I played around in Photoshop for quite a bit before settling on the process I describe below. I’m pretty pleased with the outcome, so I thought it worthy of inclusion on this blog. Hopefully some of you will find it useful too!

Ok.. So first we need an image to work with. I’m going to use this one I took of a friend’s BMW Z4:

With the image open, the first step is to increase the midtones of the image.
This gives a more “fluid” look to the car and provide a better base for the alterations we will be making later on.
To do this, select the “Layer->New Adjustment Layer->Curves” menu option and then alter the curve so that the midtones are increased, as below:

Now we need to fade out some of the colour in the image, and we do that by adding a new black “Fill” layer.
To do this, select the “Layer->New Fill Layer->Solid Colour” menu.
On the “New Layer” dialog that will appear, name the layer “Black Fill”, select “Hue” from the “Mode” dropdown box and set the “Opacity” to 40%, as shown below:

Click “OK”, and then select true black (0,0,0) as the solid colour when prompted. Your image should now appear slightly less colourful than when we started, but don’t panic – that’s normal and intended!

Next we need to flatten the image (Select the “Layer->Flatten Image” menu), and then switch over to “Lab Colour” mode, which we do by selecting the “Image->Mode->Lab Color” menu option.

Once in Lab Colour mode, click on the “Channels” window (or select “Channels” from the “Windows” menu list), then click on the “Lightness” channel to select just it.

Click the “Filters->Sharpen->Unsharp Mask…” menu, and set the “Amount” and “Radius” values to 50, leaving the “Threshold” value set to 0 (zero), as shown below, then click “OK”.

Time to move back to RGB mode, so first select the “Lab” channel, then click the “Image->Mode->RGB Color” menu option. We should now be seeing a nice sharp image of the car with quite crisp and clean colours, with any reflections on the paint or windows showing much clearer:

Our image is beginning to take on that Top Gear feel, but we’ve one or two things left to do to make it that bit more special.

One of those things is to add a colour gradient, by clicking on the “Layer->New Fill Layer->Gradient…” menu option. Change the “Mode” dropdown value to “Overlay”, then click OK on the “New Layer” dialog.
In the “Gradient Fill” dialog window that appears next, we want to tick the “Reverse” checkbox, then click on the gradient box to set our own colours.(you need to click the coloured gradient, not the down arrow next to it)

This will open the “Gradient Editor”, with the default “foreground to transparent” gradient already loaded.
First of all, click the “New” button to ensure we don’t mess up any presets, then change both the start and end colours to a nice blood red; Click on the two sliders that appear below the gradient bar and choose a medium-dark red.
(If you’re unsure, they are the two sliders that appear blood-red in the image below)

Once you’re happy with the gradient, click OK to add it to the image.

Nearly there! Next we need to add the Top Gear Vignette to the image.
To do this, first ensure the “Background” layer is selected, then choose the elliptical marquee tool and, using the rulers for reference (CTRL+R), select from one “tick” in at the top left corner, to one “tick” in at the bottom right corner, as shown below:

Next, invert the selection (“Select->Inverse” menu option) and then feather it by 250 pixels (“Select->Modify->Feather…”) so that we are only selecting the outside edges of the photo.
Finally, select the “Image->Adjustments->Levels” menu and move the midtones slider to the right, until you get a nice level of vignette around the outside of the image. (The exact amount varies with each image, so play around until you find a value that suits your image.)

The final step is to increase the polarisation effect within the image, which will boost the detail and reflections on the car. To do this, we will be using the shadows and highlights adjustment tool. Select the “Image->Adjustments->Shadows/Highlights…” menu, then set both “Amount” and “Tonal Width” to 40%, then click OK

And there we have it! To complete the job, simply flatten the entire image, by selecting “Layers->Flatten Image” and you’re done!

Here’s how my sample image turned out:

Let me know how you get on and if you’ve got any improvements  

1
2

SELECT COUNT(1) 
FROM <<TABLENAME>>

And I got my row count back (incdidently, it was just over 325,000), but it was taking almost 10 seconds to return. Not Good!.

As I needed to do this in codebehind for a website, I didn’t want to loose ten seconds of loading time to what seemed to be a fairly easy query.

After a bit of hunting around in the new system tables (namely sys.Indexes and sys.Partitions), I managed to piece together the following little query, which runs nicely on the same table in under a second – MUCH better!

