Luckypenguin
Hall of Famer
- Location
- Brisbane, Australia
- Name
- Nic
A little while back I was asked by a coupleof forum members to give a bit of a description of my processing workflow and how I achieved a certain "look". I've finally gotten around to doing a workflow description which follows below. Everything described here is based on a Lightroom workflow (v4.4) and running on a Windows 7 operating system. A lot of what is described here has analogs in other editing softwares but some of the terms used by each may vary from those mentioned here and some of the particulars of the software may be different. For information's sake the image used as an example was taken with an Olympus OM-D E-M5 and Olympus M Zuiko 9-18mm f4-5.6 ultra wide-angle lens set at 9mm.
My first stop, and something I only started doing not all that long ago, is to drag and drop the image onto the map. I don't particularly have a use for this but it's easy to do so why not?
Next step is to go into the develop module where all the work begins
If I was to export the raw file without any edits it will look like this. The point here is that the undeveloped raw file is obviously just the starting point, not the finished product. You'll notice that this particular image was taken partly in shade and partly in direct sun so a choice needs to be made on how to expose it correctly to deal with the dynamic range of the scene. For me there is no real choice when exposing an image; I'll always try to push the exposure as far into the highlights as I dare without burning any highlights that I wish to retain. This is what I call "shoot-to-process", where I'm not worried about whether the image is exposed correctly but whether it is exposed to give me the best file for processing. Specifically, I want to minimise or even eliminate any pushing of the file which will induce noise, particularly in shadows. More often than not my OOC jpegs don't look very good, but this is by choice not by bad management. The Olympus used here is the perfect camera for optimising the exposure for three reasons. 1, it has a solid dynamic range to work with including a good amount of room in the highlights to pull back data. 2, it has an extended dynamic range jpeg mode which gives me a good representation of the dynamic range that I might eventually be working with in the raw file. This is important because it much more accurately defines the highlight and shadow limits than a standard, contrasty jpeg. 3, as well as the usual histogram it has what are referred to as "shadow and highlight blinkies" where it will overlay blown highlights with red and black shadows with blue. Finally 4, the image displayed on the screen IS the same exposure as the image that gets recorded to card. It is not some equalised image designed to give a consistent exposure to the live view image.
Looking at the histogram, you can see what happens when you expose to the right. I will be relying on the ability to recover the highlights from the raw data
The first real processing that I will do is load my preset for this particular camera. I won't go into the particulars here since as I go through I'll be fine tuning most of the preset values.
Looking at the image after applying the preset you can see that it is looking better already. Here is the histogram before and after loading the preset. You'll notice that the highlights have pulled back to the left slightly and the centre is now more evenly distributed.
Usually from this point I'd start working my way through the Basic Module but the orientation of the image immediately looks wrong to me so I'm going to deal with this first. In framing the image I am pointing the camera up slightly which is going to induce perspective distortion, and unless I was a little drunk at the time (possible) it appears as though I've made the mistake of lining up the right edge of the vertical. In some circumstances this may give an interesting effect, but otherwise it is wrong, wrong, wrong. It is the centre of the image should be vertical. Going to the Lens Correction Module (and making sure to tick Constrain Crop so the edges are automatically trimmed), I can hover my mouse over the Rotate slider and get a grid overlay.
In this case the centre of the image is five squares in from either side, so this is the line at which I want the image to be vertical.
Next up I want to deal with the perspective distortion where the image appears to be leaning in at the top. This is adjusted using the Vertical slider. Generally it is considered more natural to leave just the slightest hint of lean to the image, but you can also choose to go straight up and down if you wish.
The full image
I'm not happy with the edges of the image so I'm going to give it a crop to remove the distracting edge details on the left
The full image again. I'm still undecided how far I want to crop the right edge so I'll come back to this later.
At this point I'm now ready to start looking at the exposure. I do this by looking at the image itself of course but also by keeping an eye on the histogram, since this provides a visual representation of the distribution of the exposure. I'll go into the methods of adjusting the exposure in some detail since this is "where it's at" in terms of raw editing. In Lightroom you get five sliders to adjust this: Exposure, Highlights, Shadows, Whites, and Blacks. These are ordered logically from a workflow perspective but not from where each slider influences the histogram. From left-to-right (darkest to lightest) it goes Blacks, Shadows, Exposure, Highlights, Whites. To see the area of influence of each slider you can hold your mouse over each slider and the corresponding section of the histogram will be highlighted.
