Luminous Landscape Forum
Raw & Post Processing, Printing => Digital Image Processing => Topic started by: Rajan Parrikar on June 01, 2016, 05:07:06 am
-
FYI: major enhancements are claimed.
http://relaunch.piccureplus.com/update-version-3/
-
Does this application do anything better than I can get from Lightroom?
-
Does this application do anything better than I can get from Lightroom?
Hi Mark,
In principle, all Deconvolution based sharpening applications at least have the potential to do better than Lightroom's implementation (which is pretty mediocre, e.g. in comparison with an old time favorite FocusMagic).
The difference with Piccure is that Piccure is supposed to apply spatially variant restoration of image detail, IOW it treats corners differently compared to the image center, in function of the local amount of blur caused by various aberrations and defocus.
Unfortunately, previous versions also created some dark halo artifacts and it was relatively slow. The latter issue has apparently been addressed (according to their announcement) with the new Version 3. Whether the restoration quality has improved remains to be seen.
At least the upgrade seems to be free for previous owners of a license, which is better than their normal upgrade policy.
Cheers,
Bart
-
Thanks for those useful perspectives "in principle" Bart. Now, the key thing I would be interested in hearing about is actual comparative results from user experience.
-
At least the upgrade seems to be free for previous owners of a license, which is better than their normal upgrade policy.
For the past 18 months, all updates were for free. Those were not minor updates, but they were rather significant. It was two major releases: 6 months after v2.0 came out with a 50% speedup and 18 months after v2.0 came out with about 66% speedup for Quality+ and 10-30% for other settings in v3.0. In addition to a refined user interface, Retina support was added and DxO and CaptureOne are now supported. We offered a 75% discount to users of version 1 at the time of the 2.0 release. In software, all costs are incurred months/years prior to a release and accumulate to a "big negative cumulated cashflow". So we essentially did work for free for the past 18 months - because we believe our customers deserved it. You can like the software or not - it's your personal choice. But those "free lifetime update policies" of many other software vendors often turn out to be very mediocre improvements at best. We don't promise these free lifetime updates, we are expensive - but we offer fast support and really develop the software further. Version 3.0 is a lot better than version 2.0 (at least in our opinion and the customer feedback so far). I don't think there was any update to the software you mentioned in the past 2 years? So at least to that solution we should compare favorably.
As whether one may need a solution like piccure+ ("who needs software if only good glass and technique matter") - we do believe adaptive deconvolution is the way to go (that's why we keep investing). Consider that:
- Most cameras have optical low pass filters deliberately blurring the image irrespective of the lens used
- Most lenses suffer from back-/front-focusing
- There is considerable copy-to-copy variation in lenses
- Fast lenses easily lose 20+% towards the edges at open aperture - coma, spherical aberration and other aberrations are hard to correct with unsharp masking
So you can easily end up with only 50% of "the sharpness you paid for". We hope you enjoy the free update - we put a lot of effort into it. And don't hesitate to shoot us an email if you run into problems: [email protected]. Please don't forget to update your graphic card drivers (nvidia and on-board) if you run into any problems with the GPU acceleration (and restart the system).
@Mark: there are actually many comparisons on many forums and a lot of bloggers wrote about it (just search). To cut a long story short: if you get good results with unsharp masking and you are happy with it - piccure+ probably won't help you much. If you still think "the image lacks some sharpness - particularly towards the edges" you may give it a try. It complements other software solutions rather than replacing them - because it does something unique. Personally I would recommend anybody to try it with his own images, because the improvements depend a lot on your camera/lens combination and your overall photography preferences. If you do a lot in the studio at f5.6 or if you shoot sports images with 17 MP in JPG or band gigs at ISO 25000 - piccure+ may not be your top choice... If you shoot full frame (or above 20 MP RAW) and fast apertures and know what OLPF, coma and diffraction mean - there could be a chance you may find it useful. It's not made for "kit equipment" - but rather the more expensive one. A lot of medium format photographers use it as well...
Best,
Lui
Co-Founder
-
Thanks for those useful perspectives "in principle" Bart. Now, the key thing I would be interested in hearing about is actual comparative results from user experience.
Mark, the "in principle" better performance depends on the implementation. LR/ACR's implementation is quick and dirty and produces lots of artifacts when pushed a bit (by shifting 'Detail' closer to maximum).
Almost all decent implementations do better, RawTherapee is free and can apply Deconvolution also on TIFFs, not only on Raws, FocusMagic is very solid, Topas InFocus is very effective (but easy to over-do, i.e. user dependent), Topaz Detail not only offers a Deconvolution "Deblur" control but offers lots of additional detail controls. Franzis Sharpen Projects also does a good job, but I don't like it's Raw conversions. Just my 2 cents.
It would of course be preferable to have such quality in a simpler workflow, not having to use external programs.
Cheers,
Bart
-
Thank you both. If sharpening is the key incremental benefit I probably don't need this application, because I obtain very sharp photos from my Sony a6300+Zeiss lenses with moderate capture and output sharpening in LR. Some form of deconvolution is helpful to rescue detail when I mis-focus (i.e. pilot error), but I have never been satisfied with the ultimate quality (from a photographic perspective) of those rescue attempts (using demo versions of such software) and therefore usually trash photos that are out of focus when they should be in focus. Now that too may be pilot error - maybe I need to spend more time focusing on deconvolution technique (pun intended); but that will be for another day :-)
-
Thank you both. If sharpening is the key incremental benefit I probably don't need this application, because I obtain very sharp photos from my Sony a6300+Zeiss lenses with moderate capture and output sharpening in LR.
Hi Mark,
The default image capture quality of a good lens, and by using proper technique (e.g. tripod for stationary subjects), produces an image with roughly a Gaussian blur of 0.7 (or worse) after demosaicing.
