1.33x factor: is it really an issue?

[RichardJ]

Quote ..... "For those who are all so concerned about DoF and the "look" of this lens or that lens - here's a portrait. Can you identify which focal length, maximum aperture, and shooting aperture were used?".......


Andy, That's easy: 50mm, f2.8 and film (!!?)

I understand that your question was rhetorical and I agree with your sentiments. Why the fuss? We all know that film produces beautifull images but digital capture is the future, and the advent of the digital rf age does not suddenly steal from us the freedom of choice. If you prefer life where lenses behave "normally" then, fine, remain with film and a 24 x 36mm format.



We all have a favourite prime lens (28, 35 or 50) so, with the advent of DRF we have four options:


A) Buy a new prime lens (ie your favourite divided by 1.33)

B) Take a few steps back

C) Stick to film,or

D) Wait (years probably) for a full frame sensor.

DSLR users are light years ahead of us; I'm sure they are not having these kind of debates anymore. They're probably too busy, out taking photographs.

Yours, too exhausted waiting for the "full frame" techno thingy and too excited by the (RF) future

Richard

(ps. I met a pro commercial photog last week who told me that none of his colleagues (as far as he was aware) used film any more. This was not, in his opinion, a choice relating to speed of post production nor a desire to "keep up" with changing times or editorial needs but merely a choice based upon preference. Even though his opinion is not necessarily a representative sample I was, nevertheless, slightly taken aback.)

 

[40oz]

So does this mean that the numbers on the lens for DOF will be inaccurate?

Not as far as I know. Consider that the lens focuses the image independent of the size of the film or sensor. The issue would be that someone used to a certain depth of field at a given aperture with a 50mm on film will be using a 35mm as a digital "normal," and the depth of field will be greater. Thus making it harder to isolate a face in a crowd, for instance.

It's basically a half-frame-type camera, with all the obvious drawbacks of moving to a smaller format. Contrary to (what seems to be) popular belief, going digital doesn't negate the quality hit you take when you decrease the size of the captured area. There's a limit to what a lens can resolve. The film or sensor has no affect on this limit, it just records what detail the lens is capable of transmitting. If you currently run a half-frame mask in your M, you'll be right at home with what the camera can do. It's extrememly unlikely that anyone is going to find they get more detail out of the digital M than the same lens on a film body. Essentially, you are just going to blow up your shots more than you would with film.

 

[rvaubel]

So does this mean that the numbers on the lens for DOF will be inaccurate?

In general, for the same field of view, the DOF goes up a little. On the other hand, for the same lens, the DOF would be the same, the crop is just tighter. In any case the effect is pretty minimal for a 1.33X sensor. I notice it a little bit with a 1.6X sensor and compensate by opening up the aperture by one stop (if I want a thin DOF). For the M8, the difference will be to small to matter much

What is missing from the lens lineup is a fast, wide. Below 28mm, there is nothing faster than F2.8 . This is a problem that needs to be remedied. Someone need to introduce a 21mm/2.0

Rex

 

[jaapv]

Not as far as I know. Consider that the lens focuses the image independent of the size of the film or sensor. The issue would be that someone used to a certain depth of field at a given aperture with a 50mm on film will be using a 35mm as a digital "normal," and the depth of field will be greater. Thus making it harder to isolate a face in a crowd, for instance.

It's basically a half-frame-type camera, with all the obvious drawbacks of moving to a smaller format. Contrary to (what seems to be) popular belief, going digital doesn't negate the quality hit you take when you decrease the size of the captured area. There's a limit to what a lens can resolve. The film or sensor has no affect on this limit, it just records what detail the lens is capable of transmitting. If you currently run a half-frame mask in your M, you'll be right at home with what the camera can do. It's extrememly unlikely that anyone is going to find they get more detail out of the digital M than the same lens on a film body. Essentially, you are just going to blow up your shots more than you would with film.

