• Please use real names.

    Greetings to all who have registered to OPF and those guests taking a look around. Please use real names. Registrations with fictitious names will not be processed. REAL NAMES ONLY will be processed

    Firstname Lastname

    Register

    We are a courteous and supportive community. No need to hide behind an alia. If you have a genuine need for privacy/secrecy then let me know!
  • Welcome to the new site. Here's a thread about the update where you can post your feedback, ask questions or spot those nasty bugs!

Topaz Labs InFocus, a new deconvolution based sharpening plugin

Hi folks, FYI,

http://www.topazlabs.com/infocus/

Finally a dedicated sharpening tool that also works on 64-bit platforms.
It also gives some control to reduce potential sharpening artifacts. When used on a Luminosity blending Layer, it gives an immense amount of control over what gets sharpened, and how much.

While there is no option to use user-defined point spread functions (PSFs), there is a mode that attempts to construct such a PSF from image content. It does need some (blurred) edges in the image to estimate the PSF. It's slightly less effective than a custom PSF, but still gives very acceptable results, e.g. restoring sharpness from diffraction blur or other complex PSFs.
Example of how to use blur estimation.

Cheers,
Bart
 
Thanks for that, Bart. I'll have to try it. The blur reduction looks amazing.

Hi Murray,

Only the deconvolution type of algorithms can restore sharpness from blur. All USM based methods just enhance edge contrast. I've been playing with InFocus a bit, to get familiar with it, and overall I like what I can do with it. I have yet to compare it in a more rigorous way with the alternatives I already use, but the 64-bit capability is a great workflow benefit.

Deconvolution does tend to increase noise as well (although less than real image detail), so one needs to be careful, and I prefer to use it on a luminosity blending layer to control what and how much is ultimately sharpened.

PK sharpener is a useful tool, but it is just a way to intelligently select presets with the standard Photoshop tools. I had hoped they would also add real deconvolution sharpening to enhance Photoshop capabilities, but it seems that adding 64-bit platform capability was the main goal.

Cheers,
Bart
 
It's on sale, so I bought the plug-in.

After experimenting with several photos to get a feel for how it worked, I ran it through its paces on an image taken today. At some point while moving the sliders the plug-in stopped responding - couldn't even toggle between preview and original. Also couldn't make the OK or Cancel buttons respond. Finally hit escape, saved the image, and then rebooted the PC. I reopened the original image and launched the Infocus plug-in. It still just hangs.

Hmmm... I'm hoping it's just cockpit error so I'll fiddle around a bit before contacting Topaz.
 
It's on sale, so I bought the plug-in.

After experimenting with several photos to get a feel for how it worked, I ran it through its paces on an image taken today. At some point while moving the sliders the plug-in stopped responding - couldn't even toggle between preview and original. Also couldn't make the OK or Cancel buttons respond. Finally hit escape, saved the image, and then rebooted the PC. I reopened the original image and launched the Infocus plug-in. It still just hangs.

Hmmm... I'm hoping it's just cockpit error so I'll fiddle around a bit before contacting Topaz.

Hi Tom,

From what I've seen on the Topaz forums, they are very keen on supporting their products. Just leave them a message, if you can by describing which sliders were being used when the issue surfaced. I'm confident that you'll get a tip before long on how to unlock the plugin. Perhaps just re-installing it will do the trick? I did see that there might be an issue with some AMD processor equipped machines, for which they already issued a fix.

I've been getting acquainted with the InFocus functionality myself, and things are becoming predictable. I've also tried a stress test by resampling an image from my 1Ds3 to 300%. and sharpening that. That obviously was a bit slower, but it was handled without problems on my 64-bit Windows 7 Ultimate machine.

Cheers,
Bart
 

Asher Kelman

OPF Owner/Editor-in-Chief
Bart,

In the flow of things, where would a blur before sharpening occur if one is going to use this too?

Asher
 
In the flow of things, where would a blur before sharpening occur if one is going to use this too?

Hi Asher,

I'm not sure I understand the question but there are numerous occasions for blur to occur. From an optical point of view, the residual lens aberrations + diffraction, and the AA-filter and the discrete sampling will cause a specific blur point spread function. There may also be a certain amount of defocus involved, and/or camera shake.

Then on top of it all there is the Raw data reconstruction of the Bayer CFA, which poses a trade-off between softness and artifacts.

What deconvolution sharpening does, is attempt to regroup all the photons and bits that wandered off to neighboring pixels, for each pixel, back to their original position. That is a daunting task, and it won't be perfect, but we can in many cases get close to our ideal.

