There are many sources of information describing the scientific and technical aspects of modern digital camera functionality and design. And there is also a fair amount of proprietary technology used by various camera manufacturers which is patented intellectual property and not readily available to the public. Certainly the vast majority of camera users who simply strive to make lovely images with a digital camera can still learn to do so without multiple advanced degrees in optics and digital electronics. But a grasp of certain fundamental digital imaging concepts can still help us to improve the results we get from our cameras. Suffice to say, this article is written for the layman and intentionally sacrifices technical depth for the sake of simpler understanding. I hope it will be helpful to some of our readers.
What are Megapixels?
At the heart of every digital camera, is a relatively small electronic sensor chip which contains a vast array of tiny photosensitive elements called pixels. This sensor array plays essentially the same role that a frame of film does in film cameras. These tiny photo-sites collect the light coming through the camera lens, which is then recorded as data and processed electronically to ultimately become our digital photos. One of the most commonly misunderstood specifications of digital cameras is the quantity (measured in millions) of these tiny pixels contained on the sensor… aka the number of megapixels (mp). Camera buyers often assume that a 12mp camera is better than a 10mp camera, or that one with 24mp is superior to one with 16mp. And while there certainly can be advantages to higher resolution pixel arrays, the size and spacing of these pixels and the precision of the data they record is also crucial to the quality of the images ultimately rendered. Various classes and models of cameras use different sized sensors as well as different numbers of megapixels, and larger sensors generally offer superior image quality and dynamic range at the pixel level. So a full-frame DSLR camera with its large sensor will typically record a higher quality image than a Micro Four Thirds (MFT or m4/3) camera using a medium sized sensor, and a much higher quality image than a small-sensor compact or camera phone. These differences in sensor sizes are generally much more significant to image quality than the number of megapixels they contain.
What is Pixel Peeping?
A lot of images are never viewed at the full-resolution recorded by the camera sensor however. Indeed the vast majority of digital photos are viewed only on relatively low-resolution displays like computer monitors, TV or cell phone screens. These display devices typically have resolutions of 2mp or less, so to display the images in these smaller media formats a re-sampling of the image is done essentially discarding a large percentage of the data originally captured by the sensor. Under these circumstances it may be difficult to detect the image quality differences between the outputs of various cameras despite major differences in their size or cost.
So why do so many photographers and even casual camera buffs continue to long for that next “better” camera model with the superior “image quality”? Well there’s more to a quality image than pixels of course, but that’s what this article is about… pixels.
There are endless disagreements amongst photographers about the usefulness of viewing images at full resolution, (often referred to as pixel peeping). Many people feel that pixel level quality is unimportant since their images will never be presented at 100% size. Many also know that the quality of the images their cameras produce at full size is not too good, and so they avoid examining the quality of their photos at the sensor level. But there is a powerful case to be made for pixel peeping, and I’d like to share some thoughts on that now.
First, the sharpness and focus of any given photo taken is often not easily discerned when viewed at reduced resolutions. If a series of shots is taken of the same scene (especially hand-held), the focus and motion blur caused by camera and/or subject movement will often vary significantly amongst them. Here’s an example…
At reduced resolution these consecutively shot frames may appear to be of equal quality:
Choosing which frame of any such series to process and/or print is obviously better done at high-res. The same is true for assessing variations in color, exposure and artifacts like noise and lens induced aberations (eg: corner softness and fringing). All of the qualities and/or deficiencies in any given photograph are better and more easily evaluated at high resolutions. This is why slide film photographers have always used a loupe or magnifier to analyze and select their best shots for processing.
For bird and wildlife photographers (or anyone shooting distant objects at longer focal lengths) there can be another important reason to pixel peep… cropping. Even when using the longest telephoto lenses, wildlife is often still too far away to compose a shot where the subject bird or animal is large enough in the frame to present them as we’d like. In these situations cropping is frequently used to allow the target object to be shown larger in the finished image. Using post-processing software to remove significant portions of the captured frame lets us present distant subjects larger in the processed photo, but doing so also means displaying the kept portion significantly closer to the full resolution of the sensor. The high-resolution image quality of the sensor thus becomes much more important and evident as deeper crops are employed. Less detail is lost to re-sampling, and any artifacts or imperfections in the pixel level image is also revealed.
