Photography tips
This page contains a selection of key technical tips. To see a wider range please visit my blog and click on technical under the post categories on the right of the page.
Metering and exposure
The amount of light falling on a subject, or incident light, is what determines exposure. Please read that sentence again. This is what you measure with an incident light meter (by pointing the light dome at the source of light) and is the most reliable method for determining exposure. It's much more convenient though (imagine if the tabby cat above was a tiger) to have a light meter built into the camera. This meter, however, can only measure reflected light from the subject - from this the camera must then infer the amount of incident light illuminating the subject and this is where the problems start. Replace the tabby cat above with a white cat and then a black cat. We'll keep the incident light constant and therefore the exposure will also remain constant - remember, exposure is determined by incident light. The white cat will reflect more light to the camera than the tabby cat, which will reflect more than the black cat. Cameras are programmed to assume that the average reflectance for a scene is 18% - equivalent to a mid-grey tonality - which indeed does work very well for most scenes. The camera does not understand that the three scenes with the cats have different brightnesses due to differing amounts of reflected light, it assumes that the incident light is different in each case and thus determines different exposures for each of the scenes. The solutions are:
i) use an incident light meter (not so practical when the cats are tigers).
ii) meter off a mid-grey subject in the same light - either use a grey card or learn to spot mid-grey tonality by eye. You can zoom in on a particular area to meter or spot-meter, if you have the option.
iii) exposure compensate. In the case of the white cat the increased reflectance fools the camera into thinking that the incident light is stronger than it really is, so the camera ends up underexposing - therefore you need to 'open up' the camera and increase the exposure. The trick is knowing by how much. In the case of the black cat the camera overexposes and so you 'stop down' the camera.
iv) if you shoot digitally I'd suggest you get a feel for the histograms that your camera generates and calibrate the brightness of your camera's LCD to match that of your monitor - this gives you a good notion of your exposure when you're in the field.
Depth of field
The most important variables in controlling DOF are:1. The aperture of the lens. As the aperture becomes bigger, the DOF becomes smaller.
2. The magnification of the subject. As the subject increases in size in the image, the DOF becomes smaller. You can easily illustrate this by picking up an object and moving it closer to your face - watch the background blur as your eyes' DOF shrinks. You can increase the magnification of your subject by either moving closer to it (reducing the focal distance) or by using a telephoto lens (increasing the focal length).
The magnification of the subject depends upon the focal distance (distance between subject and the focal plane of the camera) and the focal length of the lens. Note the key difference between focal distance and focal length! It is possible to keep a constant magnification of the subject, and hence DOF, whilst changing the focal length of the lens by making changes to the focal distance. This may seem like nit-picking but it's worth getting your head around.
You can convince yourself that DOF depends on magnification, rather than focal length, by taking a series of shots of a subject at various focal lengths and keeping it the same size in the frame. The DOF appears independent of focal length.
Let's say you're using Canon's 24-105mm f/4.0 L IS zoom lens. You focus on your model at a focal length of 105mm in aperture priority mode at f/4.0. You take your first shot, which with these settings will have a modest DOF. You zoom out 10mm to 95mm, making your subject smaller in the frame, which means you have to move in closer to restore your model to their original size. You repeat this until you're at a focal length of 25mm. If you were to estimate the DOF in each shot by eye you'd say it was constant.
It isn't exactly the same, however. At shorter focal lengths the DOF does increase minimally, even if the subject stays the same size, which means that DOF is not completely independent of focal lengths. The impact of this is small compared to the impact of magnification though.
The following is also worth noting. Most photographers are interested in a shallow DOF for blurring backgrounds. DOF extends in front of and behind the focal distance, but the ratio between the two changes with focal length. At shorter focal lengths the DOF extends approximately 1/3 in front of the focal distance and 2/3 behind. At longer focal lengths this changes to 1/2 in front and 1/2 behind. Therefore longer focal lengths have an advantage for blurring the background.
It's important to note that moving closer to a subject will affect its appearance due to changes in perspective. It is possible to have a blurred background when shooting a model at shorter focal lengths, but the reduced focal distance will distort their features due to the impact on perspective. If you want to shoot people with a blurred background then use a telephoto lens.
