[Trek of Joy]

9 hours ago, Dude_ger said:

I don t get the point, you always can tell which picture is fullframe and which is apsc or mft or whatever.  It s not about the light, it s the look, and this look you can t replicate using faster lenses.

This is not true, and it would be easy to find samples that disprove what you're saying. There is no such thing as the "FF look". 

Chris

[Castorp]

On 23/05/2018 at 12:32 AM, Damphousse said:

Okay.  This is what you said...

All the Nikon and Canon 50mm 1.8 lenses I'm familiar with are around $100 or less!  I just don't understand this apples to oranges comparison.  You bring me a 50mm 1.2 Canon lens and yes it will be $1,100 but now it is capable of bokeh that the Olympus can't do.  So either way you are wrong.

 

I think dismissing the inherent barrel distortion, etc with wide angle lenses particularly when you don't know the exact specific characteristics of each lens in the test is a bit disingenuous.  Everything you are seeing in those pictures is not 100% due to distance.  The exact ratio of what is affecting what we can debate but just glossing over a fact that doesn't support your argument isn't the way to help someone learn.

I also stated that M43 has its place.  I own a BMPCC and it is super 16.  There are pros and cons to the camera.  If someone asks me I tell them the truth.  All these things are just tools.  They all have their place.

No.  You are wrong, at least about OEM M43 lenses.

It might be easier to make a smaller lens but that is not the only factor in the equation.  Four things...

1)  It is easier to make a smaller lens.  POSITIVE

2)  To get the same field of view you have to use substantial wider angle lens.  NEGATIVE

3)  To get the same level of bokeh you have to use a larger aperture.   NEGATIVE

4)  ALL nonCanikon lenses are produced at lower volume (excludes smartphones) and do not have the same mass production cost benefits.  NEGATIVE

So one positive and three negatives.

Also there is something you are being very disingenuous about.  M43 OEM lens manufactures use software interpolation to correct defects in their lenses.  In many cases they make optically inferior lenses and use in camera software to do the last bit of heavy lifting.  One can argue about how much difference there is between the software corrected and uncorrected images but full frame lens manufactures going the extra mile to correct as much as possible with glass needs to be acknowledged.  People need to tell the whole truth and do apples to apples comparisons.

Cutting corners to hit a particular price target and then relying on a software crutch is fine but don't turn around and say a fully optically corrected full frame lens is some kind of rip off.

When you see lists of the sharpest lenses there are two things you don't see M43 lenses and a bunch of lenses wider than 35mm.   Ever ask yourself why?

Tbh I'm not following your post. Feels like you're putting words in my mouth? The point I was making is that when the PRO Oly lenses came out people where comparing to 200$ Canon/Nikon lenses saying it was expensive. I don't think it's expensive at all when you look at the quality of image those things produce. 

I don't own a m43 system. I comment on what I have seen and the images from those lenses are stellar. There are few 135 lenses at similar or less cost I have seen that compare. The closest one I can think of is the 58mm f1.4G Nikkor which makes beautiful images. It is also more expensive. The Oly is pin sharp at 1.2 and way nicer than the Oly 1.8 when both are at max aperture. The same can not be said for the Canon f1.2 compared to it's 1.8 sibling at max aperture. I couldn't care less about the Canon bokeh when the Oly lens is so much nicer in all other aspects of rendering. There's a few Nikon lenses that do it and those huge Zeiss lenses. All more expensive. That's what was trying to make a point about - in general, (perhaps with some exceptions,) lenses are more expensive the larger the format for similar quality 

I'm guessing we shall see bokeh be much less written about as soon as computational imaging makes amazing bokeh possible on smartphones et c. Software interpolation and digital image processing is a great thing. I don't see any problems with it whatsoever in digital photography. If I want an analogue image, with analogue corrections, I'd much rather shoot film. To argue for analogue corrections on a digital medium makes zero sense to me. I think some of the photo industry have one leg stuck in the old field. Like how the first cars where half horse carriages. 