1
2
3
4
5
6
7

SELECT P.rows AS [RowCount]
FROM sys.indexes I
   INNER JOIN sys.partitions P 
      ON P.object_id = I.object_id 
      AND P.index_id = I.index_id
WHERE I.index_id < 2 
   AND P.object_id = OBJECT_ID(N'<<TABLENAME>>')

I have recently been working with a multiple web project solution in Visual Studio. For ages, I’ve been switching between each web project by stopping debugging, setting the other as the start up project, and then starting debugging again. As you can imagine, this makes testing interoperability between the two quite a drag.

One of my colleagues showed me that if you set a specific port for Visual Studio’s development server, you can then start the main project (in the normal fashion), but then also start the second project by right-clicking on it, and selecting “run” from the debug menu.

This was a good step forward for me, as previously I wasn’t sure it was possible to achieve debugging across two projects if there was no direct reference defined. However, being the person I was, I wasn’t quite satisfied with having to start one project, then manually starting the second once it had started. So I had a bit of a dive through the various properties and configuration screens within Visual Studio and discovered a method that allows you to start multiple projects, in a defined order, by simply hitting F5.

To do it, right click on the Solution, and choose properties from the context menu. Once the new window appears, ensure “Common Properties”->”Startup Project” is selected from the list on the left.
Then, on the right side of the screen, select the “Multiple startup projects” option. Finally, change the “Action” cell in the table for each project you want to start up when you begin debugging.
If you need to specify or change the order in which projects are selected, you can select the table row for that project, and using the up/down arrow on the right of the page, move it up and down in the startup order.

Quite a neat trick that could have saved me a good deal of ball ache in the earlier parts of this project!

I’m now the sole developer working on arguably one of the largest projects for our company, and things are going well. However, I keep falling into traps and pitfalls created by one of the projects last programmers.

One of his “favourite” tricks seems to be creating a business layer method that populates an object based on data retrieved from a database. Nothing odd there.. but it’s when you have the need to re-use one of his functions that you end up falling right into his carefully seeded trap!

For example, we have an object called “CustomerSite”, which as the name suggests, has various properties that pertain to a customer’s site – 15 individual properties to be exact.
Whilst writing an additional management page for the project today, I had the need to get a “CustomerSite” object from the database, based on the site code.
A quick look through our data access methods reveals a pre-written method “GetCustomerSiteBySiteCode”. Excellent; No need to write a new stored proc and no need to write the retrieval C# code!

However, in hindsight I knew this particular programmer had written both the retrieval method and the stored proc, so I should have checked ahead of time, but alas I did not…

So I fire up my new shiny page and all goes well, right up to when my codebehind calls the “GetCustomerSiteBySiteCode” method. At which point, my page throws a wobbly with invalid use of null errors throwing hand over fist.

Initially, I start checking the database to make sure the data is actually fine for the customer site I selected. With that being fine, I check my code to make sure I’m not doing anything daft. Happy that I’ve not, it’s time to place a breakpoint and step on through…

It’s at the point I step into “GetCustomerSiteBySiteCode” that I realise I’ve been stung; Despite the “CustomerSite” object having 15 properties, only TWO of them are being set by this method – the rest are simply set to null.

After altering the method to actually populate all 15 properties, I checked the stored procedure, to find two select statements, the first with all fifteen fields commented out with the note “Probably want to include these fields later, but haven’t the need to use them yet.”

To quote Peter Griffin “This really grinds my gears”. It was the programmer who created the “CustomerSite” object to begin with, and it’s always had all 15 properties from the get go. It’s simply a case of laziness and lack of forethought; there’s every reason to believe that this method would be called elsewhere, especially as when he was working on the project, it was a multi-person project.

Grr.

We’ll start with the same photo as before:

As before, the image appears quite “flat” and lifeless. We can try and fix this with a very quick and (usually) efficient contrast tweak using a “soft light” layer.

To this, open your image up in Photoshop, then create a duplicate of the original background layer (right-click on the background layer and select “duplicate”..).
Now, with the duplicate layer selected , choose “Soft Light” from the dropdown list of layer types:

At this point, your image may already look better, but it some cases (including this image), the soft light can be a little overpowering on full opacity.
Obviously, every image is different, so there’s not a “one size fits all” opacity setting, so you’ll need to experiment with each image; Try dropping the opacity down to around 70%, or until you gain a good contrast.

I settled for 70% for this image, and the results I gained were a good deal better than the original:

So, first of all we need an image to work with. Here’s one of my little cousin:

As you can see, the image appears very “flat”.  I know it’s not exactly a fantastic photo to begin with, especially as it’s from a 1mp camera phone, but it’s good enough to demonstrate this trick!