Only deconvolution can restore the original scene resolution (before lens aberration, diffraction, filterstack with or without OLPF, Bayer CFA, photosite aperture shape, all add some blur), and at the same time as removing micro contrast blur, it removes the "veil of drabness" by lifting the entire MTF response, it's as if you've upgraded the existing lens to a more expensive model. Good implementations improve the signal quality more than the noise, i.e. produce a better S/N ratio, which is more robust to subsequent postprocessing.
In extreme cases it may be able to recover seemingly lost detail, but that usually requires moderate to low ISO captures to start with.
Some form of deconvolution is helpful to rescue detail when I mis-focus (i.e. pilot error), but I have never been satisfied with the ultimate quality (from a photographic perspective) of those rescue attempts (using demo versions of such software) and therefore usually trash photos that are out of focus when they should be in focus.
Mis-focus can cause 2 issues. First it poses a challenge for the deconvolution algorithm to separate amplification-/read- and demosaicing noise from the intended signal which is also noisy (photon shot noise). The defocus blur may have reduced the signal's micro-contrast to the level of the combined noise floor itself. Second, there may be reasonably focused parts of the image that get over-processed if the locally non-existing blur of more blurry parts of the image is removed. Masking or adaptive deconvolution may be able to solve the latter issue.
Cheers,
Bart
-
Thank you both. If sharpening is the key incremental benefit I probably don't need this application, because I obtain very sharp photos from my Sony a6300+Zeiss lenses with moderate capture and output sharpening in LR. Some form of deconvolution is helpful to rescue detail when I mis-focus (i.e. pilot error), but I have never been satisfied with the ultimate quality (from a photographic perspective) of those rescue attempts (using demo versions of such software) and therefore usually trash photos that are out of focus when they should be in focus. Now that too may be pilot error - maybe I need to spend more time focusing on deconvolution technique (pun intended); but that will be for another day :-)
Mark,
I too am not impressed with the ability of deconvolution to restore out of focus images. That approach is best limited to when one can not retake the image and for forensic work such as being able to read license plate numbers despite the deconvolution artifact.
Even with the best lenses, sharpening is necessary to reduce the blur incurred during demosaicing of the raw file. With my Nikon D800e and Zeiss 135 Apo Sonnar, I can get very sharp images at f/4 with ACR/LR sharpening, but Focus Magic does a better job and is worth while with one's best images, especially when one must stop down to f/16 or more for depth of field. Even Zeiss lenses are susceptible to diffraction. :)
Does your A6300 have a low pass (blur) filter? If so, I would expect deconvolution to be superior to LR/ACR.
Regards,
Bill
-
Hi Bill,
It probably does have an AA filter, being 24 MP on an APS-C sensor; but technology is everything and from what I can see on some 13*19s I printed last night in my 4900 on Ilford Gold Fibre Silk, sharpened in LR, you wouldn't know it. The detail is all one wants or needs. Doubtless I could go to 17*25 (as I did on another photo last week) and it would be just fine. My main interest is what the prints look like viewed in normal viewing conditions, and my optometrist "certified" my vision as 20/20 the other day :-).
-
Hi Mark,
The default image capture quality of a good lens, and by using proper technique (e.g. tripod for stationary subjects), produces an image with roughly a Gaussian blur of 0.7 (or worse) after demosaicing.
Only deconvolution can restore the original scene resolution (before lens aberration, diffraction, filterstack with or without OLPF, Bayer CFA, photosite aperture shape, all add some blur), and at the same time as removing micro contrast blur, it removes the "veil of drabness" by lifting the entire MTF response, it's as if you've upgraded the existing lens to a more expensive model. Good implementations improve the signal quality more than the noise, i.e. produce a better S/N ratio, which is more robust to subsequent postprocessing.
In extreme cases it may be able to recover seemingly lost detail, but that usually requires moderate to low ISO captures to start with.
Mis-focus can cause 2 issues. First it poses a challenge for the deconvolution algorithm to separate amplification-/read- and demosaicing noise from the intended signal which is also noisy (photon shot noise). The defocus blur may have reduced the signal's micro-contrast to the level of the combined noise floor itself. Second, there may be reasonably focused parts of the image that get over-processed if the locally non-existing blur of more blurry parts of the image is removed. Masking or adaptive deconvolution may be able to solve the latter issue.
Cheers,
Bart
Thanks for those insights Bart.
Mark
-
Just thought I'd post this information. To get the second speed boost from Piccure+ version 3 your Nvidia graphics board needs to have a CUDA "Compute Capability" of at least 3.0. My current board Nvidia Gtx 580 comes in at a 2.0, but the Piccure+ version 3 still runs faster than Piccure+ version 2 did.
I just purchased a new graphics card so will see if that really kicks up the speed of Piccure + version 3, or back to Best Buy it goes.
Here's a list of Nvidia boards and their compute capability:
Tom
https://developer.nvidia.com/cuda-gpus
-
Just a few words. Yes the upgrade is very dependent on Nvidia card and levels.
On my GTX970 the results are impressive cutting down the processing by 8/10ths at least. On my GTX960 open CL fails but even without the processing time is cut in half. The GTX960 seems to have a lot of issues with various software and it's going to be replaced soon.
Does Piccure +, work that is subjective to the user but I love the results and have pretty much replaced Focus Magic with Piccure + for all my cameras especially Fuji and Phase One.
Paul C
-
I test drove the previous version of Piccure+ and the new improvements seem like a good thing. I liked the effect of Piccure+ on some stuff but on some stuff the dark edge synthesis was a no go.
Never the less, I would already own a license to Piccure+, and use it when it showed its strengths, but $79 for only one install caused me to choose not to make a purchase. I need two installs but will not pay the $119 for that luxury. I hope that someday Piccure will allow for a second install and only cost $79.