I have a particular dislike to format snobbery. Perhaps it is rooted in old university times when large format was the only true format and nothing else counted. Breaks my heart to hear beginners think that they need this or that format to make good work. They're indeed all tools. The difference in image quality between todays digital formats is much smaller than it was between yesterdays film formats.

The point with Andrew's  original post was, as I understood it, to say that all formats are legit. They're all great for different reasons and they all have their specific advantages and disadvantages.  

On 23/05/2018 at 2:46 AM, noone said:

You are missing the point.

The Sony Zeiss 55 1.8 IS a premium lens...

I don't think I ever discussed Sony lenses. I have very limited experience with them and if you say it's nice I'm sure it's nice. Personally Sony cameras and lenses leave me stone cold. I'm aware that's entirely subjective and I'm sure they're great. It's just that I don't care at all. I guess I'm romantic like that - there has to be something in the image or tool that catches my imagination. Apologies. 

[noone]

3 hours ago, Castorp said:

Tbh I'm not following your post. Feels like you're putting words in my mouth? The point I was making is that when the PRO Oly lenses came out people where comparing to 200$ Canon/Nikon lenses saying it was expensive. I don't think it's expensive at all when you look at the quality of image those things produce. 

I don't own a m43 system. I comment on what I have seen and the images from those lenses are stellar. There are few 135 lenses at similar or less cost I have seen that compare. The closest one I can think of is the 58mm f1.4G Nikkor which makes beautiful images. It is also more expensive. The Oly is pin sharp at 1.2 and way nicer than the Oly 1.8 when both are at max aperture. The same can not be said for the Canon f1.2 compared to it's 1.8 sibling at max aperture. I couldn't care less about the Canon bokeh when the Oly lens is so much nicer in all other aspects of rendering. There's a few Nikon lenses that do it and those huge Zeiss lenses. All more expensive. That's what was trying to make a point about - in general, (perhaps with some exceptions,) lenses are more expensive the larger the format for similar quality 

 

 

I don't think I ever discussed Sony lenses. I have very limited experience with them and if you say it's nice I'm sure it's nice. Personally Sony cameras and lenses leave me stone cold. I'm aware that's entirely subjective and I'm sure they're great. It's just that I don't care at all. I guess I'm romantic like that - there has to be something in the image or tool that catches my imagination. Apologies. 

Personally, I think some of those cheap $200 lenses on larger formats DO compare with the 25 1.2.

 

That Nikon 58 1.4 isn't a lens that had sharpness as its first objective I think and is an acquired taste that many love and many don't.

 

I mentioned the Sony Zeiss 55 1.8 first as it is a 135 format lens (fact) and it is as nice as the Oly 25 (subjective) and it is cheaper (fact).

I have owned a few M43 lenses and liked them but I preferred having a couple of those, a couple of Sony lenses and using Canon lenses adapted as I used them on both Sony and M43 (and some even on Canon briefly too).

I use whatever lens works for me from any maker on the cameras I am using at the time.     What is cheaper and what can be compared IS subjective and also varies from lens to lens and you can not say it applies to all.

 

[nigelbb]

On 22/05/2018 at 1:20 AM, Cliff Totten said:

Admittedly,...this is going to sound really stupid. However, I think the word "Micro" was a bad choice for the title of this system. People are OK with a 4:3 aspect ratio but nobody wants a "Micro" sensor. All it does is remind people every time it is said that it's a "small" sensor.

Imagine if it was named the "Super 4/3" system"? "Super" 4/3 would have diverted attention away from its size and "psychologically" is a much better marketing name.

Yeah....I know this sounds stupid but sadly, I still think this idea holds true.

Super 4/3....much better "ring" to that name.

 

You need to understand the history. Originally Olympus & Kodak introduced the Four Thirds system which was for DSLRs with a mirror box https://en.wikipedia.org/wiki/Four_Thirds_system. Subsequently Olympus & Panasonic introduced the Micro Four Thirds system using the same size sensor but for mirrorless cameras. It was called "Micro" because the cameras & lenses were significantly smaller than those from the older Four Thirds system https://en.wikipedia.org/wiki/Micro_Four_Thirds_system.