The important thing to note here is that each slider won't only adjust the data within the highlighted section of the histogram, but it is weighted to that section. Each slider influences the entire histogram distribution but with diminishing effect the further the data is away from the highlighted section.
One of the key benefits of raw processing is the ability to access the full dynamic range of the sensor, but simply pulling down the highlights and pushing up the shadows doesn't produce a pleasant looking image, because high dynamic range and contrast don't go together. You might say that Lightroom and other editing programs offer contrast adjustment, but I don't use it because it is a very crude method of exposure control. You still need the appearance of contrast of course, and this is where I want to be adjusting those five sliders to ensure that I have a histogram that stretches all the way from left to right. An image should always have a true black, and almost always have a true white somewhere.
Getting back to my example image, not having a true white isn't an issue since I already have too many. Lightroom offers a neat way of showing black and white pixels. To show which parts of the image are true white (i.e. blown) you simply click and hold either the Exposure, Highlights, or Whites sliders and hold down the Alt key. The image will turn black where it is not over exposed and white where it is. It will also display colours which represent where individual colour channels have been blown. White means that every channel is blown.
By cropping in on one of the offending sections we can see this in more detail.
To recover these pixels we need to go to the highlights and whites sliders and pull them to the left (negative values). This is a trial and error process. You can follow the histogram, you can use the blown pixel alert (the Alt key trick), but ultimately it is looking at the image itself that should be your guide. In this case I've used the values as shown.
Checking my blown pixels I now get this. I haven't eliminated all of them, but as I alluded to earlier I believe that an image should ideally just kiss the right hand edge of the histogram. Not all the time, but most of the time.
The entire image is shown below...
...and the same crop that we looked at earlier, but now less retina searing and with some detail returned.
The histogram as it currently stands. You'll see that the spike on the right which represents the brightest, whitest parts of the image has now been pulled back from the very edge of the histogram.
To be continued...
My first stop, and something I only started doing not all that long ago, is to drag and drop the image onto the map. I don't particularly have a use for this but it's easy to do so why not?
Join to see EXIF info for this image (if available)
Next step is to go into the develop module where all the work begins
Join to see EXIF info for this image (if available)
If I was to export the raw file without any edits it will look like this. The point here is that the undeveloped raw file is obviously just the starting point, not the finished product. You'll notice that this particular image was taken partly in shade and partly in direct sun so a choice needs to be made on how to expose it correctly to deal with the dynamic range of the scene. For me there is no real choice when exposing an image; I'll always try to push the exposure as far into the highlights as I dare without burning any highlights that I wish to retain. This is what I call "shoot-to-process", where I'm not worried about whether the image is exposed correctly but whether it is exposed to give me the best file for processing. Specifically, I want to minimise or even eliminate any pushing of the file which will induce noise, particularly in shadows. More often than not my OOC jpegs don't look very good, but this is by choice not by bad management. The Olympus used here is the perfect camera for optimising the exposure for three reasons. 1, it has a solid dynamic range to work with including a good amount of room in the highlights to pull back data. 2, it has an extended dynamic range jpeg mode which gives me a good representation of the dynamic range that I might eventually be working with in the raw file. This is important because it much more accurately defines the highlight and shadow limits than a standard, contrasty jpeg. 3, as well as the usual histogram it has what are referred to as "shadow and highlight blinkies" where it will overlay blown highlights with red and black shadows with blue. Finally 4, the image displayed on the screen IS the same exposure as the image that gets recorded to card. It is not some equalised image designed to give a consistent exposure to the live view image.
Join to see EXIF info for this image (if available)
Looking at the histogram, you can see what happens when you expose to the right. I will be relying on the ability to recover the highlights from the raw data
Join to see EXIF info for this image (if available)
The first real processing that I will do is load my preset for this particular camera. I won't go into the particulars here since as I go through I'll be fine tuning most of the preset values.
Join to see EXIF info for this image (if available)
Looking at the image after applying the preset you can see that it is looking better already. Here is the histogram before and after loading the preset. You'll notice that the highlights have pulled back to the left slightly and the centre is now more evenly distributed.