The idea that the relatively high cost reflects the fact that it has received a lot of upgrades seems valid until I consider that the history has seemed, to me, to have started as an unusable curiosity, then it was maybe good or maybe unusable, and now it may be better than that. It seems like it may just now be evolving to where it should have been at the onset.
I am going to wait to see if Piccure wants to expand its customer base by offering $79 for two installs.
-
Two vague sentences on quality and all the rest on commercial conditions. It could be helpful if you would expand and enhance your views of the application's strengths and weaknesses, say compared with Lightroom, in respect of image quality. That would provide a more useful context for evaluating the commercial issues.
-
Spending a few minutes with the free trial demo will explain more than I will.
-
Spending a few minutes with the free trial demo will explain more than I will.
In my case I would need a lot more than a few minutes because I tend to dig into stuff, but thanks for the advice - not that I hadn't thought of that, but I was hoping to hear more from others who have.
-
Never the less, I would already own a license to Piccure+, and use it when it showed its strengths, but $79 for only one install caused me to choose not to make a purchase. I need two installs but will not pay the $119 for that luxury. I hope that someday Piccure will allow for a second install and only cost $79.
The idea that the relatively high cost reflects the fact that it has received a lot of upgrades seems valid until I consider that the history has seemed, to me, to have started as an unusable curiosity, then it was maybe good or maybe unusable, and now it may be better than that. It seems like it may just now be evolving to where it should have been at the onset.
The people working on piccure+ all have PhDs in physics, computer science and are among the scientific leaders in the field of blind image deconvolution. In order to make something like piccure+ available, you need a strong team with experience in programming GPUs, multi-core processors, memory management, doing cross-platform software development and know all the respective APIs (piccure+ can be integrated in more programs than any other solution). The "large" players in the market have tried for years to aggressively hire away our key developers. The technology is used in much much more expensive camera systems in modified ways... For a rational business person it would make very little sense to research, develop and market such a technology over several years and subsequently sell it in the niche market of DSLRs (the market with the steepest sales decline high-tech). But the people behind piccure+ like photography and they wanted to develop something that "they could show their family" instead of simply licensing it away... In addition to that, in software you invest everything upfront - and don't know whether you will ever make a dollar of revenue, particularly with a new technology. There is a big risk involved as well - and most academics/scientists would rather work for a lot of money in a large software company than be faced with this amount of risk and uncertainty...
In the end you need to see the "119.99 USD benefit" for yourself. However, since most of our customers spend several thousands USD on their equipment they are willing to pay that price for the "package" they get. But not everyone sees that benefit and we are ok with that choice.
I could expand on how we differ from a technological perspective - but I think the bottom line is simply: our solution may make your images sharper than other solutions in the market - and recover "several thousand dollars of sharpness" you lost due to manufacturing variation and design choices. It's no magic bullet for everyone, but can help many people.
Have a nice weekend.
Best,
Lui
-
Mark et. al., attached is a sample. Zero sharpening on the left, sharpened on the right using the parameters as set. The image is a 1D X file using an 11-24mm @ 11m, f/7.1, ISO 2500, with shadows lifted a bit. It's at 100% view, of course.
(In this image, if denoise is set to off, the noise in the shadows noticeably increases, so I turned it on.)
The sharpened image is clearly much sharper, but suffers from a fatal flaw - the high contrast edges look dreadful, e.g. on the three women's hijabs (head scarves).
For images that don't have high contrast edges, or edges of the kind caused by using an extremely high aperture lens wide-open, Piccure+ can produce extremely impressive results. But for those images with the flaw, it's a royal pain in the butt to fix it in Photoshop, which may or may not be worth it.
I reported this problem long ago. I don't know if it can be fixed or not. I hope the developers can fix it!
-
That's very useful Damon, thanks. A couple of observations: firstly, since the screen grab is an enlargement of a small portion of the image area, and the effect you mention is present but rather subtle, I wonder how visible it would be say in a 13*19 or even 17*25 inch print of the whole photo. Large magnifications of small areas (less politely known as pixel-peeping, but I'm not going there :) ) is useful for identifying underlying issues, but their significance could be another matter. This depends on practical results in prints. Secondly, in both renditions I notice what looks like a fair bit of colour noise in the pavement and the windshield of the automobile notwithstanding that noise mitigation is deployed. If you use LR, out of curiosity it would be interesting to see whether this same photo fares better or worse there in respect of both sharpening and noise mitigation.
-
The sharpened image is clearly much sharper, but suffers from a fatal flaw - the high contrast edges look dreadful, e.g. on the three women's hijabs (head scarves).
For images that don't have high contrast edges, or edges of the kind caused by using an extremely high aperture lens wide-open, Piccure+ can produce extremely impressive results. But for those images with the flaw, it's a royal pain in the butt to fix it in Photoshop, which may or may not be worth it.
I reported this problem long ago. I don't know if it can be fixed or not. I hope the developers can fix it!
Hi Damon,
Have you tried reducing the rendering control (to something closer) to zero, i.e. more emphasis on detail than contrast?
I've 'played' around a bit with Version 3.0 on some different files, and I like the potential ability of P+ to correct spacially variant Capture blur, and the fact that the blind deconvolution is of decent quality. However, it's the fully automatic detection with very little manual override capability that produces results that can be improved upon by other deconvolution solutions, or more extensive sharpening suites.
Cheers,
Bart
-
Hi Mark and Bart, regarding printing, I'm afraid I have no idea because due to currently being in Tajikistan I haven't been able to make any prints for several years.
Regarding image noise, yes there's plenty -- welcome to the world of pre-2016 era Canon sensors! :D The version on the left has luminance NR set to 9 (detail 50) in LR and shadows at +12.
If rendering is set to zero, the problem with high-contrast edges is exactly the same, i.e. it makes no difference what value it's set to.
Attached is roughly the same crop, but this time from LR with LR capture sharpening set to the "Faces" preset (rather than set to zero as with the version above).