[Tiago Rosa-Rosso]

 

[tupp]

From this thread's excellent EOSD article:

Quote

Since the moment they were born, the Panasonic GH5 and other Micro Four Thirds cameras have suffered from a misconception on the specs sheet created by the smaller sensor. This misconception is that they can never replicate the look of a full frame camera. It’s not true and the pictures above prove it.

This paragraph suggests that equivalence principle is completely true -- in other words, that there is no difference between the looks from different sized formats (more importantly, no difference in the looks from the optics designed for different sized formats).

 

 

From the most recent excellent EOSD article on the Fuji GFX 50S:

Quote

The great advantage of medium format though is that you can maintain a very beautiful shallow depth of field without using such mushy wide apertures. You can stop down your adapted F1.4 DSLR lenses to F2.8 for higher resolving power and still enjoy the beautiful rendering and three dimensional depth of field of a very wide aperture on full frame.

This passage asserts that there is a difference in looks between formats, due to the "mushiness" inherent in having to use larger apertures with smaller format lenses, in order to get the same mathematical DOF as that of lenses made for larger formats.  In other words, the equivalence principle is not valid, and optics for different sized formats yield differing looks.

 

It seems confusing...

 

 

On 5/25/2018 at 7:09 AM, Timotheus said:

Agreed ;-)

You're saying you can tell the camera's apart that took the two pictures from Andrew's original post? Nah man.

The differences between the two images are significant.  I can tell them apart, and it appears that another poster in this thread can also discern the differences.

[Deadcode]

10 hours ago, tupp said:

The differences between the two images are significant.  I can tell them apart, and it appears that another poster in this thread can also discern the differences.

Bravissimo brotha', you can tell a difference between a lens manufactured by Sony and a lens manufactured by Voigtlaender. Sick skillZ!

If he use the same lens with focal reducer you will not see any difference between the two images. I have tried it.

[tupp]

4 hours ago, Deadcode said:

Bravissimo brotha', you can tell a difference between a lens manufactured by Sony and a lens manufactured by Voigtlaender. Sick skillZ!

Marvelous mister,  you completely missed the quote in my message which showed that I was merely responding to another poster who suggested that one could not "tell the cameras apart."  Sm-o-o-o-o-o-th!

 

The striking differences between the two images do not involve the character of particular lens brands -- the discrepancies shown result from fundamental differences in DOF/focus.

 

Actually, I am not sure that I could say which lens was a Voigtlaender or a Sony, because I am not familiar with either brand.  On the other hand, I would guess that the Sony lens is a FF lens and that the Voigtlaender is made for 4/3.  If so, the photo with shallower DOF is probably from the larger format (Sony?) lens, because such has been the case in every equivalency test so far (except for a peculiar one linked earlier in this thread).

 

A lot of folks in this forum insists that the equivalency principle is absolute --  that lenses made for differing formats will look exactly the same, given the proper apertures and focal lengths.  Evidently, you disagree that the equivalency principle is  absolute, as your sarcasm indicates that the difference is obvious between these two lenses which are designed for different formats.

 

 

4 hours ago, Deadcode said:

If he use the same lens with focal reducer you will not see any difference between the two images. I have tried it.

I mostly agree, but that point is irrelevant to whether or not the equivalency principle is valid.

[no_connection]

It is absolute. But for some reason lens manufacturers don't seem to follow the theoretical magic lens that was used to make it.

And since lenses are made with other aberrations in mind large aperture wide angle lenses will always struggle for small sensors. Try make a lens for a phone sensor with the same DoF as a FF 50mm f1.4. You might argue that it's an extreme example but does show there is a upper limit (or lower) so it has to follow a line or curve of some sort. And when pushing that extreme FF will always win out in that regard. Does it matter when you stay inside the extreme? No, the equivalency is happy to work just fine when not pushed to the edge of oblivion.

[tupp]

@no_connection

The equivalency principle is not absolute if it doesn't work in every instance.  You seem to agree that it does not work in every instance.