This deconvolution sharpening can also be applied to upsampled images, since the process of upsamling introduces a certain amount of softness. I did a stress test and resized an image to 300% and sharpened that result, which improved significantly and would produce a much better large format print.

Cheers,
Bart
 

Doug Kerr

Well-known member
Hi, Bart,

Hi Asher,

I'm not sure I understand the question . . .

It sounds to me as if Asher is speaking of intentionally applying a blur to the image before applying the deconvolution sharpening.

I seem to recall that there are cases where intentional application of blur is suggested as one step of some process (not for the artistic propose of making the resulting image blurry), but I can't remember what it is, or why that is said to be useful.

It may have to do with downsampling an image.

Best regards,

Doug
 

Alain Briot

pro member
Topaz in focus adds blur estimation and removal to sharpening. I'm not sure if this was already available in other sharpening packages. In any case, its only useful if your photographs are blurry to start with. otherwise I'm not sure if Infocus improves on previous sharpening packages such as Photokit (my personal favorite) except that it works in 64 bits, which the next version of Photokit will as well when it is released end of this year.
 
I tried Topaz InFocus for blur reduction on a few images. These are just my first quick impressions and may be misplaced.

My first try was a live music image that I was wanting to print today. It concerned me because it wasn't as sharp as I would want, probably due to misplaced focus rather than motion blur. It was an action shot taken on a Nikon D3 at 3200ISO and I couldn't use anything like the blur radius shown in the video or I got excessive reticulation. So I eliminated all noise, found a setting where I could get some blur reduction and reintroduced noise from an unmodified layer. The effect of that was an increase in sharpness (or reduction in blur) but after I toned that back in areas where it didn't particularly help, the effect was relatively subtle. I then applied sharpening using Topaz Detail2 and got better results than I had with PK Sharpener but then I was only using my standard method without adjustment there, so I could probably have gotten better sharpening results with PK sharpener as well.

I have just purchased a 14-24mm and applied Topaz InFocus blur reduction to some quick test images I had taken rather casually hand-held. In some cases there was significant improvement, in others nothing much. It appears to be useful but perhaps not quite the magic bullet the video may imply. Perhaps it may work best with certain types of blur involving straight lines (Topaz examples of launch, numberplate).

We will see what the new version of PK Sharpener brings soon enough. A significant advantage to me with the current one is that I can quickly run an action for my standard method and quickly adjust it with layer opacities and blend if. I may be wrong but it appears to me that Topaz sharpening does not lend itself as well to semi-automation.

Regards,
Murray
 
I tried Topaz InFocus for blur reduction on a few images. These are just my first quick impressions and may be misplaced.

My first try was a live music image that I was wanting to print today. It concerned me because it wasn't as sharp as I would want, probably due to misplaced focus rather than motion blur. It was an action shot taken on a Nikon D3 at 3200ISO and I couldn't use anything like the blur radius shown in the video or I got excessive reticulation. So I eliminated all noise, found a setting where I could get some blur reduction and reintroduced noise from an unmodified layer. The effect of that was an increase in sharpness (or reduction in blur) but after I toned that back in areas where it didn't particularly help, the effect was relatively subtle. I then applied sharpening using Topaz Detail2 and got better results than I had with PK Sharpener but then I was only using my standard method without adjustment there, so I could probably have gotten better sharpening results with PK sharpener as well.

Hi Murray,

Thanks for the feedback. In my limited, but growing, experience with InFocus it is usually best to use the generic type of blur removal. The plugin will produce artifacts much faster than FocusMagic, and while the artifact reduction slider will reduce them it also adds softness (mostly to non edge areas). I can imagine that image noise will frustrate the restoration possibilities, but that will also be the case with alternative remedies.

I have just purchased a 14-24mm and applied Topaz InFocus blur reduction to some quick test images I had taken rather casually hand-held. In some cases there was significant improvement, in others nothing much. It appears to be useful but perhaps not quite the magic bullet the video may imply. Perhaps it may work best with certain types of blur involving straight lines (Topaz examples of launch, numberplate).

I've also tried recovering some more pathetic cases, and InFocus struggles more than FocusMagic, but since FM will not run on all systems that's perhaps a moot point. With reasonable blur amounts the improvement with InFocus can be rather good. Just avoid creating significant artifacts that require significant (>30) artifact reduction amounts. Generally well behaved images will only need something like 0.70 -1.00 radius generic deblurring. It can also be useful to cascade 2 runs of InFocus with moderate settings.

We will see what the new version of PK Sharpener brings soon enough. A significant advantage to me with the current one is that I can quickly run an action for my standard method and quickly adjust it with layer opacities and blend if. I may be wrong but it appears to me that Topaz sharpening does not lend itself as well to semi-automation.