Finally, for a more favorable view of Piccure, I've attached a crop of a 100% view with the same body and lens at 11mm, this time tripod mounted, ISO 100. Under controlled conditions like this, it really shines. (sorry about the file size but I felt exporting at 80% jpg was necessary to preserve the fine detail)
-
Thanks Damon - interesting - I was once upon a time gearing up some work from home on a major project in Tajikistan, but that particular arrangement fell through :-) , so I never got there.
OK, back to photos, thanks for taking the trouble doing that. It would be good to see your LR settings in the Detail panel. Looking at the two images side by side, I don't see a huge difference in respect of sharpness and noise - the Piccure is a bit "crisper" all round, but detail not necessarily better (however, the detail in the ceiling of the other photo is impressive). Also not much difference in noise. These kind of tests are helpful for evaluating the merits of a new piece of software.
-
You're welcome Mark. The LR settings are nothing special - just as I described, plus highlights pulled down which I forgot to mention.
Attached is a 100% crop of this image: https://www.flickr.com/photos/damonlynch/21084192720/in/album-72157605715069583/
This attached image is my final copy, i.e. after I've done all post-processing, including the use of Topaz Detail. Again the lens is at 11mm in this image, using a tripod of course (actually the camera was just off the floor, with the legs as flat as they'd go). For this kind of image in my opinion Piccure in combination with other tools does a very fine job and is worth the effort.
-
Thanks Damon. Good sample and nice job. At the risk of trying your patience, may I ask how this same image would fare in LR?
-
Thanks Damon. Good sample and nice job. At the risk of trying your patience, may I ask how this same image would fare in LR?
To be honest I can't show you (at least not without a lot of work) - my regular workflow these days is to write over the top of the intermediary 16 bit TIFF, which in the case is the result of using LR first, SNS-HDR second. The reason I don't save the intermediary TIFFs is that I found I never use them should I need to revisit an old file. I never use them because the software I use to work on them always improves. For example in this image my guess is the very latest SNS-HDR would do a better job that the version I used to produce the image you see.
-
Understood Damon, thanks. Of course when putting that question I had in mind my very much more simplified workflow which is to remain in LR from ingestion to print unless I need to do something to a photo that LR simply can't handle because it doesn't have the tools or the tweaks within tools needed for the purpose. 99% of my files never leave the raw state and the history is permanently available as long as the LR catalog and XMPs are viable. It's a different approach. I have a bunch of plugins to LR that require conversion to a TIFF version, so I only deploy them when I expect real value-added. This approach, by the way, really forces one to drive LR's capabilities to the limits - kind of a stubborn mindset to stay with the mothership, which is instructive in its own right. Cheers.
-
I probably end up working with TIFFs much more than you do Mark, but I hear you loud and clear regarding the downsides. The loss of state is probably the biggest. I wish every step could be faithfully recorded in XMP, ready to be recreated at moment's notice with the same software, just like we can in LR. Alas!
-
And how do you fare with electricity to run the computers and accessories over there? When I was working in Nepal (on that very sector BTW) it was a recurrent nightmare.
-
And how do you fare with electricity to run the computers and accessories over there? When I was working in Nepal (on that very sector BTW) it was a recurrent nightmare.
Mostly good in the capital and larger cities. Much better than it used to be.
Sometimes I wonder about the possibility of photographic tours here -- because so few visitors come here, the chances of getting a unique shot no one has done before are considerably higher than in many other places. There's plenty of great scenery here, and some wildlife too, e.g. snow leopards. But the big problem, as always, is political instability and the possibility that a tour would be halted because of it.
-
Thank you both. If sharpening is the key incremental benefit I probably don't need this application, because I obtain very sharp photos from my Sony a6300+Zeiss lenses with moderate capture and output sharpening in LR. Some form of deconvolution is helpful to rescue detail when I mis-focus (i.e. pilot error), but I have never been satisfied with the ultimate quality (from a photographic perspective) of those rescue attempts (using demo versions of such software) and therefore usually trash photos that are out of focus when they should be in focus. Now that too may be pilot error - maybe I need to spend more time focusing on deconvolution technique (pun intended); but that will be for another day :-)
Mark,
I too am not impressed with the ability of deconvolution to restore out of focus images.
Hello Mark and Bill. I believe that's because deconvolution cannot restore out of focus images: defocus and dof cannot be modeled accurately as convolution with a blurring kernel. Of course one may apply any type of micro/mid/macro local contrast adjustments to attempt to improve the 'perceived sharpness' of such an image, such as USM or deconvolution of small/large radius. But by then it's no longer image restoration, it's simply an adjustment to make the final product look more pleasing.
In images captured with excellent technique there are very few quantifiable sources of blur left (diffraction, AA, pixels, various lens aberrations). Restoration by deconvolution works best when those are known (in type, if not quantity).
Also the a6300 does not appear to have an AA filter, its green channel hitting Nyquist around 0.25 cycles/pixel with an FE55mm at f/5.6.
Jack
-
Also the a6300 does not appear to have an AA filter, its green channel hitting Nyquist around 0.25 cycles/pixel with an FE55mm at f/5.6.
Jack
Interesting observation, because there is a great deal of uncertainty bred of unnecessary secrecy about this matter. In any case, with many of these cameras the most determinative variable for resolution of scene detail is the lens, not the sensor.
-
Mostly good in the capital and larger cities. Much better than it used to be.
Sometimes I wonder about the possibility of photographic tours here -- because so few visitors come here, the chances of getting a unique shot no one has done before are considerably higher than in many other places. There's plenty of great scenery here, and some wildlife too, e.g. snow leopards. But the big problem, as always, is political instability and the possibility that a tour would be halted because of it.