 

So far, almost all equivalency tests show significant differences between the look from lenses designed for different formats, with the results mostly showing shallower DOF and less mushy focus falloff with larger format lenses.  None of these tests are "pushing the extreme."

 

There exist variables other than focal length and aperture (and aberration) that affect DOF and focus falloff.  You and others in this thread touched on one such variable -- there is a limit to the amount of optical resolution that can be crammed into an image circle for a tiny sensor.  Like many physical properties, this variable probably doesn't have a hard threshold, but instead is a matter of degree, with the image slowly degrading as the image circle decreases.   There are plenty of other possible variables, too.

 

If you think that the only variables that matter are the ones expressed in the equivalency principle (focal length and aperture) and aberration, merely consider apodization optics.  Here are two photos taken with two Fuji lenses of the same focal length and set to the same aperture, with identical optical elements, except one of the lenses includes an apodization filter:

Note the discrepancy in the softness of the distant background (click on the photos to enlarge them).  Every variable is identical between these two shots, except that one lens contains a filter, and yet there is an obvious difference in the DOF.

 

The equivalency formula does not take into account all of the variables that affect focus and DOF, therefore it is not absolute.

[Deadcode]

1 hour ago, tupp said:

The equivalency formula does not take into account all of the variables that affect focus and DOF, therefore it is not absolute.

I still dont understand what you want to prove.

Same lens with focal reducer on smaller sensor will give the same look. 

Different but faster lens will not give you same look but not because of the sensor, but the characteristics of the lens.

Focus fall-off is lens dependant. Helios 44m-4 will give more creamy blur than Zeiss 55 1.8. Panasonic 25 F1.4 will never be as creamy as Carl Zeiss Jenna Tessar 2.8/50. But not because of the sensor size... SLR Magic 25 F0.95 will be smoother than Canon 50 1.8 STM... it's all about lens characteristics.

 

[BTM_Pix]

52 minutes ago, Deadcode said:

Same lens with focal reducer on smaller sensor will give the same look. 

This is the test I did for the 2017 edition of this thread

Nikon 70-200mm f2.8 on a full frame (Nikon Df), a crop frame (Nikon D500) and an MFT (Panasonic GX80).

Shot at 70mm f2.8 on the crop, 100mm f4.0 on the FF and 70mm f2.8 on 0.7x Speedbooster on the MFT.

These images show an identical equivalent depth of dirt on my wall.

 

[Timotheus]

On 5/29/2018 at 9:21 AM, tupp said:

The differences between the two images are significant.  I can tell them apart,

Claiming to be able to tell them apart, with the actual lens and camera data provided is bit easy, no?

Everybody agrees there can be subtle differences between different lenses, like in the example of Andrew, and also the one you provided (noting that adding an APO filter is quite the change to a lens!).

Let's keep it simple and general: using equivalence math, you can pretty precisely predict and match the results of certain lenses on certain sesnsor-sized camera's. That's the essence. For evidence: see @BTM_Pix above (and countless others).

I feel this point gets unnecessarily muddied when focusing on all kinds of specific details of rendering of a specific lens.

[no_connection]

8 hours ago, tupp said:

If you think that the only variables that matter are the ones expressed in the equivalency principle (focal length and aperture) and aberration, merely consider apodization optics. 

Look, if you flip the terms of the equation the only thing equivalency will tell you is how to get the same diameter lens with different focal lengths. It's simply a number that relates things to each other and make them comparable.
You are 100% correct in that I think it's not that relevant and as you have shown lens construction or "features" is more important. However the lenses you showed do NOT have the same aperture which is evident in the upper right corner, I would suspect that f4 is the "mean" aperture of the apodization filter which makes sense as you loose light to the darkened edges of the filter. I do like the idea of such a filter and want to try to make one for my 50mm f1.4 that have aperture to "spare" so to speak.

For what the equivalency formula is it holds true, but at the same time you have to understand what it does not do.

 

[tupp]

 

23 hours ago, Deadcode said:

On 5/29/2018 at 11:24 PM, tupp said:

The equivalency formula does not take into account all of the variables that affect focus and DOF, therefore it is not absolute.

