2 way keystone (horizontal and vertical) with 4 independent guides
The fact that the four guides are linked together makes your two way keystone tool practically unusable. You have to make them independent from one another! Your competitor is doing it, I'm sure you can do it too.
-
Thomas Kyhn has many requests over many years and consolidated into below request..
https://support.captureone.com/hc/en-us/community/posts/360014158198-A-new-keystone-tool
Below search result shows all requests under "Community" section..
https://support.captureone.com/hc/en-us/search?utf8=%E2%9C%93&topic=360000154057&query=keystone
But except cosmetic (color) change to guideline no other changes were made..
1 -
Yes, this is much needed and is included in the summary of keystone requests (first item) that Prasad Palaniyandi links to above. See this separate topic also.
1 -
"Your competitor is doing it, I'm sure you can do it too. "
Hopefully, if and when something is provided, it offers better results than the "competitor" (or its copycat implementations) manage.
Just my opinion of course. Based on what the various User guides and online resources offer as examples of how to use them.
0 -
"Hopefully, if and when something is provided, it offers better results than the "competitor" (or its copycat implementations) manage."
I haven't yet had any issues with LR's keystone tool (and I've used it quite a bit), and don't remember seeing anyone complain about it. Like everything else, I'm sure it could be better, but even in its current state it has none of the issues and limitations of Capture One's keystone tool, at least none that I'm aware of.
1 -
Couldn't agree more. If there's one great tool in LR, it's the keystone. Fortunately for C1 user, it is possible to use that keystone tool in Photoshop under Camera raw filter. But I prefer aligning my image at the beginning of my workflow, so it would be best if C1 could up is game on that matter.
2 -
The problem I see is that the examples used as part of the tutorials or training or User guide - at least the stuff I have seen, deliver a very distorted result for the subject.
LR manages to mask somewhat so that unless one considers it carefully one may not spot the anomaly.
Indeed in some situations where "making an impression" seems to be the primary objective, the results may find favour with many - just as turbocharged HDR can create a memorable image that is nothing like the original subject in appearance. I have no problem with that but no need of it either.
The "clone" Keystone in Affinity is much the same - except the example images in the User Guide (as of about a year ago, may be different now?) provided an image that made a statement with "corrected" verticals and horizontals but looked totally wrong architecturally if one thought about it. An impressive manipulation but a totally distorted result.
I was reminded of Bill Brandt's rocky beach images where a low camera position and a wide angle lens could produce "huge" foreground boulders from relatively small pebbles but looked OK juxtaposed with carefully selected parts of a human anatomy positioned behind them.
There is a place for artistic photography of course. I have no problem with that whether in portraiture, landscapes or architecture. I'm just not sure it should be presented as something to take precedence over reality when presenting options for - in this case - Keystone correction.
-2 -
Interesting, but this is not an artistic or conceptual forum... 2 way keystone in C1 is unusable, punto.
0 -
And the 2 way keystone is useful for one point perspective cases, i.e. when your facing an object, let say a building.
0 -
@SFA – So you haven't actually used LR's keystone tool? If this is the case, it should be safe to say that the confidence with which you dismiss it lacks any basis.
0 -
Not for a while.
However if it works differently to the examples provided in the various online "User Guides" and tutorial resources - and significantly differently to the Affinity clone version which I have experimented with - why would Adobe present anomalous examples to promote the functionality?
I can, for example, understand that someone might want to present a multi storey building, shot from the ground at eye level, as having completely vertical verticals and completely horizontal horizontals, as if it habe been shot from the centre of a side half way up the height and had minor "tweaks to be made (after lens correction) to present it as desired.
I can understand why a real estate agent might want to use wide angle lens to make a keystone adjusted image of a room look bigger than it is in a sales presentation.
However, if one is trying to record a view of something accurately but cannot shoot from the perfect location for doing so I, speaking personally, do not think that manipulating pixels to make extreme adjustments is necessarily the best way to approach the requirement. The challenge really comes with shots that have to be taken close to the subject or where space is severely constrained - interiors for example.
Sure you can take a trapezoid in an image that was shot in 3 dimensions and force it to become a rectangle in what appear to be 2 dimensions but unless it is an area devoid of any significant detail there is a good chance it will look "wrong" compared to expectation.
For some purposes that might be perfectly acceptable as an end result. For others - not so much.
If a "true representation" of a 3 dimensional subject needs to be converted to a 2 dimensional representation as if it was shot from a different position at a different set of angles then I rather suspect that a lot more information would be required for the calculation than is available from a regular camera image.
At a distance that may not matter much. Our brains will correct the discrepancies anyway. The problems arise when using extreme angles or shooting close to a large subject. (And a few other influences that are probably outside the scope of this part of the discussion.)
-1 -
wut??
1 -
@SFA – Whether or not you like parallel verticals, the effect of keystone correction, etc. etc., is entirely irrelevant in this context. You've brought up these personal preferences countless times before and they haven't become more relevant since the last time. It is evident that quite a few Capture One users, myself included, are dissatisfied with the keystone tool, and for someone who clearly has no interest in improvements as far as this tool is concerned to repeatedly brush aside the issues other users are experiencing while presenting at immoderate length their own considerations about the tool in general comes across as more than a little arrogant.
0 -
Thomas,
Thank you once again for your comments. I certainly respect your tenacity in persuing this matter which is clearly something of importance to you and the "Quite a few" other users that share your concerns if not your staying power. Perhaps there is also a silent majority out there that has decided not to join the discussion.
My point is not so much about not improving things as it is about improving things.
