Requirement 0 in the digital area

You are here: Nature Science Photography – Contrast – Contrast in photography

The easiest way to meet the premise set up at the very beginning is to save as .jpeg right in the camera. This format is the digital equivalent of the standard AgX image carrier and, in addition to gamma correction, already contains a tone curve that varies slightly depending on the manufacturer, which ensures that the print fulfills all the factors listed above and looks „good“. For all those who want to customize their pictures, we will go through it point by point below.

Compensation of Stray light

We can perceive stray light as a consistent amount of light overlaying the image in the focal plane. In relation to the linear data of a digital imaging system, stray light results in a raised black point and a corresponding positive shift of all other pixel values. Graphically, the stray light moves the curve upwards as a whole without changing its shape. To compensate for this effect, we need to subtract a constant value from all pixels. This can be done by adjusting the black point, the histogram, the tone curve, the brightness or contrast. The subtraction yields the best results when the image’s data values remain in linear order. That is, before gamma correction or conversion to a nonlinear color space. – So going after a finished .tif or .jpeg with any image editing program is not so good. Manipulating .raw data in a .raw converter, on the other hand, is, because such applications show a gamma-corrected preview – more on the meaning of gamma correction below – but apply their calculations to the linear dataset. Therefore, changes in values of the same size affect shadows and highlights of the same size. This is not the case when manipulating gamma-corrected data. However, calculating the preview takes the .raw converter a moment (it is not real time) due to the need to first convert the editing into gamma correction, which requires significant computational effort.

In practice, the correction of a 2% stray light is sufficient. So with a maximum pixel value of 255, the black point should be 255*0.2 = 5. – Surprisingly, this is the default value of the black point (shadow slider) in Camera Raw and now we know why that is! Even if this value has proven itself, it is useful to watch the preview window because the optimal setting always depends on the image content. As you increase the black point, the image usually becomes more vivid and gains depth and three-dimensionality. But if you go too far and set too many values to zero (clipping), the shadows quickly become unsightly.

Compensation of ambient brightness

Processing the images into prints eliminates the need for ambient brightness compensation. Even in the rare event that the data needs to be exposed on reversal material to slides, the film imagesetter automatically performs the necessary correction.

Increase of image quality

Assuming a calibrated working environment, we can rely entirely on our eyes to increase image quality in the digital realm and judge each correction immediately – no more trial and error with different gradations, as in the wet lab. We already know the name of the corresponding tool: the tone curve, also known as the gradation curve, is the preferred tool. It is arguably the most powerful and flexible image editing tool, though also the most intimidating for many photographers. Just as characteristic curves help each silver film and paper achieve distinctive characteristics, so does the digital tone curve, because it affects the two main effects of light: tonal values and contrast.

Similar to tone correction, the tone curve lets you pick out tonal values and selectively stretch or compress them. However, tone correction only gives you control over black, white, and the middle gray. With the tone curve, you rule over up to 16 points along the curve, which are inserted by double-clicking. The result of a manipulation can be visualized by following an input tone value of the horizontal x-axis vertically up to the curve and then left to the corresponding output tone value on the y-axis. So a diagonal line through the center of the graph leaves all tone values unchanged. And as long as we don’t introduce a curve with a negative slope, any manipulation maintains the global tonal hierarchy. Tonal values that were brighter than others before the manipulation will be brighter afterwards, though not necessarily by the same amount.

If you trace any two tone values in figure 44, you will notice that their distance increases (they are stretched) when you steepen the curve and decreases (they are compressed) when you flatten it the other way around. In terms of contrast, stretching means less contrast, and compressing means more contrast. Figure 45 illustrates this and shows the two most used curve types: the S-curve and the
inverted S-curve . The S-curve increases contrast in the midtones at the expense of highlights and shadows, while the inverted S-curve does the opposite. Notice how these changes affect the image: With the S-curve, detail in the midtones becomes more prominent, but detail is lost in the highlights and shadows. For the inverted S-curve, as I said, the opposite is true.

This brings up an important keyword: Increasing contrast in one area always comes at the expense of another. This is the most important concept in terms of the tone curve. You cannot increase the contrast of one tonal range without simultaneously decreasing it elsewhere. All photos have a „contrast budget“ that we cannot exceed, and the tone curve simply redistributes the contrast within that budget. The question arises: why should we do this if we have to balance every manipulation somewhere? There are two answers to this.

Our goal in this operation is to make the image look „good“. Since the contrast of the midtones is perceptually the most important, we increase it until the image looks good to our satisfaction. Just as with most silver films and papers, an S-curve serves us for this purpose.

The second answer relates to the requirements of our output materials, as most photos have a wider dynamic range than what can be reproduced on paper. With the tone curve, we have a means of making the best use of this limited dynamic range and deciding where to apply the inevitable compression. Since the contrast of the midtones must be raised for the result to be acceptable, the shadows and highlights usually bear the brunt of this tone value compression.

In many cases, five anchor points in the shadows, midtones, and highlights plus a white point and black point are sufficient to make the necessary corrections. Note that even slight shifts in these points can result in significant changes in the image. Abrupt changes in the slope of the curves will almost certainly lead to tonal breakup (posterization) because they will overextend the tonal values in areas of shallow transitions. For this reason, you should also observe the histogram parallel to the tone value curve. So moderate changes that result in smooth curve shapes usually work best. To have the maximum amount of control, you can enlarge the tool window.

To correct the tone curve changes at any time, use an adjustment layer (in Photoshop: Layers Tab – New Fill Layer or Create Adjustment Layer – Gradation Curves) to bring the changes in.

The issue of the relationship between contrast and color saturation persists, as an increase or decrease in contrast correspondingly affects color saturation. Figures 46 to 49 illustrate this connection. Compared to the unprocessed original image, we see a strong change in color saturation in the RGB mode and a less pronounced change in the brightness channel of the Lab mode. This difference stems from the fact that contrast enhancement in RGB mode directly converts the pixel values of the three primary colors red, green and blue to higher, i.e., darker, values (and a darker red is a more saturated red), but the relationship in the Lab model is indirect. It comes like this. Lab divides the color information into two parts: the brightness part, L*, and the two parts of chromaticity, a* and b*. However, the correlation between the color saturation and the result of the chroma / brightness division remains unchanged. Increasing the contrast in the brightness channel sets all pixels to higher (darker) values, but if a* and b* remain the same, the result of the division changes. However, the effect is not as significant as it is in the RGB model.

Simon Tindemans has thankfully written a series of Photoshop actions (LuminanceCurve & LightnessCurve) and a plugin (Tonability), respectively, which serve to avoid the side effects of the normal tone value curves on color saturation. They convert pixel values from one lightness to another without changing the R:G:B ratio. On his website (www.21stcenturyshoebox.com), he gives detailed instructions on how to integrate the tools into a workflow with Camera Raw. Although the majority of users prefer the increase in color saturation associated with contrast enhancement. This is probably because our viewing habits are so shaped by the behavior of AgX image carriers, which, until the introduction of the most modern color couplers, were determined by the connection „contrast plus = color saturation plus“.

Next Gamma correction the first – distorting linearity

Main Contrast

Previous Practical consideration of the variables dynamic range, sensitivity, full well capacity and pixel area

If you found this post useful and want to support the continuation of my writing without intrusive advertising, please consider supporting. Your assistance goes towards helping make the content on this website even better. If you’d like to make a one-time ‘tip’ and buy me a coffee, I have a Ko-Fi page. Your support means a lot. Thank you!