Err.. This post disregards that Leica M lenses outresolve film or sensors considerably... The limiting factor is not the lens but the film. And if you do your math, the difference in DOF will be about 1/2 a stop for a 1.33 sensor, or, if you prefer, 2/3rds of a step in focal length. Not really enough to matter in real life. What will be noticable though is that DOF and Bokeh will be more defined on a sensor than on film. As another poster mentioned, DSLR users are beyond this discussion by several years.

BTW, Rex, wouldn't a 1.8 18 be nicer? It would be quite a feat to squeeze such a lens into a RF mount, but if anybody can do it, it would be Leica.

 

[rvaubel]

Err.. ...... Leica M lenses outresolve film or sensors considerably... The limiting factor is not the lens but the film.

That's what gives the M8 the real possibility of outresolving the 5D. Also, the M8 will not be hobbled by the presence of anti-aliasing filter and the attendant in camera sharpening software required to mitigate its blurring effect. Clean, sharp, unprocessed RAW files is a major benefit of the Leica approach.


Rex

 

[rvaubel]

.............. wouldn't a 1.8 18 be nicer? It would be quite a feat to squeeze such a lens into a RF mount, but if anybody can do it, it would be Leica.

jappv

Yes, a 18mm/f1.8 would be a perfect lense for the rangefinder existing light light guy. And it would be possible, but not easy or cheap. That's why I suggested a 21mm/2.0 I keep harping on the fact that Olympus already offered such a lens for their OM system over 30 years ago. Not only was it fast, it was also very compact, smaller than the Leica 21mm/2.8 ..... and it was a retro-focus design for christ sake!!

Come on Leica, Zeiss, and Cosina. You can do it!!

Rex

 

[40oz]

Err.. This post disregards that Leica M lenses outresolve film or sensors considerably... The limiting factor is not the lens but the film. And if you do your math, the difference in DOF will be about 1/2 a stop for a 1.33 sensor, or, if you prefer, 2/3rds of a step in focal length. Not really enough to matter in real life. What will be noticable though is that DOF and Bokeh will be more defined on a sensor than on film. As another poster mentioned, DSLR users are beyond this discussion by several years.

All I was trying to say was the lens focuses at the film plane, regardless of the size of the film format/sensor. What is or is not in focus is not determined by the recording medium. Maybe I'm missing something? Is the depth of field different on the edges of an image vs. the center?

I was also trying to indicate that a smaller format is not the direction most go in order to increase detail in an image, for good reason.

As far as lenses exceeding the capacity of film, well, yes they do. To exploit the maximum resolution the film (say, 100 Lp/mm for TMax100), apparently you need a lens capable of 3 to 4 times the resolution. Now, whether this math applies with digital sensors is a subject for someone else, but clearly you need enough pixels per millimeter to surpass the ability of film, which would be more than 200 pixels per mm. In an 18X27mm sensor like the M8, this would require a ~19.5 million pixel sensor. So obviously, no matter what lens you put in front of the sensor, you aren't going to surpass available films until you greatly increase the density of pixels in the sensor.

If you could increase the size of the sensor and project a larger image onto it, you could capture more detail from a scene, similar to going to a larger format film. For a given target, put more pixels under it than film can resolve line pairs, if you simply wish to surpass the theoretical limit of films. The lens is immaterial until you have a sensor capable of capturing the details.

This discussion is simply silly idle chatter, I'll happily admit 🙂

The bottom line is digital cameras take as good a picture as anything else. Some people prefer the look of certain mediums over others. It matters little which system is best at capturing test patterns. All that matters is what you do with it.

 

[thurows]

Cropping is only an issue if you make it one. If your stuck on the crop factor why not shoot a 6X7 rangefinder, a 40 is like a 20 on 35mm. If the camera make sense to you, digital without having to buy a bunch of new lenses in a compact package, buy it. If you're going to cry that your 20mm only covers 24mm then plunk down$7,000 for a Canon EOS-1DS Mark II which weighs 42.9 ounces for just the body. It's a tool, adapt and over come or wait for your film to get processed and printed and scanned before it gets to your computer.

