One of the commonly used sharpening methods in digital image-processing is unsharp masking. This method detects the edges within an image and increases them – rendering an effect of perceived sharpness. Deconvolution technology, on the other hand, tackles sharpness and image blur in a more complex but beneficial way.
In image processing, deconvolution is a restoration process that essentially allows a blurred image to be re-focused. It is able to achieve this by detecting the state of the image before the distortion occurred and approximately reversing the process that caused the optical blur to occur. While unsharp masking increases the perceived sharpness of an image, deconvolution increases the actual sharpness based on information which describes some of the likely origins of the distortions when capturing the image. With deconvolution, lost image detail may be approximately recovered.
Topaz InFocus uses this same deconvolution technology to mathematically reverse image blur while increasing the actual sharpness. InFocus can enhance the clarity of an already well-focused as well as deblur an out-of-focus or motion blurred image. The ability to re-focus a blurred image (after the fact) is a challenging but valuable capability. This technology can be applied to both enhancing the clarity in an already well-focused image, as well as de-blurring out-of-focus or motion blurred images.
A common side effect of the deconvolution process is deconvolution artifacts. These artifacts can appear as a collection of artifacts around lines of details in an image. Topaz InFocus includes a parameter specially designed to eliminate these artifacts.
Topaz DeNoise 5 is powered by a leap forward in noise reduction technology – IntelliNoise™ – which was developed in-house by Topaz Labs after years of research into image processing. The goal of the project was to discover a noise reduction solution that was actually capable of recovering detail as well as removing noise. The effort was spearheaded by a Ph.D in Computer Vision with twenty years of experience in the field, and the culmination of that effort means that the IntelliNoise™ technology exclusively powers Topaz DeNoise 5.
The big benefit of IntelliNoise technology is that it allows higher levels of noise reduction than existing methods while still keeping the photo looking natural. The idea is to make your ISO1600 photos look exactly like you shot it in ISO100 on a tripod. Compare the DeNoised ISO1600 and ISO100 below, and roll over for the original ISO1600 photo.
The DeNoised ISO1600 photo has the exact same amount of noise as the ISO100 photo, yet looks natural and unprocessed. Note the preserved detail and sharp lines with none of the telltale signs of blurriness or blotchiness. It’s as if you shot the photo at 100 ISO to begin with, just without any of its disadvantages like a longer shutter speed or smaller depth of field! This is not currently possible with any other technology.
One of the new DeNoise 5 features is a debanding tab on the user interface, used to remove banding noise from images. This new debanding technology tackles dual-directional banding noise, eliminating the long lines that extend across images and restoring image color, detail and integrity.
What is Banding Noise? Banding noise appears in images as colorful parallel stripes that extend across an image. Banding noise can be horizontal or vertical – depending on the orientation of the camera at the time of the shot. Occasionally you may come across an image that exhibits both horizontal and vertical banding noise.
There are two forms of banding noise: short banding noise and long banding noise. Short banding noise appears in isolated areas of an image, while long banding noise extends across the entire image.
What Causes Banding Noise? Banding noise is a camera-induced artifact caused by the camera’s sensor. It is typically unpredictable and unavoidable. Banding noise is frequently seen with the use of high ISO speeds, in shadows and in images that have been significantly brightened.
Before DeNoise 5, banding noise removal was challenging, as there were no tools to remove it and very little resources for preventing it. With DeNoise’s debanding tools you can now remove horizontal and vertical long banding from your images.
Click below to see entire image
DeNoise 5 is a free upgrade to customers who own a previous version of DeNoise or a bundle that includes DeNoise. To access the upgrade, simply download the new version from our downloads page and install it with your existing license key.
Have any questions or comments about noise reduction, DeNoise 5 or the new debanding tool? Ask away below!
Much has been said about Topaz ReMask, but now there’s one more thing to talk about: ReMask is about to become faster, easier, and prettier right in front of your eyes. We’ve been working Topaz ReMask v2, which includes a user interface for easier and more intuitive use – and even enhances the already high masking speed and mask quality of the original.
Just to give you an little sneak peek of what it will eventually look like, here’s a shot of the current alpha user interface:
(note that this may undergo significant changes by the actual release of ReMask 2)
The overall workflow will be like this:
Paint the KEEP (green), CUT (red), and PROCESS (blue) areas with a combination of fill tools and brush tools.
Hit the process button. The mask immediately updates right there in the user interface.
Refine the mask if you wish by specifying more parts of the image to keep or remove. Refinement much easier because you can see exactly what you’re doing right there on the screen.
Extremely simple, highly intuitive, and still blazing fast – just the way masking should be. Here are a couple of other cool features in the new version:
Vastly improved and specialized hair masking algorithms.
Saves result to a layer mask in Photoshop if you have one enabled on the layer, or just cuts image out of its background if you don’t.
Four views that show all aspects of your mask at all times for the utmost control over your final result.
Undo/Redo functionality and full keyboard shortcut support.
Highly improved memory management – uses own memory instead of sharing Photoshop’s memory. This makes it a little faster and a lot stabler than original ReMask.
Personally, when I tried using the new version of ReMask for the first time, I was actually extremely surprised by how easy it was to use. Astronomical props go to our developers for releasing a completely redesigned and reworked product in such little time. This is not a little minor update, this is practically an entirely new product – looks like it, feels like it, and works like it. I usually don’t like to gush catchphrases but I was sincerely blown away by the difference between the new and old ReMask – it’s simply amazing. PSP and Elements will also be supported by ReMask 2 in addition to Photoshop.
Despite all these improvements, we really appreciate our early adopters and would therefore like to offer ReMask v2 when it’s released as a completely free upgrade for existing customers of ReMask v1. Your support and feedback is what makes the constant improvement of our products possible.
Thanks again! ReMask 2 is scheduled for a mid-November release, so you don’t have too long to wait.
This post was originally posted on November 5, 2009
Twitter is great as a communications tool. I was answering questions and discussing our software with Twitterers (…Tweeple? Twitterholics?) by the first month. The communication barrier is much lower than email or phone – a simple @topazlabs and we’ll get your message or feedback instantly and be able to respond equally as quickly. And on your end, you’ll be able to receive our updates instantly and unobtrusively. For example, if we have a discount that we announce through Twitter, you’ll know about it without having to open up an email or check our website.
So on that note, we’re giving away a Topaz Photoshop Bundle to three different Twitterers at the end of September! To enter the contest:
1. Follow @topazlabs on Twitter.
2. Tweet this message:
RT @topazlabs: 3 free Topaz #Photoshop Plug-In Bundles! To enter raffle: follow @topazlabs and retweet! Details here: http://is.gd/2Lxbd
We’ll be picking the winners by taking three random @topazlabs followers (generated from a randomizer) and then making sure they have the above tweet in their timeline. We will announce the winners at the beginning of October. Good luck, and comment here with any questions!
This post was originally posted on September 1, 2009