Conversely, down-sampling the images captured by high resolution sensors reduces both the fine detail and the artifacts seen when viewed at reduced sizes.
Finally, for those who post-process their images, nearly every adjustment that is performed in software can be made more judiciously and with greater precision when done and evaluated at high resolution. The results of better selection and processing afforded by working on your photos at higher resolutions is most beneficial when printing or displaying them at larger sizes. But it can improve the quality of your images at whatever size you view or present them, in any media format, and no matter which digital camera was used to take them.
There’s always a lot going on in the yard this time of year, and we’ve been doing a lot more birding than blogging lately. But I wanted to just do a quick post now to report that our dominant male Ruby-throated Hummingbird, Buzz, has returned once again from his long Winter migration. Some of you may remember Buzz from our posts last season, telling of our discovery of his band and his subsequent reunion with his bander, Lanny Chambers. https://kenn3d.wordpress.com/2012/07/19/a-20-second-buzz/
Anyway, here is Buzz… back again and taking up his claim on the territory.
From time to time I’ll quote from and/or link to various forum threads we’ve participated in. This article is based on excerpts from recent conversations at Birdforum.net http://www.birdforum.net/showthread.php?t=240202&page=25 Discussions like these include questions and answers from the range of experience, with newbies and skilled photographers freely sharing info and advice. These forums are a valuable resource to us, and are highly recommended.
The Case for Manual Exposure:
I prefer to shoot in Manual mode, generally avoiding all the automatic features of the camera save for AF. I begin by assessing the prevailing light and manually choosing an appropriate minimum ISO for my intended target area. My preset C1 settings set ISO80, Manual mode, full-zoom with aperture full-wide, so in decent light I generally only need to frame the shot and adjust the shutter to expose. With practice this becomes quicker than you might imagine, and not too different than trying to guess how much EV compensation will be required to correct for however wrong and in which direction any given AE metering will likely be. An often overlooked advantage of Canon SZ cams is the “live-view” EVF preview which simulates both exposure and DOF (wysiwyg) when shooting in Manual mode. This puts exposure control back in the hands of the photographer. This behavior is unique to the Canon superzooms afaik, and it’s a priceless feature imo… despite the other short-comings of the EVF.
I’m sure most users will prefer to use one or more of the semi-auto modes in various situations for a variety of valid reasons. Temple prefers Av mode with both the SX50 and SX40, and I also use Av mode occasionally when speed is of the essence. I believe Canon has done a good job with SX50 mode processing in general. They provide users at all skill levels with a capable and flexible camera. But I shoot all our Canon cameras in Manual mode most of the time. So with due respect to the preferences and techniques others prefer, I’ll just explain my own thinking on this and hope not to cause any angst. And then I’ll share some example images which illustrate the benefits of manual exposure.
So, just what’s so different about shooting in Manual mode, as compared to Auto, Program, Av, Tv, or even Sports mode for that matter?
The big difference for me, is that in all the AE modes, exposures are determined by algorithm. And however fast or clever that code may be, it simply does not see with the eye of this human photographer … More often than not AE does not expose my images as I would prefer. Further, depending on the complexity and lighting in the scene, any slight reframing (intentional or otherwise) prior to focus lock can cause the camera to alter exposures (often dramatically) even after EV comp has been dialed in. And, when you half-press to lock focus the AE program also locks the exposure in accordance with whatever “it thinks” is correct or best for the framing in place at that focus point, without regard to what the photographer sees or wants to see. Any subsequent reframing (intentional or otherwise) before releasing the shutter does not automatically re-adjust the exposure accordingly even though the composition may have changed significantly, and no manual re-adjustment or further compensation is possible without also refocusing. So if you lock focus on the Raven’s eye in bright light, expose the deep textures of his black head, and then reframe to include the white Mouse in his talons…you can be pretty sure that all the detail in that poor Mouse’s fur will be blown out in your image. If the exposure is not correct when focus is acquired, the captured frame will not be exposed correctly. These complex auto-exposure calculations are all done pretty quickly I grant you, which can be a major advantage, but it’s obvious that this method simply doesn’t always work so well, with blown highlights, blocked shadows and harshness or imbalance in overall lighting even when clipping does not occur… all being very common in auto exposures. And I find it all rather just too fussy anyway.