 

[Matthew Runkel]

Of course crop factor is an issue. Whether it is a gating issue depends on personal preferences and resources. The original poster, who imagines "everyone" is weighing the purchase of a $5000 camera, does not consider the issue significant because all that is required to approximate his film setup is the purchase of two new Leica lenses. Others would find it a hardship to add two new lenses on top of an expensive body.

When the crop factor discussion occurs on SLR forums, most posters ignore that the standard range of prime lenses and their fields of view was not chosen at random. Leica's 1.33 crop factor honors the established sequence of fields of view, which is nice. Still, not everyone already owns the lenses whose cropped FOVs correspond to those of their preferred film lenses. Some people only own their favorite film lenses and, like the original poster, will need to think about additional purchases.

On SLR forums, a significant subset of users who favor long lenses rejoice about the crop factor extending the reach of everything. By contrast, very few rangefinder users would consider everything getting "longer" as an unmitigated good.

Although the 1.33 cushions the blow to the lens lineup by ratcheting existing lenses up to different recognizable FOVs, this mapping is imperfect. The 135mm lenses are out, too "long" for accurate focusing. And I'm pretty sure there are good answers to rhetorical questions like "why don't they just make a 24/1.4?" I believe that as focal length decreases, the difficulty of creating fast lenses increases greatly, which is why you just don't see things like 21/1.4 lenses. Probably it would be easier to create fast wide lenses with image circles that don't cover the 35mm frame, but that goes against Leica's philosophy that every lens should generally be usable with every body.

 

[Mark Norton]

Good comments; I've never understook why dSLR users get all enthusiastic about the increased reach of their long lenses; they gain nothing that they could not have got by cropping a full-frame image.

Right now, 1.33 is all we can have, if you want the image quality to be good, the camera to be compact, the battery life to be reasonable and the cost to be below stratospheric.

As regards lenses, it would be nice to have a faster wide-angle, but as Leica prices show, going faster tends to be more expensive than going wider. I'm sure we'd like a 28mm f1.4 or a 24mm f2, but I think they would be expensive. In the dSLR world, Nikon have a 28mmf1.4, a bit long in the tooth now, which is big, heavy and expensive.

 

[rvaubel]

....... all that is required to approximate his film setup is the purchase of two new Leica lenses. Others would find it a hardship to add two new lenses on top of an expensive................. And I'm pretty sure there are good answers to rhetorical questions like "why don't they just make a 24/1.4?" I believe that as focal length decreases, the difficulty of creating fast lenses increases greatly, which is why you just don't see things like 21/1.4 lenses. Probably it would be easier to create fast wide lenses with image circles that don't cover the 35mm frame, but that goes against Leica's philosophy that every lens should generally be usable with every body.

Yes, you hit the real dilema for the new 1.33x sensor user. There is a definite "lens gap" in the wide, fast range. And this is supposed to be the heart of rangefinder territory. However, you are going a little far in suggesting that a 21mm/1.4 would be needed to fill a gap in the film range. After all, the fastest 28mm (film equvalent of 21mm) is Leica's f2.0 and Cosina's F1.9 Not that I wouldn't luuuv a 21mm/1.4!!

The other "gap" is any affordably lens to replace a 21mm?2.8 focal length. About the only thing out there is the 15mm heliar. Its a nice lens but very slow. At least its cheap! Well I guess you could always spring 4K for the Zeiss 15mm/2.8😱

I think the key is going to be your last point. That is to create a fast wide lens that just covers the 1.33X sensor. That would be make the design of a 18mm/2.0 a lot easier, compact and affordable. However, the film people would come out of their tree.

Rex
I'm a film guy too but really wouldn't need a 18mm/2.0 for FF. The Heliar is OK for that role.

