Working at a software company means that I’m more than familiar with phrases like “advanced technology” and “breakthrough algorithms”… sometimes too familiar! Although these are actually decent descriptions, sometimes it’s nice to know exactly how the software works in order to be able to use it most effectively, instead of just knowing that it’s “advanced”.

Now, this is immediately relevant because we just released Topaz Detail, a plug-in that specializes in detail enhancement without halos. It does use pretty new technology – for example, a detail enhancement product that doesn’t create halos would have been impossible to develop just a year ago.

So immediately people emailed us wanting to know how it worked. One of the more common questions was why we implemented a pre-processing technique only in Topaz Detail rather than our other software. Well, the answer is so simple that it can be related in haiku form:

Internal splitting
Precise size-based enhancement
Topaz Detail rocks!

I know that probably answered all your questions. Here’s a more complex answer anyways, manifested in a delightful blue diagram:

Topaz Detail internal flow. Click for larger. (You may have to scroll right)
Topaz Detail flow. Click for larger. (You may have to scroll right)

Click on the diagram for a larger and more voluptuous version. I’ll walk you through the internal steps that Detail takes from start to finish.

1. First, you have your original image. You now access Topaz Detail and appreciate the fine aesthetics of the pre-processing screen.

2. Pre-processing starts. Topaz Detail separates your image into one based on chrominance (color information) and one based on luminance (grayscale information). These will undergo different processing methods and will be affected by different sliders.

3. The luminance information is further internally broken down into three detail layers based on size, and a base layer. Manipulating these four layers make up the bulk of the detail enhancement functionality of the software. After this step, pre-processing ends and the Topaz Detail user interface pops up. (This is also the portion of pre-processing that runs the slowest… probably because it has the most advanced breakthrough technological algorithms.)

4. The user adjusts the large, medium, and small detail sizes, and each individual detail layer is affected appropriately. If you like, you can see what each of the four individual layers looks like. To see the base layer by itself, turn the large, medium, and small sliders to 0. To see each individual detail size layer, drag that particular one all the way to the right and set the contrast to 0 (which will switch the base layer to neutral gray).

5. The luminance layer is re-combined with the adjusted small, medium, and large detail layers as well as the base layer.

6. Highlight and shadow protection, as set by the user, is applied to the luma layer.

7. The internal luma and chroma layers are re-combined to form your startlingly good final processed image.

Steps 5, 6, and 7 take no time at all. This means that, although there can be a sometimes annoying wait for the initial pre-processing stage, there will be no actual waiting after that. All adjustments are made instantaneously after the necessary initial pre-processing.

This is the complete internal workflow that Detail uses. Now that you know, it becomes quite obvious how it can sharpen and enhance detail without creating any halos – simply because it doesn’t really “sharpen”! All it does is increase the intensity of the appropriate detail layer, which gives a sharpening effect without any of the negative artifacts normally associated with sharpening or detail enhancement.

This post was originally posted on August 18, 2009

If you have a beloved car and a beloved camera, it’s practically your civic duty to use one on the other. So, to do this I waited till dusk, drove (roared) to a decent location, and whipped out my Canon 40D. Then came the Photoshop Car Retouching!

Then I opened up Photoshop and played with it a bit. I tried to keep it natural instead of going too overboard with the retouching.

First thing I want to mention – if you’re shooting car photography, most of the time it looks better if you shoot it on your elbows and knees. Shooting below the car, or some other creative non-everyday angle, will produce a much better photo than just shooting it at standing height. So wear something that you wouldn’t mind getting dirt on, and watch out for ants.

So, I got home and opened up the image in Photoshop in all its low-contrast goodness. The very first thing I did was notice the unsightly blotches that adorned my otherwise beautiful baby – the little nicks, scratches, and grime clumps that show up in many car images. Usually to get rid of these I would use the healing brush, which makes it hard to work around edges and other detail.