In Manual mode there is never any need to fight with (or compensate for) the vagaries of the Auto Exposure program. Whatever metering mode is chosen only appears as a light meter guide display and has no effect whatsoever on the exposure of your image. I can adjust ISO, shutter speed and/or aperture to obtain the desired exposure and DOF for my scene based on what I see in the viewfinder. I can lock focus wherever I choose and reframe as needed without fear of the camera altering any setting I’ve chosen. And I decide what compromises should be made to optimize the quality of my shot.
You must select a reasonable ISO for the prevailing light of course, but you can choose the best (lowest) ISO appropriate to the available lighting and never have to bemoan the Auto-ISO program having chosen a higher one than necessary. And you must also know what minimum shutter speeds you need for a steady camera… which with good technique and the SX50s excellent image stabilizer can often be slower than would be chosen by the camera in AV mode. But by not relying on Auto-ISO and Auto-Exposure algorithms the photographer can pretty easily control all these settings to properly adjust his image prior to releasing the shutter, without any unexpected or unwanted intervention from the processor. And he can have much better confidence that what he sees in the viewfinder is both what he wants, and indeed what he will get.
The camera usually makes all of these settings decisions faster than any human could, and my choice to shoot in Manual mode means that whatever exposures I capture (and any missed opportunities that I don’t capture), will be my fault. But I usually prefer the results I get by making these decisions myself.
The following slides were created to illustrate how the high-resolution IQ of the SX50 expands the optical focal lengths (fov) presented in cropped images. But they also provide examples of common scenes which for reasons of focus, lighting and shadow, could be problematic for the AE modes to expose properly.
In this segment we share some further descriptions of our method for processing the RAW images we’ve taken with the SX50. The examples illustrated here were converted with Digital Photo Professional, which is the Canonware editor included with the camera on CD. We have also made some conversions with the DXO Optics Pro 8 trial, and hope to explore Lightroom 4 as well shortly. We may have more to say about these other software editors also in a future article.
First, let me say that just as we prefer to shoot in Manual mode, we also always process each image individually and manually rather than use any scripted, batched, or automated processing. Every image is different, and if we select any given capture for processing, it will be processed based on its own individual characteristics. So you won’t find any prescribed settings or processing recipes here. I will however, outline our workflow and share the reasoning for the steps and methods used.
Converting RAW to TIFF
We begin by bringing the chosen RAW (.CR2) file into Digital Photo Professional, and selecting “Before/after compare” in the View menu.This allows continuous judgements to be made as we develop the image. The available adjustment tools under the RAW tab are arranged in a logical sequence from top to bottom in the tool pallette. In this example the Brightness adjustment slider is reduced to correct for the somewhat harsh lighting in the shot… (note the recovery of highlight detail this provides).
Several options are then provided to correct White Balance. The drop down menu defaults to “Shot settings”, which is the WB selected in the camera when the image was taken. We usually set the in-camera WB to either Daylight or Cloudy depending on weather conditions, because Auto-WB can change unpredictably based on the specific light and color within the focus area while shooting. In this image however, I found the Auto WB menu choice provided more natural color for this scene, and so the Daylight setting in the camera was over-ridden. Alternatively, an eye-dropper tool allows selection of the white point by clicking within the image, or the Tune button can be used to fine-tune WB based on a standard color temperature wheel. Customized settings are saveable as presets for future selection, but we don’t really use this feature.
The next section provides a drop-down menu of preset Picture Styles which automatically set all of the individual settings below the histogram to a variety of predetermined “styles”. We use this selector to choose a best starting point for each image (in this case “Faithful”) and then further fine tune the settings as needed. Clicking the image toggles the view scale between “fit to window” and 100% views (we recommend using high resolution views when adjusting sharpness settings), … and additional viewing scale choices are available both in the View menu and by way of cntl-key shortcuts.