Exposure Compensation
1/400, f/2.8, ISO 1250, -1 EV, 200mm (70-200mm f/2.8 L IS)
1/80, f/2.8, ISO 3200, +1 EV, 16mm (16-35mm f/2.8 L)
Since including EXIF data with photos I've had a few questions about exposure compensation.
Cameras meter using reflected light and are programmed to assume that the average reflectance for a scene is 18% (equivalent to a middle grey tonality). The result of this is that all scenes will have middle grey tonality unless you compensate for, or bias, the exposure.
If a scene is darker in tone, you need to reduce the exposure so that the scene remains darker, and is not over-exposed as middle grey.
If a scene is lighter in tone, you need to increase the exposure so that the scene remains lighter, and is not under-exposed as middle grey.
Most scenes do have mid-grey tonality (which is why cameras are programmed this way) but the essential skill to develop is to recognise when they don't. A couple of examples for you.
In the top shot I felt that tones darker than middle grey predominated so I reduced the exposure by one stop (-1 EV) to avoid over-exposure. In the bottom shot lighter tones predominated, so I increased the exposure by one stop (+1 EV) to avoid under-exposure. It takes time to get a feel for how much to adjust exposure by - like many things in life, it's a question of practice. Check out the EXIF data I include in the selection of photos I publish on my blog from every wedding.
An alternative method is to spot-meter a small area in the scene which has middle grey tonality. Whichever method you chose, however, requires you to develop your eye for tonality. This is an essential skill to develop as a photographer. It truly does lead to enlightenment!
Focusing Tips
I've had a couple of readers ask me how I get sharp images in low-light conditions. The following are some points to consider when focusing:
1. Focus on points of good contrast - eyebrows and eyelids for a portrait shot, for example.
2. The central focus point is the most accurate and, with lenses of widest aperture f/2.8 or less, additional sensors come into play (see diagram above). In low-light conditions I use only the central focus point.
3. The outer focus points are fine with good lighting conditions but I find that they are not as reliable in low-light conditions.
4. I tend to avoid focusing and then recomposing at wide apertures. Depth of field (DOF) can be extremely small and recomposing can leave your subject out of focus (see diagram above). This is a judgement call though, since it depends upon the precise aperture, and very significantly, distance to the subject. You need to develop an instinct for judging DOF.
5. As a result, when I encounter tricky conditions, I often 'shoot and crop'. Use the central focus point and take the shot without recomposing. Crop during post-production to adjust composition. The greater resolution of the 5D Mark II will be invaluable for this method.
6. Light levels can drop so low that auto-focus is unable to work. In this case an AF assist beam from a Speedlite or the ST-E2 transmitter is necessary. If your eyesight is very good, you can try manual focus.
Contrast & Curves
And if you're a digital photographer so should you, since it gives you complete control over the contrast in your images.
Here's a shot of my chum Jack with the corresponding 'Curves' dialogue box in Photoshop. In light grey within the graph is the RGB histogram. To make it simpler to understand what curves can do, let's discard the colour information from the shot by converting it to greyscale.
The horizontal (X) axis runs from 0 (black) to 255 (white), encompassing all shades of grey. It corresponds to the 256 tones being input into the 'Curves' process. The histogram shows the relative number of pixels at all 256 of these tones. The majority are towards the darker end of the scale - the image can be described as 'low key'. The vertical (Y) axis is identical to the horizontal axis except that it represents the tones being output from the 'Curves' process. At the moment our 'Curve' is the line indicated in red. Choose any point on the X-axis, the input tone, travel vertically until reaching the red line or 'Curve', and then travel horizontally left to the Y-axis to see the output tone. At the moment 0 maps to 0, 255 to 255, 127 to 127 etc - the image remains unchanged.
Let's introduce some extreme contrast! The red path now shows the 'Curve' that I've input - look at the affect this has on the image. All input tones from 0 to 130 have been mapped to an output of 0 (black) and all input tones from 131 to 255 have been mapped to an output of 255 (white). The resultant image now contains only black and white - the ultimate in contrast but not so useful as a treatment for our images.
A more pleasing result is seen with the path in red shown above - known as an 'S-curve'. Darker tones have been darkened slightly and lighter tones lightened. The slope of the line in the midtone region has increased leading to a more subtle increase in contrast.
The same process can be applied to colour images but the extreme contrast procedure would make your image look rather odd (which is why I converted to greyscale). In Photoshop curves can be applied to colour channels independently giving great control over colour balance.