PK Sharpener is useful because it makes the process of selecting built-in Photoshop functionality easier, but it doesn't add sharpness, just edge contrast. Deconvolution sharpening, such as from InFocus, depends on the exact characteristics of blur, which requires a bit of tinkering to remove. There will always be some blur, e.g. from the Raw conversion process (do switch sharpening and noise reduction off during that stage!). Bayer CFA demosaicing involves trade-offs between softness and artifacts, so there is always something left to really sharpen. Lenses+diffraction and AA-filters will also add their blur signature. So even with good technique there is a potential for improvement.

Cheers,
Bart
 
Hi Bart,

You're right about Topaz making an effort to resolve my hanging program problem. First attempt was to download the latest drivers for my Wacom tablet. This didn't help. Next attempt was a new version of the program to download and try. This too didn't work. At this point I am confident they will either eventually resolve the issue or refund the purchase price. Time will tell.
 
Hi Bart,

You're right about Topaz making an effort to resolve my hanging program problem. First attempt was to download the latest drivers for my Wacom tablet. This didn't help. Next attempt was a new version of the program to download and try. This too didn't work. At this point I am confident they will either eventually resolve the issue or refund the purchase price. Time will tell.

Hi Tom,

Thanks for the feedback.

Indeed. They are apparently trying to resolve the issue together with Wacom. There seems to be a conflict with the Wacom drivers, and I read that an issue with some AMD systems is identified:
http://www.topazlabs.com/forum/product-technical-support/6438-wacom-tablet-compatibility-problem-fix-download-links.html.

Things will be resolved somehow. It's at these times that companies can prove what they're worth.

Cheers,
Bart
 
Hey Bart, success!

Topaz issued a Wacom fix that required replacing all of their programs. DeNoise worked fine previously so I didn't bother installing the new Wacom version. This proved to be a mistake as the new version of Infocus couldn't work without updating both programs. So, that's now been done and all is good.

This new plugin could not have come along at a better time of year, and I'm looking forward to a lot of experimentation.

An early usage attempt, indicating nothing particular except that it was damn cold this morning hereabouts -

130652832.jpg


Much yet to learn...
 
Hey Asher, I'll eventually post before and after shots after I get the hang of this program. At the moment it's way too soon to be doing that sort of thing.
 

Asher Kelman

OPF Owner/Editor-in-Chief
Let me add to this the revelation that one does not need Photoshop to run the Topaz in focus or any other of Topaz image enhancement plugins.

pieces.jpg


Free program avaialble here for either Mac 10.4, Mac 1.5+ or Win 32/64

With this software, one can work in one's favorite image browser/ editor and then call up the Topaz features at your whim and fancy!

This brings tremendous increase in capability to iphoto, Lightroom or Aperture. So Bart and tom and anyone else, see if you can download the free Fusion software and try out the In focus from within Lightroom, iPhoto or Aperture.

Asher
 
Does anyone have examples where InFocus provided better sharpening benefits than the usual USM approach? I've tried with several photos with out of focus elements, but have yet to find one where the deconvolution approach offered by Topaz is any better than my tried and true USM recipes.

I am not saying the sharpening InFocus results are not good - they are - but I've yet to see evidence of their advantage over USM. It is quite possible that I don't yet understand the program, and so this just might be an example of a learning curve that needs a bit more climbing. I am all ears and eyes.
 

Doug Kerr

Well-known member
Hi, asher,

Let me add to this the revelation that one does not need Photoshop to run the Topaz in focus or any other of Topaz image enhancement plugins.
If these are plugins, what do they plug into?

If they can run on their own (not as plugins to Photoshop or some other image processing application) why are they described as "plugins"?

For example, PTLENS has two manifestations, as a plugin that works with Photoshop and many other image processing apparitions, and as a free-standing application (not described as a plugin).

Perhaps I am misunderstanding what you are saying.

Best regards,

Doug
 
Hi Doug

If you have an image in Photoshop, you can duplicate the background layer and apply Topaz adjustments to it through the filter menu. Hence, a plug-in.

In Lightroom, if you define Topaz Fusion Express and a second editor, you can launch from an image to any of the Topaz products and it comes back in Lightroom as a TIFF with a single layer. You can of course also round trip through Photoshop and include Topaz enhancements.

Regards,
Murray
 

Doug Kerr

Well-known member
Hi, Murray,

If you have an image in Photoshop, you can duplicate the background layer and apply Topaz adjustments to it through the filter menu. Hence, a plug-in.
Yes.

In Lightroom, if you define Topaz Fusion Express and a second editor, you can launch from an image to any of the Topaz products and it comes back in Lightroom as a TIFF with a single layer.
Aha!