I still dont understand what you want to prove. 

Well, I have demonstrated that the equivalency principle doesn't take into account all of the variables that affect DOF and focus, therefore it is not absolute. ?

 

 

23 hours ago, Deadcode said:

Same lens with focal reducer on smaller sensor will give the same look. 

Yep.  I acknowledged that very point above.  However, that point is irrelevant to the DOF equivalence discussion, and the validity of equivalence is the primary gist of OP's article from which this entire thread follows

 

23 hours ago, Deadcode said:

Different but faster lens will not give you same look but not because of the sensor, but the characteristics of the lens.

Focus fall-off is lens dependant. Helios 44m-4 will give more creamy blur than Zeiss 55 1.8. Panasonic 25 F1.4 will never be as creamy as Carl Zeiss Jenna Tessar 2.8/50. But not because of the sensor size... SLR Magic 25 F0.95 will be smoother than Canon 50 1.8 STM... it's all about lens characteristics.

I mostly agree, but I would add that images from lenses designed for larger formats often look different than images from lenses designed for smaller formats using "equivalent" settings.

 

 

22 hours ago, Timotheus said:

Claiming to be able to tell them apart, with the actual lens and camera data provided is bit easy, no?

The lens for the larger format has strikingly shallower DOF.  It's amazing that some can't see that.

 

 

22 hours ago, Timotheus said:

Everybody agrees there can be subtle differences between different lenses, like in the example of Andrew,

Again, the differences between the two images is striking.

 

Also, it is often the subtleties that make the difference in imaging.

 

22 hours ago, Timotheus said:

the one you provided (noting that adding an APO filter is quite the change to a lens!).

In regards to the lenses of the two images in question:

• the focal length is an exact match;
• the aperture is an exact match;
• the refractive optical elements are identical;
• the only difference is that one lens has a non-refractive filter.

How can you reconcile the equivalency principle with such a dramatic difference in DOF, when the aperture and focal length are the same? 

 

 

22 hours ago, Timotheus said:

 using equivalence math, you can pretty precisely predict and match the results of certain lenses on certain sesnsor-sized camera's.

Please show how the equivalence math accounts for the dramatic DOF difference in the apodization example that I posted above.

 

 

22 hours ago, Timotheus said:

For evidence: see @BTM_Pix above (and countless others). 

So, you maintain that  an equivalence comparison in which all images were shot with the same zoom lens proves that there is no difference in look/DOF between optics designed to cover different sized formats?  Doesn't one need to compare optics designed to cover different sized formats to make a valid conclusion on whether or not formats have a certain look?

 

In regards to the "countless other" equivalency tests, most of them have been discussed on this forum, with the equivalency "absolutionists" similarly glossing over little "details," such as  comparisons done with the same zoom lens.  In almost every such test that used lenses made for different formats (except for a peculiar one linked earlier in this thread), the larger format lens always has shallower DOF.  In addition, another common flaw in equivalency tests is that there is nothing but air between the foreground and background, so there is no way to tell how the DOF rolls off.

 

 

16 hours ago, no_connection said:

However the lenses you showed do NOT have the same aperture which is evident in the upper right corner, I would suspect that f4 is the "mean" aperture of the apodization filter which makes sense as you loose light to the darkened edges of the filter.

No.  The whole point of that apodization demonstration is that the aperture and focal length are EXACTLY identical -- but the DOF is very different.  If you click-to-enlarge the photos, you will see considerably more DOF differences than those shown in just the upper right corner.

 

By the way, that apodization example originated in this article.

 

 

16 hours ago, no_connection said:

I do like the idea of such a filter and want to try to make one for my 50mm f1.4 that have aperture to "spare" so to speak. 

I think that apodization filters are always internal in a lens.  I believe that Fuji had a lens in which different apodization filters could be inserted.

 

 

16 hours ago, no_connection said:

For what the equivalency formula is it holds true, but at the same time you have to understand what it does not do. 

The formula only gives the "mathematical" DOF.   It does not account for DOF rolloff nor for the mushiness at wider apertures with smaller format lenses (nor for apodization).