Histograms
Each pixel in a digital image has a corresponding light-collecting 'photosite' on the sensor of the camera that took it (on the 5D Mark II there are 5616 x 3744, or 21 million, of them!) In the majority of digital cameras each photosite can only record light intensity for one of the primary colours (red, green and blue), assigning it a number between 0-255 (256 values = 2 to the power of 8 = 8-bit). For each photosite information about the other two primary colours is deduced from adjacent photosites using a process known as 'demosaicing'. The Foveon sensor, featured in Sigma's cameras, is the exception - it records the intensity of R,G & B light at each photosite. This information can then be converted and displayed as a digital image or graphically as a histogram.
I took the above 'snap' of my daughter Emily this morning, deliberately including some vibrant primary-coloured objects, to give clear histograms.
Pressing the 'info' button whilst the image is displayed on the LCD of the 5D Mark II gave the above screen which shows a brightness, or luminance, histogram.
Pressing the 'info' button again leads to separate histograms for R, G & B light (the order in which these screens are displayed can be adjusted by a setting in the menus).
In Photoshop the above histograms can be displayed for the image - the 'RGB' and 'colours' histograms are in addition to those that the camera displays.
The horizontal axis on all these histograms represents brightness values which run from 0-255. The vertical axis represents the number, or frequency, of pixels occuring with this brightness value.
It is not always possible to expose a shot so that it is cleanly captured within the brightness range of these histograms - the dynamic range can be too great for the camera (think outdoors on a bright sunny day). In this case some of the tonal detail of the shot will be lost or 'clipped' - it will be either pure black or pure white. One of the benefits of shooting RAW is the greater tonal range that can be captured.
I tend to pay most attention to the brightness, or luminance, histogram as it corresponds most closely to the scene as our eyes see it (the 5D Mark II displays a luminance histogram but many cameras only display the RGB variant). The histogram is calculated by taking a weighted average of the R, G & B light at each pixel, with more emphasis placed upon green light as the eye is most sensitive to this. As a result, notice that the luminance histogram bears most resemblance to the green-only histogram.
The separate colour histograms can be used to see if individual colours have been clipped.
The RGB histogram is purely a summation of the colour histograms and can be used to assess colour clipping - but as a result of this simple maths, 'pixel location information' is lost. The histogram can show strong clipping at the far right of the histogram, due to bright R, G and B pixels at separate locations, without there being any pure white pixels in the image. A white pixel requires values of 255 for all three colours at the same location.
The 'colors' histogram is an overlay of the separate colour histograms and is a convenient way of visualising all three at once - Lightroom displays this.
Manual Mode
Exposure is determined by incident light. Cameras, however, measure reflected light. Therefore, to estimate the incident light from the reflected light measurement, cameras assume that a scene has an overall reflectance of 18% (equivalent to a tonality termed middle grey). In most cases this works very well, but when the overall tonality of a scene is not middle grey, a situation which occurs frequently at weddings, exposure inaccuracies result. Exposure compensation is one way of correcting for this and is a technique that photographers need to be familiar with. Under the correct circumstances though, an alternative approach is possible - using the manual mode of your camera.Let's say you're shooting outside at a wedding reception. All the guests are standing in open ground, not shaded by trees or buildings. It's a lovely sunny day and there are no clouds in the sky. In this case all of the guests are being lit by the same amount of incident light and the exposure settings will therefore be identical for any shot you take under these conditions. Once you've found an exposure that you're happy with, using the camera's reflected-light meter and by checking the histogram, you can put these settings into your camera's manual mode and shoot away without worrying whether a scene's reflectance deviates from middle grey. If you keep an eye on your camera's light meter, you'll notice that when taking shots of the bride in her white dress the camera will think you're overexposing (which would have resulted in underexposure if left in an automatic mode such as aperture priority, requiring positive exposure compensation) and that with the groom in his dark suit the camera will think you're underexposing. You know, however, that your exposure is spot on and that it's the camera's reflected-light meter being fooled. If you do this, don't forget to regularly review your exposures, just in case a cloud has crept in front of the sun and reduced the amount of incident light!
Manual mode is also extremely useful when shooting indoors in low light. Let's say, for example, that the maximum ISO setting you're prepared to use on your camera is 1600, that the minimum shutter speed you're prepared to use when shooting people is 1/60 sec, and that the maximum aperture of your lens is f/2.8. These personal criteria determine the maximum amount of ambient light that you can capture.