Interesting, but unfortunate, because primary concerns of a tour or workshop operator are reputational risk, liability and liability insurance. Anything that elevates those risks above a comfortable level would make it challenging to organize and market such events. Seems more like an environment for hardy souls with a high tolerance for risk and somewhat deep pockets to cover the uninsurable possibilities.
-
Hello Mark and Bill. I believe that's because deconvolution cannot restore out of focus images: defocus and dof cannot be modeled accurately as convolution with a blurring kernel.
Hi Jack,
Those are two different things and I can add some other detractors too. The point is that mathematically, convolution can be exactly reversed by a deconvolution, provided we can use high enough precision calculations. It's like dividing and multiplying, e.g. 1/3*3 = 1 , although some calculators insist on a different outcome due to lack of precision. Convolution is not much more than multiple multiplications and additions, and deconvolution does the inverse order of operations.
The real issue is finding the exactly correct Point Spread function (PSF) that caused the defocus , and then we also need to deal with noise. In practice we do not know the exact convolution PSF, so this then becomes a process of (user assisted) blind deconvolution with noise regularization. And even then, we will find that precision is not enough to fully restore all detail, although we can come a long way.
Attached I've added an example of a synthetic image, a noisy image would be harder to find a model for:
1. A sinusoidal star image,
2. An image convolved with a defocus kind of PSF,
3. The PSF,
4. The result after deconvolution.
I stopped at 10000 iterations of the special Deconvolution algorithm, because this small image already took some 47 minutes to complete, a full size image would take days.
t's not a perfect restoration, but it does achieve a significant restoration of what seemed to be lost. I'm just showing that there is more possible that acutance enhancement.
In images captured with excellent technique there are very few quantifiable sources of blur left (diffraction, AA, pixels, various lens aberrations). Restoration by deconvolution works best when those are known (in type, if not quantity).
Yes, but then there will still be other factors, like the demosaicing of the sensor signal and of the noise.
Also the a6300 does not appear to have an AA filter, its green channel hitting Nyquist around 0.25 cycles/pixel with an FE55mm at f/5.6.
The presence of aliasing artifacts will again make it harder to restore the original image signal.
Cheers,
Bart
-
Interesting observation, because there is a great deal of uncertainty bred of unnecessary secrecy about this matter.
Well, this is where the uncertainty about the presence or absence of an AA in the a6300 ends :)
(http://i.imgur.com/cHqbcvZ.png)
There is no AA signature in the MTF data of the a6300+FE55mm at f/5.6 (DPR studio captures), green channel in the vertical and horizontal directions shown.
Jack
-
Thanks Jack; I'm not familiar with the technique you are using to illustrate this point. Perhaps if you don't mind supplying a few lines of explanation about this test, and putting up a chart from a camera that is known to have an AA filter it would be of educational value to me and perhaps other readers here. Thanks in anticipation.
-
Thanks Jack; I'm not familiar with the technique you are using to illustrate this point. Perhaps if you don't mind supplying a few lines of explanation about this test, and putting up a chart from a camera that is known to have an AA filter it would be of educational value to me and perhaps other readers here. Thanks in anticipation.
Hi Mark,
That's a plot of the spatial frequency response (sharpness at varying levels of detail) of the shown camera and lens as captured. If a current implementation of an AA is present its job is to make the curve tend towards zero around +/- 0.7 cycles/pixels, so that aliasing-inducing frequencies above Nyquist (0.5 c/p) are minimized. You can see the difference an AA makes in this well known duo (Canon 5DS with AA, 5DSR sans AA):
(http://i.imgur.com/Mnd8dcN.png)
The earlier curve of the a6300 does not show the effect of an AA filter.
Jack
-
Hi Jack,
Those are two different things and I can add some other detractors too. The point is that mathematically, convolution can be exactly reversed by a deconvolution, provided we can use high enough precision calculations. It's like dividing and multiplying, e.g. 1/3*3 = 1 , although some calculators insist on a different outcome due to lack of precision. Convolution is not much more than multiple multiplications and additions, and deconvolution does the inverse order of operations.
The real issue is finding the exactly correct Point Spread function (PSF) that caused the defocus , and then we also need to deal with noise. In practice we do not know the exact convolution PSF, so this then becomes a process of (user assisted) blind deconvolution with noise regularization. And even then, we will find that precision is not enough to fully restore all detail, although we can come a long way.
Attached I've added an example of a synthetic image, a noisy image would be harder to find a model for:
1. A sinusoidal star image,
2. An image convolved with a defocus kind of PSF,
3. The PSF,
4. The result after deconvolution.
I stopped at 10000 iterations of the special Deconvolution algorithm, because this small image already took some 47 minutes to complete, a full size image would take days.
It's not a perfect restoration, but it does achieve a significant restoration of what seemed to be lost. I'm just showing that there is more possible that acutance enhancement.
Hi Bart that's a good demo and I agree with your comments on aliasing and noise.
However you have shown a simulation of convolution and deconvolution with a known PSF in 2D, a tractable problem. On the other hand my understanding is that in practical photography situations defocus cannot be modeled as a convolution with a PSF because it cannot be known without having a precise 3D map of the image in the first place: in practice the scene's depth/phase information within the region of interest gets scrambled to the point of no return. So any attempts at recovering the original image need to make strongly simplifying assumptions that when applied almost never satisfy - as observed by Mark, and I am sure most anybody who has tried to recover anything other than very mild defocus with current technology. Even then, my guess is that other than in ideal situations the improvement is generically pleasing but not accurate.
Also interesting is the opposite problem, trying to add focus blur to an image and make it look credible. Check the background: credible?
(http://i.imgur.com/OuwkKaz.jpg)
Not to this critical peeper. However, interestingly, that we can recover quite accurately.