 

[rvaubel]

As regards lenses, it would be nice to have a faster wide-angle, but as Leica prices show, going faster tends to be more expensive than going wider. I'm sure we'd like a 28mm f1.4 or a 24mm f2, but I think they would be expensive. In the dSLR world, Nikon have a 28mmf1.4, a bit long in the tooth now, which is big, heavy and expensive.

Mark

Actually, the Canon is a 24mm F1.4 . I have one and, let me tell you, it is big and heavy. But it is also one of the most incrediable low light experience you can have. On my 20D it is equivalent to the legendary 35mm/1.4 But on the 5D it has no equivalent. Now if Leica would just come up with a 18mm f1.4 🙂 !!!

Rex

 

[Olsen]

- I would think that 1,33 crop factor will be less of an negative issue on a Leica Rangefinder. My first reason to have a FF-DSLR (1Ds II) is the viewfinder which is the largest and brightest in the business. I can't use crop factor SLR's if sharp pictures are to be expected of me. Their viewfinder is all too small and dark for my aging eyes.

When shooting with the 1Ds II and extreme wide angle, I stop the lense down to get rid of soft corners and light fall off. So, a 'fast' wide angle on a M8 is not on my wish list. You will soon find out that it is easer and better to compensate by turning the ISO knob,- a mind boggling thing for anyone who have done photography for 30 years. - I will never be spoiled by this opulent luxary.

The 1Ds II produces some awsome high quality - noice free - pictures on high ISO that has changed my photography totally. In the analogue world I was a '100ISO guy'. Not anymore. Now I am not afraid of using 1600ISO at available light.

Super fast super wides for available light photography? Forget it! They will produce some awfull pictures on a digital camera with curved lines and dark corners. It is far more important how M8 is performing regarding noice on high ISO and how it can be corrected. If at all.

'That is the question', to site Hamlet.

To what I have heard is that M8 is far from 1Ds II's noice free high ISO performance in this respect. That is far more worrysome, - and dissappointing. I am looking forward to see sample pictures of M8's performance in this respect. There will soon be a call for a M9 with, not neccessarily more pixels, but lower high ISO noice.

Further; I will go for the new zoom 16-18-12 mm 4,0 - which I, most likely, will stop down even further, to get rid of light fall off, which I will compensate will higher ISO. Most likely, this lense is probably one of the poorest on the MTF chart bench of the Leica range, but that does not matter. The sensor will not outperform the lense anyway. Far from it.

- I am more conserned on how much distortion it produces on the M8 and how it can be corrected. Has anyone seen any sample pictures from this 12-16-18 mm 4,0?

 

[jaapv]

I think you might be pleasantly surprised, as the issue of light fall-off is exactly what Leica adressed. And geometrical distortion? That doesn't even begin to be an issue with Leica wide-angles. Forget about your Canon stuff. They make very good tele lenses, but their wide-angles seem to be designed by the Coca-Cola company, in comparison to Leica and Zeiss. As for ISO performance, yes, the Canon files will be more noise-free at high ISO. But you will be able to hand-hold your M8 at least two to three shutterspeeds slower than your bulky Canon, gaining you two or three ISO steps. And if you go a bit longer, what about the Noctilux?

 

[John Camp]

Good comments; I've never understook why dSLR users get all enthusiastic about the increased reach of their long lenses; they gain nothing that they could not have got by cropping a full-frame image.

They get more pixels and better resolution. If you take a FF Canon 1Ds2 200mm and crop it to a 300mm FOV, you have fewer pixels across the FOV than you do with a 200mm Nikon D2x (which gives you an effective 300mm FOV), even though the D2x has fewer total pixels. If you use a 300mm Canon to put all available pixels across the FOV, the lens costs more , is heavier and is usually slower.

JC

 

[photogdave]

Aha! But the Canon's pixels are bigger with deeper pixel wells!

...I could really care less!

 

[jlw]

To answer briefly the question posed in the original subject line:

No.

At least, it's not an issue to people who simply want to use a camera to take pictures. You put whatever lens on the camera that produces the field of view you want to include and you frame your pictures accordingly.