Let's say you're in a dark wedding reception venue. To correctly expose for ambient light requires an exposure of 1/15 sec at f/2.8 and ISO 1600 - 2-stops lower than your exposure criteria. What you can do in this situation is to dial 1/60 sec, f/2.8, ISO 1600 into manual mode, underexposing for the ambient light, and then use flash to illuminate your subjects. You'll be capturing the maximum amount of ambient light allowed by your exposure criteria and not allowing your camera to give an unusably slow shutter speed, which would occur if it tried to correctly expose for ambient light.
manual, 1/50, f/2.8, ISO 1600, 0 EV, -1 FEC, 35mm
If you've not done this before you'll probably be surprised to see how much detail you capture in the background with this amount of underexposure. The image above was taken using this approach and the background is over two stops underexposed (and I've added a vignette).
1/25, f/2.8, ISO 6400, +1 EV, 16mm
If you have a DSLR which performs well at high ISO, and you're happy to use these settings, then you can expose correctly for just ambient light in very low light settings. Check out the shot above taken with a Canon EOS 5D Mark II.
Lenses for Wedding Photography
So which are the 3 must-have lenses for a wedding photographer? The lenses that I currently take with me to all weddings are the 15mm f/2.8 Fisheye, 16-35mm f/2.8 L II, 35mm f/1.4 L, 24-70mm f/2.8 L, 85mm f/1.2 L and 70-200 f/2.8 L IS. Bear in mind that I'm using a full-frame Canon EOS 5D with no crop factor. Rather than speculate as to which of these lenses I think are most important I've used Lightroom 2 to analyse the metadata from my last 10 weddings in order to see how frequently I actually use each of the lenses (I need to get out more!). Numbered weddings in the first column, percentage use of each lens in the following columns and average use in the bottom row.
A few observations:
1. Perhaps I should remove the 15mm f/2.8 Fisheye from my bag! I only used it at wedding 3 (for some shots on a beach).
2. The focal length range 16-70mm completely dominates every wedding I shoot. The 16-35mm f/2.8 L II is my most used lens. I think this reflects the compact nature of many wedding venues and my desire to shoot people with their environment. The 24-70 f/2.8 L is my next most frequently used lens.
3. The 35mm prime lens is used less frequently at weddings with shorter coverage time. I stick to the zooms to save time.
4. My 85mm f/1.2 L is primarily employed for the bride and groom portrait session and usage reflects the length of time that is allowed for this aspect of the wedding.
5. Weddings 2, 4, 5 and 7 were mainly shot indoors due to poor weather. The 70-200 f/2.8 L lens was then often unused.
6. The length of coverage, structure of the day, nature of venue and quality of weather all have an impact on lens usage.
So on the basis of this analyis my recommendations would have to be the 16-35mm f/2.8 L, the 24-70mm f/2.8 L and the 85mm f/1.2 L (or cheaper f/1.8 variant) for the portrait session - that dreamy bokeh is just irresistible for weddings. A pretty portable set of 3 lenses.
You will always be able to find a negative review of any piece of kit - people are much more likely to write one if they've had a bad experience. Lenses, in particular, show variability and it is possible to buy poor quality examples of them. If you're not happy with a lens then ask for it to be replaced - the more supporting evidence you can provide for this claim, the more willing the vendor will be to help you (theoretically).
My digital wedding workflow
In response to a reader's question I present my digital wedding workflow. It's very efficient and allows me to finish processing a wedding by the next day.
1. RAW files loaded from memory cards to a folder labelled with couple's names.
2. Imported into Lightroom.
3. Quick eyeball of all photos in library module. If a shot is spoilt by blinking or grimacing, or if someone has walked into shot, use X key to mark image as reject.
4. Verify and then delete rejected shots.
5. Work through all the images in the develop module (I can do 200 shots an hour when I'm in 'the zone' - use the paste develop settings within Lightroom to save time).
6. Tonal adjustments are the major correction to make. Tone curve to strong contrast (I apply this to all the images in two steps by copying the develop settings from the first shot and then pasting to all the rest). Recovery and exposure sliders to adjust highlights. Fill light and brightness to adjust shadows (this can introduce considerable digitial noise if the image was shot at high ISO). Contrast and blacks to further tweak contrast.