Noise Reduction, RGB tone curves, Chromatic Abberation, and Lens Distortion controls are provided under the RGB, NR/ALO, and Lens tabs. We are not currently using DPP for these corrections however, preferring to transfer minimally developed TIFF files into Photoshop for finishing. The File menu offers several options for saving and converting your developed RAW files into various formats, either for further editing or as fully developed images.
Processing the TIFF in Photoshop
After loading the converted TIFF image into Photoshop, our first step is cropping. Many and most of our finished images are cropped to some degree, even those taken at close range and full-reach. It is indeed rare (for us) that a chosen frame was composed perfectly in camera… so I’ll usually crop a little for balance and/or to remove unlovely elements from the edges and so on. This also makes further editing faster and more efficient.
Next we adjust levels. Photoshop allows finer adjustments of Highlights, Mid-tones, and Shadows (than can be done in DPP) via the histogram and slider controls in the Levels tool. Overall exposure adjustments are always best done on the cropped frame, disregarding any unused portions of the original capture.
Sharpening and Noise Reduction are both key aspects of post-processing, but unfortunately these are generally not compatible processes. The problem inevitably arises that NR will tend to reduce fine details, and Sharpening increases the visibility of noise. There are noise reduction products which go a long way toward eliminating this natural conflict, but when applied universally to the entire image some compromise is still unavoidable. So we apply these processes selectively whenever possible. Careful selection and masking allows noise reduction to be applied to the oof backgrounds separately. And then by inverting the selection, sharpening can be applied to the in-focus foreground only. Thus avoiding the sharpening of background noise, and the loss of detail in the subject. This method has the secondary benefit of subtly enhancing the perceived focus… which is sometimes seen as a kind of 3D effect.
Many photographers perform sharpening as the last process before saving the finished image btw. But we do a very slight “pre-sharpen” at full size prior to resizing for web presentation, and another “finish sharpen” pass at screen resolution.
So here then is the finished image from the conversion sample at the beginning of this article:
Processing should usually be subtle, but the effect it has on our images can often be profound. Here are a couple more “before and after” slides illustrating the processing of RAW images from the SX50:
Backyard Birding with Kenn and Temple originates in the midwest heartland of the United States, and primarily features the native songbirds common to this region. Naturally, the biggest part of our readership comes from the North American continent. But we really want to thank all our readers from around the world for visiting the blog and sharing their thoughts and comments with us. In the short time we’ve been publishing, we’ve had visitors from 130 countries all over the world, and it’s obvious that our fascination with our birds is a pretty universal human characteristic. Sharing our birds with so many bird lovers in other lands greatly adds to our own joy of them… So thanks everybody, and happy birding wherever you live.
We’ve had our SX50 for about 3 months, and until now all the images we’ve posted have been shot as what we call “detuned” jpegs… where the camera’s jpeg engine is used to render and compress the image using minimized or reduced settings for NR, sharpening, saturation, and contrast. This method is somewhat analogous to shooting RAW, because most of the processing is then done in post rather than in-camera. We have been pleased with the results we get shooting this way with all our cameras, and recommend it to anyone who prefers not to deal with RAW processing for whatever reasons.
I had tried shooting the SX50 in RAW just briefly early on, and my initial efforts yielded somewhat mixed results. There was some additional detail in the RAW files and more latitude for exposure correction. At base ISOs the noise in oof backgrounds was surprisingly slight and easily dealt with, but in-focus subjects were both noiser and more artifacted at the pixel level. It required a fair bit of additional post-processing effort to achieve visible improvements in finished image quality at screen resolutions, when compared to our processed JPEG files. Writing the much larger RAW data files to the memory card also takes a toll on continuous shot to shot performance, so we just returned to shooting JPEG only instead.
I’ve recently revisited shooting RAW with the SX50 and am actively researching software and workflow variations which may allow us to further optimize the value of this feature…
Here are a few recently taken SX50 RAW images which have been converted with Digital Photo Professional (the included Canonware) and finished in Photoshop:
I hope to try using Lightroom, Adobe Camera RAW, and some other applications which support SX50 RAW files, to see if we can develop a workflow and/or results that encourage us to shoot the SX50 RAW more often.
EDIT 3/5/2013 : Some additional RAW samples are now available here: CR2 Gallery