It would be nice to see something that could offer more than both the C1 approach and the LR (or Affinity, etc.,) approaches. I just do not see that some of the things you wish to copy over from the LR/PS approach are satisfying enough to make it worth the effort without taking in the cost and commitment of developing an entirely new tool. All of the tools that I have looked at in one way or another from the main line of suppliers, could be better.
However, to achieve the level of pixel manipulation you seek from the C1 offering may require something much more than a few tweaks to some existing code that seems to be provided to simulate, to some extent, the use of Tilt and Shift adjustments on traditional large format film camera Movements or more recent 35mm T&S lenses and their digital counterparts. And, of course, the Auto adjustments available from some Phase One hardware. Yes I appreciate that those adjustments are often more to do with depth of field and focus than keystone but the two subjects are somewhat intertwined.
Apart from your specific complaints I see some issues with C1 that are also present in LR and Affinity (to use just 2 examples).
If you look at the Help page for the C1 Keystone tool you can also just about see the problem I have referred to above. After correction the building used in the sample shot looks imbalanced. For example the window slots at the bottom (nearer the viewer) are smaller in height than the slots at the top (further from the viewer).
It is a little more subtle than the effect in the LR documentation I looked at a while back and a lot less evident than for the sample used in the Affinity guide but it is there nonetheless.
As a correction compromise it is misleading to the eye but, for that particular image shot at some distance from the subject, it is sort of OK. With other images the results can be less convincing and all the Guide examples I have seen from different software are similar in that respect.
Object size discrepancies are much more obvious in images taken at shorter distances and with greater angular correction required. A smart system would try to match the geometry, allowing for 3D objects to be adapted to a 2D representation, with aspect correction matched to the adjustments.
However, for that it may often require more measurement information than non-specialised cameras can provide. Or perhap just more than the camera manufacturers make available for general purpose use.
Taking the C1 User guide page for the Keystone correction as an example, if the image could be corrected to position the camera viewing point at the mid-height of the building a more balance vertical correction could be achieved and the horizontal correction, being a shorter side of the building, should be less evidently distorted.
That might not be the optimal viewing position for the purpose of the shot so some other position on the vertical would likely work better for the purpose of the shot. Perhaps just higher or just lower than the mid-height.
But in terms of the correction - which basically is an enhancement of the shooting position for that subject - the result should be better if adjusted relative to the middle of the image rather than the bottom.
If some size anomaly of objects that should be the same size in the image is acceptable in a sort of "as the eye might see it" way, then the old time "80%" of adjustment applied by default might work rather well.
In most subject matter and in particular when considering documents and paintings and related keystone corrections that many seem to wish could be corrected even from some extreme angles, any quality and accurate result will be dependent on much more than simple geometry even allowing for some element of non-linear adjustments to cope with scaling the content as part of the adjustment when fitting to the geometry.
It remains to be seen whether any such AI technology appears from the C1 labs as they dive into new feature developments.
Of course, Apple may get there first with the iPhone and make more traditional cameras and the software that is used for them somewhat redundant even more rapidly than has been the case just recently.
-2 -
By the looks of it, it doesn't seem like you are even using this tool at all: documentation this, guide examples that... come on now... In real life, photographers like me (it's my job) use the keystone tool to address vertical or horizontal lines that are a little skewed; not to change the whole geometry of a picture!!! Jesus! What are ou talking about! I dont want the keystone tool to go up a ladder for me, I just want it to efficiently straighten verticals and/or the horizontals in my image. Lightroom does it well, C1 doesn't.
This request is not destined to you SFA. It is for the C1 techs and product managers to see and take into account. I really don't understand what part you are trying to play in this... End of conversation.2 -
@SFA – May I suggest you start a separate topic about the general woes of keystone correction rather than repeatedly flooding specific requests with endless only tangentially related considerations?
2 -
Agree with you Ulysse, primarily I want the basic funcion to be working properly! And C1 does not even deliver that ... It's just full of bugs and no one at C1 seems to care. This is the main reason I did not take the last upgrade offer. They charge an arm and a leg - even for upgrades - (considering it "only" a RAW converter) but don't even get their basic tools right.
2 -
All of the tools that I have looked at in one way or another from the main line of suppliers, could be better
haha ! think I get it now, SFA is in the entertainment business and i'm now also sure he actually is not using C1 or any other imaging software he is referring to in practice. I have to say that when I will give up c1 for good and this forum I will miss his quirky posts a little as I do not know any other forum where a alleged user is spending so much time to derail discussions about improvements or bugs with such verve.
0 -
Thomas,
"@SFA – May I suggest you start a separate topic about the general woes of keystone correction rather than repeatedly flooding specific requests with endless only tangentially related considerations?"
I'll leave the option to create ever more threads on the same subject entirely up to you.
It's an open forum. People can express opinions. No one is forced to read them.
-1 -
@SFA – Then by all means exercise your right to keep flooding threads on specific issues with immoderately long rambling comments on tangentially related subjects.
0 -
Absolutely agree with the request. LR has hands-down the better tool. You can get away with C1s implementation but its quirky and awkward to use. Not only separating the horizontal&vertical lines help, also the loupe is very handy in LR.
1 -
Actually no - you don't get away with C1 implementation because it simply does not work as supposed. It's buggy and unusable. See an example here:
https://support.captureone.com/hc/en-us/community/posts/4408089696913-Keystone-bug2 -
Ok, I agree that these examples are really not what one should expect from a keystone tool. And I experienced that with some images as well. On images with only a small amount if distortions I usually get away with it. But from a customer standpoint that is not enough. Especially if other solutions can do better (LR or even DXO Viewpoint).
0 -
Even with subtle adjustments you often get incorrect results. See the example here for instance.
0
Post is closed for comments.
Comments
23 comments