In the case of such a "routing", is the serving product properly thought of (in that circumstance) as a "plugin"?

Of course, it is not exactly a "free-standing" app in that case either, since one must have another app (i.e., Lightroom) to access it.

Thanks so much.

By the way, I got Topaz InFocus late last night but have not yet installed it.

Best regards,

Doug
 
In the case of such a "routing", is the serving product properly thought of (in that circumstance) as a "plugin"?

Hi Doug,

Not really as a plugin for e.g. Lightroom, but it is the same plug-in that is called from Photoshop or from Fusion Express, so you could call it a plugin to Fusion Express.

Cheers,
Bart
 

Doug Kerr

Well-known member
Hi, Bart,

Hi Doug,

Not really as a plugin for e.g. Lightroom, but it is the same plug-in that is called from Photoshop or from Fusion Express, so you could call it a plugin to Fusion Express.
Yes, of course. That notion had begun to bubble to the surface of my mind before, but it never made it.

Thanks.

Best regards,

Doug
 

Doug Kerr

Well-known member
For what it's worth, I was able to install Topaz InFocus 1.0.0 to run as a plugin in my favorite editor, Picture Publisher 10.

That app nominally can use any plugin accessed via an "8bf" plugin file, but sometimes the plugin won't actually work. Here it seems to.

Way neat!

Best regards,

Doug
 
Does anyone have examples where InFocus provided better sharpening benefits than the usual USM approach? I've tried with several photos with out of focus elements, but have yet to find one where the deconvolution approach offered by Topaz is any better than my tried and true USM recipes.

Hi Tom,

Everything depends on the image data that the plugin can work on. USM just 'enhances' edge contrast, whereas deconvolution really redistributes the blur back to the initial pixel positions.

After significant testing of InFocus on various images, I've come to the conclusion that InFocus has some way to go (although I'm confident that TopazLabs will improve the performance). I'm confident that it can do better, because I have several alternative procedures to compare with, and thus know that better results can be achieved with deconvolution.

The typical difference betweeen USM and deconvolution can be seen in lower contrast areas, the local micro-contrast material structures. Deconvolution will typically boost micro-contrast detail better than the noise, thus achieving a higher signal/noise (S/N) ratio. High contrast edges are easy, real resolution is more difficult.

Having said that, InFocus uses a new type of algorithm (FYI Albert Yang is the founder of/a developer at Topaz Labs), one that apparently can easily produce artifacts in addition of detail. That means that currently it is adequate at recovering modest amounts of blur, but the more extreme cases are not handled as well as can be by some of the alternatives, notably FocusMagic (which however seems to lack any progress and may not run on various platforms).

I am not saying the sharpening InFocus results are not good - they are - but I've yet to see evidence of their advantage over USM. It is quite possible that I don't yet understand the program, and so this just might be an example of a learning curve that needs a bit more climbing. I am all ears and eyes.

Indeed, the results can be good with relatively well behaved images, and there are more controls available to skin the blurry cat. However, on images that need more help (the questionable quality ones), there needs to be a boost in performance before InFocus can be crowned as being the new King.

Cheers,
Bart
 
After significant testing of InFocus on various images, I've come to the conclusion that InFocus has some way to go (although I'm confident that TopazLabs will improve the performance). I'm confident that it can do better, because I have several alternative procedures to compare with, and thus know that better results can be achieved with deconvolution.

Apparently the feedback from the Version 1.0 introduction sent them a clear message:
http://www.topazlabs.com/forum/general-discussion/6642-infocus-sale.html#post38343

There are improvements expected for InFocus in a couple of months from now. They are now focusing their attention on a new "Remask" version, which might also be very intersting for those who need to do a lot of mask creation, including hair masks.

Cheers,
Bart
 
Hi Bart, thanks for the link to the InFocus discussion. I now have some miles on the deconvolution tires, so to speak, and am beginning to get a sense of the program's strengths. The approach has promise, and knowing Topaz has upgrades in the pipeline increases my confidence significantly.
 
BEFORE and AFTER comparison

Looking forward to the improvements here as well.

I was culling my images over the weekend, ran across this series of images that I shot under difficult circumstances that I was unprepared for (ie, no tripod).

I ran one of them through Topaz InFocus and using the Estimate Blur function at what seem to be unconventional settings (3 Blur Pixels, 50% Artifact Reduction and 100% Edge Softening and a touch of Microcontrast). The results have taken the image from unconvincing detail at any size to convincing detail at a moderate size (i think)...

100% zoom BEFORE:
topaz-infocus-1-before.jpg


100% zoom AFTER:
topaz-infocus-1-after.jpg
 
Top