Skip to main content

⚠️ Please note that this topic or post has been archived. The information contained here may no longer be accurate or up-to-date. ⚠️

Lack of Canon lens correction profiles partly Canon's fault?

Comments

4 comments

  • SFA

    Ric,

    This would certainly seem to be the case as far as I can tell. 

    One thing that struck me when the latest 1Dx came out (and I assume also applies in many ways to all Canon releases since then) is that the speed at which CR3 support became available - across all software vendors - appeared to be slower than one was used to.

    I found software that claimed to be able to render the CR3 files but it seemed to me to manage only an OK job of it.

    Using the same set of sample (pre-production body) files the samples for Out-of-camera jpgs looked head and shoulders (and most of the torso) better than the RAW files, using conversions I could access at the time.

    After checking a few of the images to make sure that they were not great for some and poor for others, it occurred to me that, if I had been in the market for that body at the time, it would probably pay me to spend time getting to understand the internal jpg processing and just skip RAW files completely.

    Clearly the jpgs were also deploying lens correction - so all boxes would have been ticked for my main volume of images.

    Going further into this I suspect that for most camera and lens manufacturers the issue of software lens correction, having become a fundamental part of lens design work some years ago, should mean that almost all modern lenses from quality manufacturers will likely be passing optimised correction data directly to RAW conversion software exactly as they do it for internally generated jpgs. 

    That does not leave much for a third party RAW developer software to do other than read the data and apply it. As a C1 source wrote somewhere a few months ago, for most lenses any C1 added profile data might relate to no more than a difference of intention for the correction of edge sharpening and, possibly, the values to apply for Light Falloff. Both of which are rather subjective matters.

    However, if for some reason the correction data is not being picked up at all (seems unlikely   .... but ....) then that is a truly serious problem. For the manufacturer, if they are making things difficult for third parties. Unless their business results suggest otherwise.

    0
  • Markus Stamm

    Yes, unfortunately Canon decided to either not disclose how to decode correction data in CR3 files or (more likely) to not include that data in the RAW files at all. Whatever their motivation was, I honestly don't get it. Software correction offers some nice opportunities especially on mirrorless cameras, since it allows the viewfinder to already show the final image with all corrections applied (unlike on DSLRs, where viewfinder image, and as such fiel of view, and the final software corrected image can differ quite a bit even on lenses with moderate distortion). That's the positive aspect.

    The really negative trend is, however, that many manufacturers now seem to underdesign their lenses optically with heavy software correction to make up for it.

    One example is the Fuji XF 35/2 mentioned in the initial post, a lens that comes with heavy distortion and vignetting and some serious focus breathing on top (the XF 16/2.8 is similar, btw.). I don't know why they went that path with these lenses, because other XF lenses show that Fuji knows how to correct these issues in glass instead of software (XF 14, XF 23/1.4, XF 23/2, XF 50/2,...).

    Nikon also released a few lenses with the Z system that come with very pronounced distortion (and vignetting)... too much for Nikon themselves to even allow the user to disable the in-camera distortion correction for some of their own lenses (Z 24-70/4, Z 14-30/4,...). Luckily, like almost any other manufacturer of modern mirrorless systems, they include the correction profile in the RAW files, so in C1 the full glory of almost fisheye distortion becomes visible (unlike as in Lightroom, because Adobe hides the native performance of the Z lenses, too, I assume they have an agreement with Nikon to do so, but that's a different story).

    However, no one currently underdesigns lenses the way Canon does, at least in some cases. The RF 24-105 IS STM does not even cover the full sensor area at its shortest focal length setting without software correction (pitch black corners, close to fish-eye level distortion). So I really don't get why they do not deliver the obviously necessary (and intended by design) correction data with their CR3 files.

    Yes, it would be nice if C1 delivered more RF correction profiles... which I am sure they will (and have to) over time. However, the initial flaw is not with C1, but Canon's decisions.

    0
  • SFA

    Markus,

    Interesting observations.

    I think after so much time with no profiles for new Canon lenses it was clear that some sort of problem exists and it had to be either what you have outlined or some sort of commercial agreement constraint. Or both.

    Do Canon still make their own lenses or have most of them been outsourced by now?

    The market for new "proper" camera equipment sales still seems to be contracting so presumably every one of the remaining manufacturers is trying scrape every yen of profit from wherever they can find it.

    The transition to mirrorless cameras means that no one has to care too much about optical accuracy through a mirror based viewfinder. All problems can be fixed (or otherwise overcome) by software now that there is more than enough computing power, memory and battery capacity available to allow it. Just look at the abilities of Mobile Phones!

    Obviously no one will see the optical limitations n thenew lenses if the lenses cannot be used on older DSLR devices. So long as the technical details are redacted no one would consider trying to make some sort of adaptor to use the new lenses with older technology. (Either that or people had considered it, experimented and decided there is no market for the results!)

    Like everything else in the world at the moment, it seems that the economic model involves an attempt at earlier than necessary obsolescence of what has gone before.  Just to keep the wheels turning. For now.

    One might consider that a rather desparate concept that is highly wasteful of paste materials and energy consumption for no real benefit. At least, no benefit I can think of that can justify the waste in the majority of cases. There may be a few exceptions, but not many.

    The added complexity that tends to come with such changes is also poor for longevity. But good for sales, so long as one can maintain a functioning market for the product or its replacement. Somehow.

    0
  • Allan OConnor

    All this may be true.

    Even if it is, it doesn't excuse Capture One's failure to tell their user base what their plans are, if they have any, concerning the Canon RF line.  They've screwed this up badly.  Their software is unuseable for the RF lenses that require significant correction unless you shoot jpeg.

    They aren't supporting almost all mirrorless lenses from the largest camera manufacturer in the world - many that have been out for over two years.  People are buying their software with the expectation that it will work. It doesn't.    Adobe seems to be able to generate RF lens profiles.

    While I love Capture One, life is short.  I don't want to be dependent on a vendor that treats their customers that way.  I'm looking to move on.

    0

Post is closed for comments.