So, once you have your image in Photoshop, right click the background layer and select “Duplicate” to give us an exact copy as a second layer.
Then, choose “Filters->Other->High Pass” from the menus and when prompted, select a radius of 10 and click ok.
You should end up with something grey looking, similar to this:

Next we need to make our new layer an “overlay”. To do this, select the dropdown list of layer types and choose “Overlay”.
This next image shows the location of said dropdown:

Now we need to apply a mask to this layer, as we don’t want to apply the high pass results to the full image. So, whilst holding ALT, click the “Add Layer Mask” button, located at the bottom of the Layers pane.

Masks work by using the colours white and black to determine if whatever is being masked should pass through, or not, respectively. By holding ALT when creating our mask, it will be created entirely black (meaning none of the layer will pass through). This is what we want, as we only need to apply the high pass results to specific areas of the image, as you’ll see next!

Ok, now we have our mask, ensure it is selected. It will have a black and blue border around it if it’s selected – see the next two images for an example:

Now, select the paint brush tool, and the colour pure white, then start to paint over the eyes, hair, mouth and teeth (or any other areas you wish to “draw attention” to within your image).

Once you’re done, you should end up with something like this:

which, if you compare against the original, seems to pop out and his face draws your attention more:

Hope this is of some use

There were plans for a company called Global Gaming Factory (GFF) to take over control of The Pirate Bay, in a 60m kronor (£3.5m) acquisition, but it looks like those plans may have fallen through as GFF haven’t (so far) proved they actually HAVE 60m kronor.

Earlier this week, one anonymous user posted a web-rip of The Pirate Bay, including all 900,000 torrents,  back on the site as a single 21GB torrent download. He (or she) stated “If the TPB deal disappoints us, we can just put it up again”.

However, various other users have already started “clones” of The Pirate Bay, using the massive torrent database download.
Sites such as http://www.btarena.net/ have already experienced overwhelming demand as the pirates set sail from their old bay in search of new homes…

Another important thing a histogram does is describe contrast within your images. According to Wikipedia:

Contrast is the difference in visual properties that makes an object (or its representation in an image) distinguishable from other objects and the background.

Or, in more simplistic terms, Contrast is a measure of the difference in brightness between the light and dark areas of your photographs.
Images with histograms that have broad tonal ranges have a significant, or “high” contrast, whereas images with histograms that have narrow tonal ranges may appear dull or flat and are known as “low” contrast.

Differences in contrast can be caused by any combination of subject matter and lighting conditions. By rule of thumb, images taken in fog will be low contrast and those taken under strong daylight will have a high contrast.

Contrast can have a significant visual impact on an image by emphasizing texture, as shown in the following image:

The left side of the image is a low contrast image, whilst the right side is high contrast. The high contrast water has more pronounced shadows and highlights, which enhances the wavelets and creates an image which “pops" out at the viewer.

In the last article, we discussed the basics of what a histogram is, what it represents and how it is generated. What we didn’t cover is why they’re important to us digital photographers. That’s where this post comes in!

Most digital cameras will have next to no trouble reproducing images that have a midtone-centric histogram (a histogram where the majority of the tonal range is within the midtone section). This is because they have an “automatic” imaging mode which uses a clever set of algorithms that attempt to estimate how bright an image should be, and these algorithms frequently result in an image who’s average brightness is placed in the midtones.
(it’s all down to the fact digital cameras have to measure reflected, rather than incident light, but we’re not going to discuss that here.)

This is usually acceptable, however images where the majority of the tones occur in the shadow range (low key) or highlight range (high key) are often unacceptable, as the algorithms try to adjust the brightness of the scene too much (moving the tonal range back into the midtones, rather than leaving them in the shadows or highlights).

To give you an example of this in action, below is an image taken by manually setting the white balance and exposure:

As you can see, the histogram shows this as a high key image, with the tonal range right up in the highlights – exactly as wanted.
However, the next image is the same scene but taken using the automatic mode within the camera:

As you can see, the camera’s algorithms have got involved and tried to shift the overall brightness down so that the tonal range is nearer the midtone section, which has resulted in an effectively underexposed image.

Whilst this isn’t great, it is a relatively easy process to fix this in Photoshop (and we will cover that later on). However, if we were taking a low key image, we could hit a big problem.