Jack
-
Hi Mark,
That's a plot of the spatial frequency response (sharpness at varying levels of detail) of the shown camera and lens as captured. If a current implementation of an AA is present its job is to make the curve tend towards zero around +/- 0.7 cycles/pixels, so that aliasing-inducing frequencies above Nyquist (0.5 c/p) are minimized. You can see the difference an AA makes in this well known duo (Canon 5DS with AA, 5DSR sans AA):
The earlier curve of the a6300 does not show the effect of an AA filter.
Jack
Thanks very much Jack - very helpful. Now I should feel happier I bought an a6300? :-)
-
Thanks very much Jack - very helpful. Now I should feel happier I bought an a6300?
Yes, unless you photograph a lot of man-made stuff like fabric and architecture. In that case no :)
-
Would applying this software to images taken with my Leica S2 70 mm lens show an improvement ?
-
Just a note:
This upgrade is a bust for me, and I am going to drop back I guess to prior release. Processing times are as slow as before on IQ100, IQ260/160 and Nikon D810.
The new app is supposed to be able to run some cycles with the GPU, (in may case identical GTX970 4GB cards). but the GPU processing part fails each time with an insufficient memory error. Not sure what that is happening, as per monitoring software, during the processing time the GPU is only using max 855mb not even 1GB. This is running both from Photoshop CC and standalone. Win 7 and Win10 32GB of ram on both machines.
At times the whole process aborts on my older machine when run under photoshop CC (I prefer this with a layer to work on fading sharpening is needed)
I realize there are way too many machine setups for Piccure to test everything. But I am using pretty high end Asus boards (which have never had an issue) and nvidia cards latest drivers and not cheap either.
So if you are running this type of setup, you may want to skip this upgrade especially if you using the files from any of the cameras I referenced. Processing time on a IQ100 16 bit image (2 part pano) is 25 minutes. Yup 25 minutes, just as long as the older version processed. And without a doubt during the run the GPU part will fail.
Overall still a bit disappointed as many know this software is not inexpensive and you have to pay per machine, I have several licences. But even with all the problems, I will still use it (older version) as the results speak for themselves. (at least for me).
Paul C
-
Dear Paul,
thanks for your email.
We have received your support ticket and we have replied yesterday night at 9:40 pm within 50 mins and today (Sunday) within 15 mins. The statements you make are not entirely correct unfortunately - there is about a 50% speedup even without using the GPU. We were still waiting for the image you promised us today (Sunday) and promised to take a look - so far we have not received the image. Our developers don't work 24/7 either - it's just the founders...
There also is the possibility to deactivate the GPU as described in detail in the handbook. Go "Menu/Deactivate GPU". Version 3.0 is still more than 50% faster on a CPU in this case than version 2.5.
We have fully refunded your order and kept your license alive. However, given that we tried really hard to fix this very specific problem of yours at 10 pm Saturday night and on Sunday - and this is the kind of communication you prefer, we'd rather not continue a business relationship going forward. I can understand that many will find this "bad and evil". But we do this because we are passionate about the technology and our product and we usually love the feedback and interaction with our customers. Money is not our primary incentive. We gave you the software for free, but we will not support this case going forward.
Have a nice Sunday.
Lui
-
Dear Paul,
thanks for your email.
We have received your support ticket and we have replied yesterday night at 9:40 pm within 50 mins and today (Sunday) within 15 mins. The statements you make are not entirely correct unfortunately - there is about a 50% speedup even without using the GPU. We were still waiting for the image you promised us today (Sunday) and promised to take a look - so far we have not received the image. Our developers don't work 24/7 either - it's just the founders...
There also is the possibility to deactivate the GPU as described in detail in the handbook. Go "Menu/Deactivate GPU". Version 3.0 is still more than 50% faster on a CPU in this case than version 2.5.
We have fully refunded your order and kept your license alive. However, given that we tried really hard to fix this very specific problem of yours at 10 pm Saturday night and on Sunday - and this is the kind of communication you prefer, we'd rather not continue a business relationship going forward. I can understand that many will find this "bad and evil". But we do this because we are passionate about the technology and our product and we usually love the feedback and interaction with our customers. Money is not our primary incentive. We gave you the software for free, but we will not support this case going forward.
Have a nice Sunday.
Lui
Lui,
Sorry that you interpreted my post as a negative statement. I thought I was very clear in both of my posts on this subject that I felt that your sharpening solution was by far the best I have used. However on an open forum such as this one I also felt that other users who have the same equipment setup as mine or close, may want to do more testing as you do offer a free trial period.
Personally, I started using your software (older version) sometime in the later part of 2015, and I based on the results, I have totally switched over to it, and made it quite clear I was more than willing to pay for multiple licenses due to the results I was able to obtain.
I also never asked for a refund, and would rather continue to pay for the software.
You are correct that I made my previous post prior to sending you the file to test on, so that may have been presumptuous on my part, but I have also worked on issue for quite a while now as I knew that if I could get the GPU portion to work, that the that the new version's improvement in processing times would be a huge benefit to my workflow.
Sorry that my comments were seen as an attack? or overtly negative as overall I would never want to stop using Piccure+ for sharpening for any of my camera files.
Edit: I also misstated my license cost as I did have the 3 actuation's licence which when I purchased it was $109.00 US. I did not purchase 2 licences.
Sincerely
Paul Caldwell
-
Paul,
What was the difference between your first post and the last post - image file size ? - where the speedup went from what was expected to very little ?
The response from the developer/founder was a bit odd given what was actually written by you in the forum thread.
-
Wow.
I referred to my tests with the previous version of piccure to remind myself how it compares with other choices. The black edge lines seem more obvious than I remembered.
https://www.dropbox.com/s/ly4cbdsyiygjj9i/Sharpen%20Tests.zip?dl=0
I would like to compare the previous version of Piccure with the latest version to see how it has improved, but my trial for the previous version ran out several months ago. It would be great if I could run both side by side, or at least one version right after another so that I can see the 50% increase in speed for myself and see if the black edge line issue has been improved.