7. Adjust white balance and vibrance. If interior shots are bathed in fluorescent light, adjustment to tint (green-magenta shift) is necessary.
8. I use the 'Previous' button at the bottom of the develop module extensively. It pastes the previous photo's develop settings. Notice that I am not making changes that are unique to a shot at this stage (such as crops and rotations) as I do not want to paste these between images.
9. I now export all the images as JPEG files (Quality 80%, resolution 240) and use the post-processing 'Export actions' facility within Lightroom. I've created a Photoshop droplet based on an action which removes digital noise using Neat Image and then sharpens the image (file size can expand to 6MB). You need the pro-version of Neat Image to do this. I've created a custom noise profile for my EOS 5D - Neat Image can read the ISO setting in the EXIF data and apply the right level of noise correction. I prefer Neat Image's noise removal and Photoshop's sharpening to the equivalents in Lightroom. 10. I have a top-of-the-range MacPro but it still takes 3-4 hours to complete the above step - I often run it overnight.
11. I import these files into Lightroom with the couple's names and the suffix 'tweaked jpegs'.
12. I now adjust cropping, straighten shots and make any other minor tweaks which are necessary.
13. I choose my picks from these to make up my web slideshow.
14. I export all the shots twice as high res (files now typically 2MB) and low res (resize 1000 x 1000, resolution 72, typically 300KB) images which are burnt to DVD.
15. If I want to do any really special treatments I can use the adjusted RAW files and export to Photoshop as 16-bit TIFFs.
Exposure values (EV)
I need to introduce the concept of exposure values (EV) for some forthcoming posts. EV is an absolute measurement of the light quantity required to give a correct exposure at a specific ISO speed (it depends upon aperture and shutter speed) - though most charts only give these values at ISO 100 (these are then also known as light values). It's useful to be able to refer to the amount of light using just one variable rather than a combination of aperture and shutter speed. In the table above EVs run from 0 to 24 (they can be negative as well). The scale is based upon stops of light - ie doubling or halving of light. For a given EV there are many combinations of aperture and shutter speed which can be used to make a correct exposure.
Some examples of EVs to help you get a feel for the scale:
EV 17 white object in full sunlight
EV 15 noon daylight
EV 13 bright, cloudy days
EV7 indoors
EV -5 scene lit by the moon
Knowing the amount of light in a particular lighting scenario allows you to construct guidelines, such as the 'sunny 16 rule' - on a clear, sunny day set your camera at f/16 and the shutter speed to 1/ISO setting - this equates to a constant EV of 15.
You'll also see these values in the technical specifications in your camera manual. For example, my Canon EOS 5D autofocus system works in the EV range of -0.5 to 18 EV.
Sensor cleaning
DISCLAIMER: The only method Canon recommends for sensor cleaning is using a Giottos-type air blaster. Any other will invalidate your warranty. The following procedure carries the risk of sensor damage.
There are many advantages of the digital format over film but one angst-inducing problem that goes with the territory is sensor dirt. Even though I keep all my lenses very clean and exercise caution when changing lenses on my Canon EOS 5D, dust and dirt manages to sneak in, and I have to clean the sensor about every 6 months. There is risk involved with this procedure but I don't want to be sending my camera back to Canon for a clean every time. Shown above is my cleaning kit - a Giottos Rocket Air blaster, a 12-pack of Sensor Swabs (type 3 for the 5D) and Eclipse E2 cleaning fluid (for tin oxide coated sensors such as the 5D) - it cost me about £55 in total.
Most digital SLRs have a sensor cleaning mode which will lock the mirror up and open the shutter curtain to give you access to the sensor. Always make sure your batteries are fully charged before attempting this - if power fails the shutter will try and close. The air blaster should be your first port of call since you don't need to come into contact with the sensor. Hold your camera with the sensor facing the floor (so that gravity is on your side) and blast it with air - keep the tip of the blaster outside the body of the camera. This will remove any loosely attached pieces of dirt and may solve your problem. If you want to check if this is the case, take a shot of a plain-coloured surface with your lens stopped down to its smallest aperture. You can then import the shot into Photoshop and play with contrast and levels to enhance any specks in the image. Bear in mind that this is a very stringent test and most sensors will probably look dirty - is the dirt really an issue? Use your camera with typical settings for a few shots - can you see any specks? If you can and the images require clean-up in Photoshop then you may want to proceed to the next step.
