rattymouse wrote:
The examples of uncorrected Q images that I have seen show massive distortion, almost to the point of looking like a fish eye lens. Distortion values of 10% or greater are bandied about. I would never suggest a lens with such a high amount of distortion is "excellent". To call a design like this "excellent" simply devalues the meaning of this word.
If the lens produced <1% distortion like the Sony then my opinion would be radically different.
Why?
In my mind excellence is based on the attributes of the results achieved, not some arbitrary opinion that some % of distortion is unacceptable.
rattymouse wrote:
Some might consider 10% distortion excellent. Most probably dont though I suspect.
But it's irrelevant if it never appears in the final image....it could be 90% for all I care, if it doesn't show. It's true that 10% distortion would be problematic ...on film, but this lens will never be used outside a closed digital system. You logic is stuck in the 20th century and film...evaluating the excellence of this lens independent of the entire imaging system is like pulling one of the lens elements out of the excellent lens on the RX1 and then saying "look how horrible this lens is...I can't believe they charge us $$$$ for this cr#p!"
CVickery wrote:
But it's irrelevant if it never appears in the final image....it could be 90% for all I care, if it doesn't show. It's true that 10% distortion would be problematic ...on film, but this lens will never be used outside a closed digital system. You logic is stuck in the 20th century and film...evaluating the excellence of this lens independent of the entire imaging system is like pulling one of the lens elements out of the excellent lens on the RX1 and then saying "look how horrible this lens is...I can't believe they charge us $$$$ for this cr#p!"
We were talking about the lens, not the camera. At least I was. Of course I understand that the two go together and that using the Q without software correction is absurd. My point, and this is the final time I'm going to state it, is I do not understand what you are paying for when the lens clearly has an extreme amount of distortion. Previous cameras had optically corrected lenses that cost a lot of money to design and manufacturer. If you remove the requirement to make optically corrected lenses, the design *must* get cheaper. I dont understand the price of this camera given that software does all the heavy lifting.
Here's the two points where we differ..
"Of course I understand that the two go together and that using the Q without software correction is absurd"
No, evaluating the lens is isolation is absurd. It is part of an integrated system.
"If you remove the requirement to make optically corrected lenses, the design *must* get cheaper"
There is no evidence of this. A optically corrected lens projects an undistorted image on the sensor. On film this was the objective and it isn't easy. On a digital image the objective may be a image with precisely 10% barrel distortion, but dramatically lower, say, LOCA. There is no reason that this would necessary be easier or cheaper to achieve since in order for the software to work properly the input must be precisely as expected. GIGO, if you will. A computationally simple transformation of the data and you end up with an image with no distortion and lower LOCA as well. The software is, in effect,the missing element, and with sufficient computational power may do a far superior job to a glass element, ... but only for digital images.
As far as I understand it's not only distortion but also CA that is software corrected. I'm not sure how severe is CA before correction though.
By the way, if you calculate the camera price compared to the M240 for instance, the lens would be actually free, so I do understand why Leica cut the budget on the lens.
Lee Saxon wrote:
I tend to agree with you about the lens choice Leica made here being not a great one, but I'm not sure we can really be sure the RX1 doesn't use correction. A lot of the corrections (even with the M240, a lot of Leica purists still deny those exist) are "baked in" even with raw, and there's no way to turn them off or even know that they exist.
edwardkaraa wrote:
I'm just curious what are the baked in corrections you are talking about. As a Leica M user myself, I've never heard anything about them. Now you make me very curious.
I wasn't aware of any distortion correction baked into M raw files. Color shift correction, yes, if it's turned on in the camera. One way to determine this would be to shoot a range of lenses on film and digital, side by side and compare.
Ratty - buying a Leica is never a purely logical decision.
edwardkaraa wrote:
I'm just curious what are the baked in corrections you are talking about. As a Leica M user myself, I've never heard anything about them. Now you make me very curious.
I don't own the camera so I'm sure you know better, but I thought it'd been discussed on the "A7r with rangefinder glass" threads that some of the M lenses which get color cast and vignetting problems on the A7r sensor exhibit the same issue to a lesser degree even on the M240 sensor but that camera has built in correction profiles which work with the 6-bit coding system.
I also thought I'd read here that the m4/3 cameras do a lot of this.
Lee, that's strictly for color shift correction. It's a requirement of the sensor's incompatibility with the short exit pupil distances of rangefinder lenses. The same thing is necessary shooting many tech camera lenses on medium format digital backs.
Yeah, I know it's not specifically distortion correction like the Q, but it's the same concept. I also thought I'd read that you couldn't turn it off (maybe that was the m4/3 bodies?). If it's optional, I don't have nearly as much of a problem with that.
Lee Saxon wrote:
Yeah, I know it's not specifically distortion correction like the Q, but it's the same concept. I also thought I'd read that you couldn't turn it off (maybe that was the m4/3 bodies?). If it's optional, I don't have nearly as much of a problem with that.
You have the option to use auto, manual and off. I always have it on off. You can manually select from a long list of lenses as well. The correction is strictly for color shading and vignetting. My lenses do not require correcting anyways.
Ratty - buying a Leica is never a purely logical decision.