Join to see EXIF info for this image (if available)
Join to see EXIF info for this image (if available)
Usually from this point I'd start working my way through the Basic Module but the orientation of the image immediately looks wrong to me so I'm going to deal with this first. In framing the image I am pointing the camera up slightly which is going to induce perspective distortion, and unless I was a little drunk at the time (possible) it appears as though I've made the mistake of lining up the right edge of the vertical. In some circumstances this may give an interesting effect, but otherwise it is wrong, wrong, wrong. It is the centre of the image should be vertical. Going to the Lens Correction Module (and making sure to tick Constrain Crop so the edges are automatically trimmed), I can hover my mouse over the Rotate slider and get a grid overlay.
Join to see EXIF info for this image (if available)
In this case the centre of the image is five squares in from either side, so this is the line at which I want the image to be vertical.
Join to see EXIF info for this image (if available)
Next up I want to deal with the perspective distortion where the image appears to be leaning in at the top. This is adjusted using the Vertical slider. Generally it is considered more natural to leave just the slightest hint of lean to the image, but you can also choose to go straight up and down if you wish.
Join to see EXIF info for this image (if available)
The full image
Join to see EXIF info for this image (if available)
I'm not happy with the edges of the image so I'm going to give it a crop to remove the distracting edge details on the left
Join to see EXIF info for this image (if available)
The full image again. I'm still undecided how far I want to crop the right edge so I'll come back to this later.
Join to see EXIF info for this image (if available)
At this point I'm now ready to start looking at the exposure. I do this by looking at the image itself of course but also by keeping an eye on the histogram, since this provides a visual representation of the distribution of the exposure. I'll go into the methods of adjusting the exposure in some detail since this is "where it's at" in terms of raw editing. In Lightroom you get five sliders to adjust this: Exposure, Highlights, Shadows, Whites, and Blacks. These are ordered logically from a workflow perspective but not from where each slider influences the histogram. From left-to-right (darkest to lightest) it goes Blacks, Shadows, Exposure, Highlights, Whites. To see the area of influence of each slider you can hold your mouse over each slider and the corresponding section of the histogram will be highlighted.
Join to see EXIF info for this image (if available)
Join to see EXIF info for this image (if available)
Join to see EXIF info for this image (if available)
Join to see EXIF info for this image (if available)
Join to see EXIF info for this image (if available)
The important thing to note here is that each slider won't only adjust the data within the highlighted section of the histogram, but it is weighted to that section. Each slider influences the entire histogram distribution but with diminishing effect the further the data is away from the highlighted section.
One of the key benefits of raw processing is the ability to access the full dynamic range of the sensor, but simply pulling down the highlights and pushing up the shadows doesn't produce a pleasant looking image, because high dynamic range and contrast don't go together. You might say that Lightroom and other editing programs offer contrast adjustment, but I don't use it because it is a very crude method of exposure control. You still need the appearance of contrast of course, and this is where I want to be adjusting those five sliders to ensure that I have a histogram that stretches all the way from left to right. An image should always have a true black, and almost always have a true white somewhere.
Getting back to my example image, not having a true white isn't an issue since I already have too many. Lightroom offers a neat way of showing black and white pixels. To show which parts of the image are true white (i.e. blown) you simply click and hold either the Exposure, Highlights, or Whites sliders and hold down the Alt key. The image will turn black where it is not over exposed and white where it is. It will also display colours which represent where individual colour channels have been blown. White means that every channel is blown.
Join to see EXIF info for this image (if available)
By cropping in on one of the offending sections we can see this in more detail.
Join to see EXIF info for this image (if available)
Join to see EXIF info for this image (if available)
To recover these pixels we need to go to the highlights and whites sliders and pull them to the left (negative values). This is a trial and error process. You can follow the histogram, you can use the blown pixel alert (the Alt key trick), but ultimately it is looking at the image itself that should be your guide. In this case I've used the values as shown.
Join to see EXIF info for this image (if available)
Checking my blown pixels I now get this. I haven't eliminated all of them, but as I alluded to earlier I believe that an image should ideally just kiss the right hand edge of the histogram. Not all the time, but most of the time.
Join to see EXIF info for this image (if available)
The entire image is shown below...
Join to see EXIF info for this image (if available)
...and the same crop that we looked at earlier, but now less retina searing and with some detail returned.
Join to see EXIF info for this image (if available)
The histogram as it currently stands. You'll see that the spike on the right which represents the brightest, whitest parts of the image has now been pulled back from the very edge of the histogram.
Join to see EXIF info for this image (if available)
To be continued...