Take the following image as an example. This scene was very shadowy and dark, but the camera’s automatic mode has increased the overall brightness:

The problem however, is that it’s lost some of the highlights from the original image; In the histogram, you can see that a large number of pixels are pure white (or very close to).  This is known as having “blown” or “clipped” the highlights.
Unlike an underexposed image, it is impossible to recover any detail from the “blown” highlights.

This is where being able to understand histograms comes into play!

Most digital SLR cameras will allow you to view the histogram of an image on the LCD back panel, giving you instant feedback that can help improve your image, without having to wait till you get the images back to Photoshop.
That quick glance is enough to tell you if the image you’ve just taken is likely to be a decently lit image.

For example, if you took the image of the gate shown above, you would quickly be able to see that you’ve got blown highlights, and as such adjust the exposure on the camera to compensate, giving you something much nicer:

In the final post about histograms, I’ll be covering contrast…
http://www.blog.ianmellor.co.uk/2009/07/30/contrasting-views/

The only crux is he’s never really used it and has asked for a bit of help.
To that end, I’ve decided to include a few tutorials in my blog to give him a bit of a head start into learning the dark art that is Photoshop!

The first tutorial I was planning to write up was how to use the level’s tool to make those little tweaks to the image brightness values that can really improve an image. However, as my colleague didn’t quite understand what we were doing and what the histogram represented, I thought it best to actually start with explaining what the histogram actually is…

According to Wikipedia, an image histogram is:

An image histogram is type of histogram which acts as a graphical representation of the tonal distribution in a digital image. It plots the number of pixels for each tonal value.
By looking at the histogram for a specific image a viewer will be able to judge the entire tonal distribution at a glance.

However, considering I feel that actually understanding image histograms is probably the most important concept to become familiar with whilst working with digital photographs, I think we best expand on that definition a little more!
(I truly believe that understanding how histograms work will not only improve your ability to manipulate images within Photoshop to greater affect, but will also improve your photography skills in general. It did for me!)

As Wikipedia states, a histogram displays tonal distribution within a digit image. With this data, you can easily check if an image has been properly exposed, if it was correctly lighted, and (more importantly from a Photoshop perspective) what adjustments will work best to produce a better final image.

A common misconception is that there is an “ideal” histogram which all images should try and mimic. This isn’t true; histograms should merely be representative of the tonal range within the scene being photographed and what the photographer is trying to convey.

Whilst there are a few different types of histograms (RGB, CMYK, etc), they all work in a relatively similar way and all produce the same basic graph-like display as shown below:

For the purpose of this tutorial, we will only be looking at an RGB histogram.

So, How do they work?

Every pixel within a digital image has been produced by some combination of the three primary colours red, green and blue (RGB).  For “8-Bit” images, each of these colours is given a value between 0 and 255 which represents how bright that colour should be. For example, if red and blue were set to 0 and green to 255, then the pixel would display a bright green colour. If all three were set to 128, then the pixel would display a medium grey colour.

An RGB histogram is built by scanning through the brightness value of these three colours for each pixel within the image and records how many are at each level from 0 to 255. These results are then used to create the graph.

Below, I’ve added a few comments to the graph shown within Photoshop to help better explain how to read the graph:

Looking at this example, we can see that the vast majority of the pixels within the image have low brightness values.  This area where the majority pixels can be found is called the “Tonal Range”.

Tonal range can differ greatly between images, so building an intuition for how these numbers map to actual pixels is  often critical – both before and after the image is taken. Below is an image that hopefully gives you some idea of how the histogram ties up to the actual image it is created from:

This image of a creek has very little highlights and deep shadows, but lots of mid-tones as shown by high pixel count curve, in the off-centre of the histogram.
Another example is this shot of Anfield, which has hardly any mid-tones or highlights, but is dominated by heavy shadows:

However, lighting is not normally as extreme as the above image. A properly exposed image, with even lighting will usually produce a histogram with peaks in the centre, and gradually dies off into the shadows and highlights, similar to the following:

Most digital camera will have no trouble automatically reproducing which has a histogram similar to the one shown in the above image.

Ok – so that’s it for now. Try viewing the histograms for images you’ve already taken and try to build up the “knack” of guestimating what your image’s histograms may look like, as this will be helpful when we move onto the next section!

(Tip: To view your image histogram in Photoshop, first open an image, then whilst holding the CTRL key, press L. Or, select “levels” from the “Image->Adjustments” menu.)

In the next post, I’ll explain what high and low key images are and why they’re important in digital photography…
http://www.blog.ianmellor.co.uk/2009/07/28/the-highs-and-lows/