-
It would be interesting to know how this product compares with DXO, which on the face of it offers similar capabilities, albeit using measured profiles, rather than calculated ones.
-
It would be interesting to know how this product compares with DXO, which on the face of it offers similar capabilities, albeit using measured profiles, rather than calculated ones.
When there is not a lot of image noise, Piccure produces notably sharper results than DxO or even Canon's DPP, the latter of which produces excellent results with its digital optimizer applied to Canon wide angle lenses.
-
Paul not sure why your machine is not working well, I have an ASUS motherboard and Nvidia GPU cards and the upgrade works just fine. I do not have CC but using PS6 and Lightroom 6. I have noticed a speedup, not measured it quantitatively but it seems faster. Still experimenting with the new version of Piccure+ but so far it is nice upgrade. As I only have a 1 machine license I am not running it on my laptop. Also using Win10, I suspect maybe there is some other conflict in your systems, maybe AV or something like that.
Alan
-
Hi Alan.
I believe it's a file size issue. Having same issues on both my main PCs where the GPU error stating not enough memory happens. One PC is brand new all new ASUS.
From watching the monitoring software the max ram that is being used is 855 to 1 GB On both cards and they are 4 GB.
Processing falls beck to the main system.
These are Phaae One files IQ260 and 100 all 16 bit sharpening with a layer so big files. But I get the same problem if I drop back to 8 bit.
But results look good once the file has finished.
Paul C
-
Well, I have had a quick test comparison with a Lightroom-processed file (Sharpening set at 50/0.7/70/20).
The result is hard to judge, because, although it produces slightly higher apparent sharpness, it is not a full raw processor, so other factors that make the resulting picture pleasing are not directly available:
- lens geometry corrections (eg, when embedded in the DNG by the camera)
- macro contrast / exposure / saturation
So you end up having to process the raw in Lightroom / ACR anyway, in which case, I assume, the piccure+ will be less effective because it is having to work with a partially processed TIFF, rather than a RAW file.
It is possible that it may be worth it for some pictures that you want to print very large but, for screen output, I don't think that the extra fiddling is worth the detour.
In conclusion, this seems like an interesting technology preview, but not really practicable.
What am I missing?
PS: Is there a way of uninstalling the program on a Mac without leaving debris? It does not seem to come with an uninstall app...
-
Hi,
The result is hard to judge, because, although it produces slightly higher apparent sharpness, it is not a full raw processor, so other factors that make the resulting picture pleasing are not directly available:
- lens geometry corrections (eg, when embedded in the DNG by the camera)
- macro contrast / exposure / saturation
So you end up having to process the raw in Lightroom / ACR anyway, in which case, I assume, the piccure+ will be less effective because it is having to work with a partially processed TIFF, rather than a RAW file.
piccure+ does one thing: correct optical aberrations and micro-shakes that cause an image to look slightly (or a bit more) blurred. It is a one trick pony only for this purpose. However, you can apply most corrections associated with the lens of a camera without any downside (as explained in detail in the handbook):
"You can pre-process the image before using piccure+ with the following operations:
- Process the RAW file.
- Correct chromatic aberrations (thus saving piccure+ processing time).
- Correct vignetting.
- Correct lens distortion.
- Denoise with DxO PRIME (several users found the results very good). Usually denoising prior to using piccure+ is
a bad idea.
Please do not:
- Sharpen the image (e.g. SmartSharpen or "Correct Lens Softness“).
- Crop the image.
- Adjust saturation, brightness, or contrast (at least do not adjust them significantly)."
It makes little sense to embed all the processing functionality a user already finds in excellent programs such as Photoshop or Lightroom. piccure+ complements existing solutions and does not replace them. piccure+ internally converts images from RAW to TIF prior to processing - applying adaptive deconvolution to RAW data would put the software more in the 1000+ USD price range given the efforts needed...
If you import images with another RAW processor, kindly keep in mind that many (most) have default sharpening applied (e.g. ACR when importing through Photoshop) - which is generally not a good idea (also for other processing).
It is correct that if you have a 72 dpi screen the results may be less impressive/visible as compared to 300 dpi print. Mac users have a real advantage with the Retina displays - and once you experience Retina, there is no "going back". Unless the computer screen is calibrated it also does not make so much sense to do much of image editing. Also keep in mind that many browsers and many Apple devices (e.g. iPhone) have problems displaying color spaces other than sRGB (e.g. just assume it is sRGB) therefore "ruining" the image. So if you just view images on a screen (and show it to others), it is probably the best idea to stay in sRGB... If you are going for print that's a different story. It depends on the lens and camera used. If you shoot f5.6 probably there is really little benefit of running piccure+...
If you want to remove it, remove the piccure+ folder and HD/Library/Application Support/IIS2
Best,
Lui
Co-Founder
-
I remain curious about the recommendation not to crop before using piccure.
It is tempting to crop before the process because you can potentially save lots of time. I have wondered; Is there an existence of symmetrical "distortions" that exists in the original full frame that is anticipated by the mathematics employed by piccure? Does cropping before hand somehow trick piccure into making corrections in an unanticipated and, or, less than ideal way?
Thank you.
-
I remain curious about the recommendation not to crop before using piccure.
It kills the opportunity to detect symmetrical distortions and sharpness fall-off. Center of image is then no longer the center of the opical projection. BTW, this also means that P+ is not very well suited for tilt and shift lenses, or view cameras.
It is tempting to crop before the process because you can potentially save lots of time. I have wondered; Is there an existence of symmetrical "distortions" that exists in the original full frame that is anticipated by the mathematics employed by piccure? Does cropping before hand somehow trick piccure into making corrections in an unanticipated and, or, less than ideal way?
Apparently that is the case. In a 'Blind' restoration it is not always possible to detect the optical center of an image.
Cheers,
Bart
-
Thank you Bart.