People who make "crop factor" an issue seem to fall into two categories: (1) people who don't like the whole digital "thing" and are looking for reasons to be critical; and (2) people who have made major investments in Canon's 24x36mm-sensor DSLRs and feel the need to produce rationalizations for their decision.

Nobody else seems to care. For example, Hasselblad just announced a new camera, the H3D, which has a 36mmx48mm sensor. Are Hasselblad loyalists complaining that it's not a "full frame" camera because it doesn't cover the classical ~57mm x ~57mm image size of Hasselblad film cameras? Not that I've noticed.

 

[mauro scacco]

I think maybe the m9 will have already full format sensor.
I general I think a 1.33 crop on a camera system that is mostly built for wide angle lenses is a big problem.
But I hope the m8 is going to be a big seller, just to put the company out of their recent troubles.


Fabian

Completely agree
I think however an m8 without full frame is an error , also because Epson could exploit this lack to product in advance an RD2 M leica mount with full format ...:bang:

What a pity, old Leica ...

 

[rogue_designer]

My problem comes in with already owning fine lenses, who's FOV I like. Especially since I shoot a lot of available darkness.

If I want to use my 50 1.2 on an M8, It no longer acts like a standard - so I have to buy a new 35 1.2 to get an equivalent fast normal. And if I want a truly fast wide? My options are very limited. (Nobody makes a 21 f1.8 or faster that I know of - which is what I would need to approximate a 35mm on the M8).

If I'm shooting both cameras - I'd have to factor in that difference every time I swapped lenses. Much easier and better for me, if they act the same. Much Much better if I don't need to buy additional lenses to get a equivalent to the kit I already carry.

 

[Olsen]

I think you might be pleasantly surprised, as the issue of light fall-off is exactly what Leica adressed. And geometrical distortion? That doesn't even begin to be an issue with Leica wide-angles. Forget about your Canon stuff. They make very good tele lenses, but their wide-angles seem to be designed by the Coca-Cola company, in comparison to Leica and Zeiss. As for ISO performance, yes, the Canon files will be more noise-free at high ISO. But you will be able to hand-hold your M8 at least two to three shutterspeeds slower than your bulky Canon, gaining you two or three ISO steps. And if you go a bit longer, what about the Noctilux?

- A D-SLR is far better starting point than an oldfashioned rangefinder design when it comes to designing low light fall off. That Leica has adressed this is natural; for them it's live or die.

It is not as simple as you state that Canon is designing 'bad' wide angle lenses. When and if you are willing to pay what the most expensive of them costs, they are very good and fully up to be compared to any Leica R. Like the 35 mm 1,4L and even the 24 mm 1,4L. Even the extreme zooms like 16-35 mm 2,8L and 17-40 mm 4,0L are among the best on the market. But on all profesional FF D-SLRT cameras, Canon or Leica, these wide angle lenses must be stopped down to perform the best.

With all it's mass,- more tha 2 kilos with batteries a high quality lense, and mirrow/shutter dampening there is few cameras that I have had that makes it possible to shoot hand held at long exposure times like the 1Ds/1Ds II. I hardly think that M8 will match that either at less than half the weight. Further; the files from high ISO settings on 1Ds II are far beond anything from film or other digital cameras I have tried. Exept for it's threatening size, - some children start to cry when they see it, the 1Ds II is 'the' available light camera.

I have a 50 mm 1,0L that I haver used a lot with my analogue EOS3. That too is a powerful available light tool. But with the change of ISO settings at your fingertip it is dusting down. On the 1Ds II it is producing files with corners that is 2 - 3 stops darker than the center at max aparture. Guess the Noctilux will perform even worse on the M8, uncorrected. Like all large aparture lenses do on large sensor cameras. These lenses have lost much of their value in the digital world. - At least on a camera that produces so bright and noice free files as the Canon EOS 1Ds II.

The Leica M8 will be lighter and more compact. Even cheaper. That is a lot.