I understand that. No one NEEDS a Leica, but many people buy them. Leica, so far, isnt for me, but I do buy expensive gear and I do so because I feel that I'm getting something that has been very well designed. These designs bring with them a cost because they either use exotic materials in the glass, or require extremely precise manufacturing, which adds significantly to the cost.
I thought of an example that might illustrate this better. Take two lenses, the 55mm Zeiss Otus lens and the el-cheapo Canon EF 50mm f/1.4. The Zeiss costs about 10 times as much as the Canon. Leave out build quality in this example and just focus on IQ. Assume that they have equal build quality. If Canon developed a software correction that could bring the IQ of the mediocre 50mm f/1.4 lens up to Zeiss Otus levels of IQ, would you pay $4000 for it then?
Canon EF 50mm f/1.4 ~$400.
Zeiss Otus 55mm f/1.4 ~$4000.
Canon EF 50mm f/1.4 w/software (IQ at Otus levels) ~$4000.
I would find it VERY hard to buy the Canon lens for $4000. I just dont want to pay for software. Not to that extent anyway.
rattymouse wrote:
I understand that. No one NEEDS a Leica, but many people buy them. Leica, so far, isnt for me, but I do buy expensive gear and I do so because I feel that I'm getting something that has been very well designed. These designs bring with them a cost because they either use exotic materials in the glass, or require extremely precise manufacturing, which adds significantly to the cost.
I thought of an example that might illustrate this better. Take two lenses, the 55mm Zeiss Otus lens and the el-cheapo Canon EF 50mm f/1.4. The Zeiss costs about 10 times as much as the Canon. Leave out build quality in this example and just focus on IQ. Assume that they have equal build quality. If Canon developed a software correction that could bring the IQ of the mediocre 50mm f/1.4 lens up to Zeiss Otus levels of IQ, would you pay $4000 for it then?
Canon EF 50mm f/1.4 ~$400.
Zeiss Otus 55mm f/1.4 ~$4000.
Canon EF 50mm f/1.4 w/software (IQ at Otus levels) ~$4000.
I would find it VERY hard to buy the Canon lens for $4000. I just dont want to pay for software. Not to that extent anyway.
I would agree with you in your scenario, but what you are describing here isn't software it's magic. No software can create IQ if the 'information' isn't there in the first place. Bringing this back to the topic of this thread, I would contend that the lens in the Q is excellent (distortion aside) , precisely because for the impact of the software correction to be so small, the lens must be resolving extremely well in its' raw uncorrected state.
rattymouse wrote:
My point, and this is the final time I'm going to state it, is I do not understand what you are paying for when the lens clearly has an extreme amount of distortion.
German salaries?
And maybe a little bit for the rescue package for Greece
rattymouse wrote:
An autofocus Leica M will be an epic fail. I can count on one hand the number of times I've heard people say that they wish that their rangefinder was an autofocus camera.
oh, there's a few of us around
We are the Contax G and Fuji Xpro1 users
that's a very, very flawed argument; that's assuming the excellentIQ coming from a glass like Otus can be replicated via software.
going back to the Leica Q...IMO, it's still an achievement in its own right, distortion or not. the details wide open is astounding...if the distortion (which CAN be corrected via software without any perceptible loss of IQ) is the price to pay for that detail (which CANNOT be achieved via software)...it's fine by me.
rattymouse wrote:
I understand that. No one NEEDS a Leica, but many people buy them. Leica, so far, isnt for me, but I do buy expensive gear and I do so because I feel that I'm getting something that has been very well designed. These designs bring with them a cost because they either use exotic materials in the glass, or require extremely precise manufacturing, which adds significantly to the cost.
I thought of an example that might illustrate this better. Take two lenses, the 55mm Zeiss Otus lens and the el-cheapo Canon EF 50mm f/1.4. The Zeiss costs about 10 times as much as the Canon. Leave out build quality in this example and just focus on IQ. Assume that they have equal build quality. If Canon developed a software correction that could bring the IQ of the mediocre 50mm f/1.4 lens up to Zeiss Otus levels of IQ, would you pay $4000 for it then?
Canon EF 50mm f/1.4 ~$400.
Zeiss Otus 55mm f/1.4 ~$4000.
Canon EF 50mm f/1.4 w/software (IQ at Otus levels) ~$4000.
I would find it VERY hard to buy the Canon lens for $4000. I just dont want to pay for software. Not to that extent anyway.
yeah, remember that you're paying a couple of Germans assemble your Leica, whereas if you go with other companies it's probably some sweatshops for minimum wage :P
For me at this stage, I'm just going to sit back and enjoy the images.
The Q is not at M level for 28mm, but it does not pretend to be (and nothing else is), and has lots of other features which the M does not, AF being the most huge, with CF not far behind. I'm not sure why one needs IS for a 28, but it has that too
Nico and others show the thing shoots pretty darn well, and even more important, the camera is a total hit. It's paid for itself in terms of publicity alone. Wired reviewed it!
It's not my Leica, but I admire many aspects about it, though not the AF
The 10% distortion seems crazy, but in the end it does make good images, and the overall form and implementation is a new benchmark for mirrorless.
As to the price: I never saw a new FF leica I could afford, so......