-
I have also wondered about the recommendation to dismiss the use of piccure at or beyond f5.6.
I became preoccupied with the dark edges that the process was synthesizing, and now I realize that I was only working on images that were already sharp. I have fast lenses but usually shoot images that I want to seem sharp at f5.6 or f8.0. Using deconvolution to sharpen further seems appropriate when I have the goal of sharp depiction. When I shoot at f1.2 f2.0 etc. I am doing so because I want the image to have a noticeable or remarkable soft fall off from the focus point.
These ideas make me wonder if the appearance of the synthesized dark edges is minimized when starting with softer images that do not already have well defined edges and if it is true that I should just regard the recommendations seriously and dismiss my aspiration to use piccure to further sharpen images that are already relatively sharp.
-
I remain curious about the recommendation not to crop before using piccure.
It is tempting to crop before the process because you can potentially save lots of time. I have wondered; Is there an existence of symmetrical "distortions" that exists in the original full frame that is anticipated by the mathematics employed by piccure? Does cropping before hand somehow trick piccure into making corrections in an unanticipated and, or, less than ideal way?
Thank you.
I mean, you can - but you do simply remove a lot of data which piccure+ can otherwise rely on. Usually cropping is done for aesthetical reasons, not in order to "dismiss valuable data". So the general recommendation is not to crop an image - just like shoot RAW as to avoid the mess JPGs create in terms of data preservation...
BTW, this also means that P+ is not very well suited for tilt and shift lenses, or view cameras.
Apparently that is the case. In a 'Blind' restoration it is not always possible to detect the optical center of an image.
We didn't say that - and as a matter of fact many professional photographers use piccure+ for tilt/shift lenses, particularly in architecture. piccure+ does not rely on any lens profiles and can thus also accommodate complexities stemming from T/S.
I have also wondered about the recommendation to dismiss the use of piccure at or beyond f5.6.
We did not say "at or beyond" but "at" f5.6. The reason is simply that as a rule of thumb at f5.6 the vast majority of all lenses is sharpest, the psf is coming quite close to a Gaussian shape and rather uniform across the image - thus diminishing the advantages of piccure+ over other solutions. Beyond f5.6 diffraction can start to become a problem (in real life at faster apertures than "diffraction limited" would suggest for a sensor size) - and can often not be corrected sufficiently with other solutions. So you could think about the "utility function of using piccure+ over other solutions" as a U-shape with the bottom at aperture f5.6. For faster apertures it makes more sense and for slower apertures it starts making sense again probably from 6.3 on(give or take, rule of thumb).
I became preoccupied with the dark edges that the process was synthesizing, and now I realize that I was only working on images that were already sharp. I have fast lenses but usually shoot images that I want to seem sharp at f5.6 or f8.0. Using deconvolution to sharpen further seems appropriate when I have the goal of sharp depiction. When I shoot at f1.2 f2.0 etc. I am doing so because I want the image to have a noticeable or remarkable soft fall off from the focus point.
... the nice thing about piccure+ is that it cannot make the "bokeh" you intentionally created sharp. But actually it strengthens the effect as the "sharp" areas become crisper, whereas the "out of DOF areas" stay "blurred". It does not shift focus.
These ideas make me wonder if the appearance of the synthesized dark edges is minimized when starting with softer images that do not already have well defined edges and if it is true that I should just regard the recommendations seriously and dismiss my aspiration to use piccure to further sharpen images that are already relatively sharp.
We don't track what people do with our software, so I cannot say for sure. But I would rather say the opposite is the case: piccure+ is rather used for very sharp images to make them a little bit better. Not to make "bad ones bearable"...
Best,
Lui
-
We didn't say that - and as a matter of fact many professional photographers use piccure+ for tilt/shift lenses, particularly in architecture. piccure+ does not rely on any lens profiles and can thus also accommodate complexities stemming from T/S.
Good to know, thanks Lui. Will have to test that though...
Cheers,
Bart
-
We didn't say that - and as a matter of fact many professional photographers use piccure+ for tilt/shift lenses, particularly in architecture. piccure+ does not rely on any lens profiles and can thus also accommodate complexities stemming from T/S.
...I remembered piccure+ was with T/S lenses - it was actually Keith Cooper from Northlight Images:
http://www.northlight-images.co.uk/reviews/plugins/piccure_plus.html
There are a few sample images towards the end of the article - taken with a TS-E17mm shift lens... Keep in mind the review is almost a year old and based on v2...
Best,
Lui
-
I downloaded and tried the latest version. It seems much improved, and I am excited about the quality. I didn't see the dark edges, but I have not specifically tried a high contrast edge circumstance where, in the past, the dark lines seemed most obvious.
I gave piccure+ another head to head comparison to what I think is the best use I can make of Topaz Infocus and it seems like my work with piccure+ has moved my state of the art forward. The quality I am appreciative of is that it seems to sharpen as good or better than my current workflow, with less build up of grain or noise.
Now if I can just get over the idea of paying an extra $40 to have the luxury of easily moving back and forth between 2 computers. At the very least it would be nice if piccure's website clearly stated if you can start with the single license and upgrade for an extra $40 after you get tired of deactivating and activating when you switch CPUs.
-
Great to hear you gave it another try.
Now if I can just get over the idea of paying an extra $40 to have the luxury of easily moving back and forth between 2 computers. At the very least it would be nice if piccure's website clearly stated if you can start with the single license and upgrade for an extra $40 after you get tired of deactivating and activating when you switch CPUs.
That's no problem. If you want to upgrade your license later, send us an email to: [email protected]. piccure+ is not a floating license (e.g. like the Adobe photography plan which comes as a subscription and you can switch between computers as much as you want) but is limited to a certain number of computers. You can deactivate a computer if you exchange it - but not indefinitely. For "one-time purchases" these "floating" licenses are not common generally either.
Best,
Lui