However, in this case I had another tool – Topaz Simplify. I actually just duplicated the layer, Simplified it, and masked in the Simplified parts where I would have regularly used the Healing Brush. The advantages to this are that edges and color are preserved while the minor blemishes get removed, unlike when using the Healing Brush.

My next step was to apply this oh so brilliant creative exposure, detail, and color software that I somehow had access to. I’m not going to tell you exactly what settings I used because I always like encouraging people to try using the sliders for themselves.

For this pass of Adjust, I ignored the sky and only focused on what made the car look good.

I tried accentuated the reflections while still keeping the car looking natural. Then I ran Topaz Adjust again, this time focusing on the sky. Afterwards, I masked out everything but the sky by utilizing clever masking techniques brushing on a layer mask.

The only thing left after this is to brighten the chrome horse and headlights, and darken the grill a little bit. Slap a small gradient vignette around the photo, and we get the final result.

And that’s it!

Simple steps:
1. Remove blemishes with Simplify and layer masking
2. Apply a layer of Topaz Adjust on car
3. Apply a layer of Adjust on sky
4. Dodge and burn select parts, add small vignette effect

Oh, and for a finishing touch, I added some Photoshop prowess, millionaire magic, and a dash of delightful dreams.

Just kidding. Hope you enjoyed the tutorial.

Please let me know if you have any specific questions about any of this, because I know I skimmed over how exactly to do a lot of what I described. Just drop me a comment for the full scoop on any questions or comments you may have.

This post was originally posted on July 27, 2009

One of the most common image enhancement workflow questions we get over here is…

At what point in my post-processing workflow should I apply noise reduction?

Why, excellent question! This is actually a very important factor in the quality of the resulting image, whether you use built-in Photoshop noise reduction tools or astoundingly good third-party noise reduction software. When to apply noise reduction is one of the most important and most overlooked aspects of effective image enhancement.

The simple answer is to apply noise reduction before applying ANY other adjustments to the image.

This includes JPEG compression, Camera RAW or even in-camera noise reduction, sharpening, exposure adjustments, dodging and burning, “Psychedelic” presets, whatever. See, in order to differentiate between noise and detail, noise reduction software have a very specific idea of what “noise” looks like and what “detail” looks like straight out of the camera. Applying any kind of adjustments to the image confuses the noise reduction software and causes it to give you a subpar result.

Demonstration time! Consider the following thousand words:

Canon 40D: f/2.8, 1/160, 1600 ISO
Canon 40D: f/2.8, 1/160, 1600 ISO

The image is very noisy and a little bit underexposed. I processed it a couple of different ways to illustrate the point of this post:

Brightened original image
Exposure adjusted, then DeNoised DeNoised then adjusted exposure
Feel free to click on the images for larger versions. The first version only has exposure adjustments – in this case Levels – applied to it. Oh no noise! 1600 ISO underexposed isn’t pretty.

In the second version, I took the RAW file and first corrected the exposure adjustment and then applied noise reduction. In the final version, I applied noise reduction as the very first step in my workflow before exposure correction. The results speak for themselves – and this was just a simple exposure adjustment!

Next time you really need to get the best results on your noise reduction, keep this in mind: no matter what noise reduction software you use, it’s always important to apply it first. I’d go as far as to turn off the built-in noise reduction and sharpening functions in Camera RAW before you import into Photoshop for best results.

There’s a lot of advice where noise reduction is put at the end of the workflow together with sharpening. Although this may be convenient, if you’re really looking for great results, put noise reduction smack at the beginning of your post-processing workflow before you touch a single pixel.

This post was originally posted on July 23, 2009

That’s right, now we can be developers, photographers, AND bloggers!

Several times a week I’ll be posting videos, Photoshop tutorials, interesting news, and small bits of blatant Topaz promotion for your enjoyment. Feel free to ask any questions in the comments or suggest future posts. After all, a blog is nothing more than a chance to put my thoughts out there and then read what you think, so I look forward to your feedback!

This post was originally posted on July 23, 2009