Some dust seems to get welded to the surface of the sensor or, more accurately, the low-pass filter which covers the sensor, and no amount of air-blasting removes them. The sensor swabs are lint-free pieces of material attached to plastic paddles that are assembled in a clean room - buy the type that matches your sensor size. The E2 cleaning fluid is ultra-pure methanol or ethanol which leaves no residue on the sensor. In combination they reduce the likelihood of you introducing further contamination, increase the chance of you cleaning the sensor first time and reduce your stress levels. Like many things in life, the procedure is not risk-free. You wet the swab with a couple of drops of cleaning fluid, carefully bring the swab into contact with the sensor, and then sweep the swab over the surface (full instructions come with the pack). I had to use a couple of swabs on one occasion when my sensor was very dirty but it did the job.
If you don't feel confident about this, you can always send your camera to Canon to be cleaned.
Perspective and portrait lenses
The relative dimensions of an object alter as you change your spatial relationship to it - this effect is known as perspective (and is fundamental to an appreciation of composition). This was amusingly illustrated in an episode of Father Ted when he tried to explain the concept to the dim-witted Father Dougal. The small toy cow he held just in front of Fr Dougal only appeared to be the same size as a real cow in the distance due to perspective - alas the idea was too subtle for him to grasp!
The photograph above was shot with an ultra-wide angle lens at a focal length of 16mm. Not a pleasant sight! To fill the frame with such a lens requires you to get very, very close to your subject (about 5cm in this case). This means that the nose looks much bigger than we're used to seeing - thanks to the unusual perspective.
The above shot was taken with a medium telephoto lens at a focal length of 80mm and is much more pleasant to look at. To fill the frame with the longer (and hence more magnifying) focal length requires the subject to be a much greater distance from the camera (in this case about 2m), giving a more flattering perspective to the face.
This is why lenses in the focal length range of 70-130mm are referred to as 'portrait lenses'. They force the photographer to stand a sufficient distance from the subject in order to fill the frame which gives a pleasing perspective to the subject's face.
Basic Principles of Colour Management
I don't claim to be an expert on colour management but I think I understand the basics (if after reading this you think that I don't, then please feel free to post a comment and correct me). However, that's all I want to write about in this post - the basic principles.
In an ideal world you'd take a picture of, or scan in, the colour chart above and it would look identical to the original on both your monitor and when printed out. For this to occur your digital camera, or scanner, needs to record the colours accurately, your monitor needs to display them accurately and your printer needs to output them accurately. As I'm sure you're all aware, the world, alas, is far from ideal.
For an example of this struggle with colour, think about a TV showroom - loads of TV sets side-by-side displaying the same channel but with each picture looking very different. Not only is there huge variability between models but also variability between individual TV sets in the same range.
The bottom line is that electronic and digital devices struggle to handle colours accurately - they need feedback to know how well they're doing. We therefore measure precisely how the device is handling colour (as the spectrophotometer would be doing if the monitor in the above picture was turned on) and then construct a correcting colour profile for that device. This in a nutshell is colour management.
I'll concentrate on colour management in the digital photography workflow in a subsequent post. I colour manage my monitor with a Gretag Eye-One device.
Colour management part II - the digital workflow
As mentioned in an earlier post, digital and electronic devices struggle to handle colours accurately. However, this can be measured and corrected by colour profiles - a process known as colour management.Let's have a look at the digital photography workflow, which for most people will consist of a digital camera, monitor and printer. You could potentially colour manage all three devices and create individual colour profiles for them, but in practice, most people will use the generic profiles provided by the manufacturers. People with more than just a casual interest in photography, however, will tend to profile monitors as their colour output changes over time.
You can callibrate all three devices though. To do this you need a colour standard (something that defines what red is, etc) such as the colour chart shown above - these are expensive and have to be replaced regularly. These charts are supplied with software containing digital descriptions of the coloured squares on them - in the RGB colour space as red, green and blue channels (there's also saturation and luminosity information) which have values between 0 and 255 ie 8-bit. There are various RGB colour spaces with different colour ranges or gamuts - sRGB, Adobe, ProPhoto - for home digital work stick to sRGB. So to create a colour profile of your digital camera you'd take a lot of photos of the colour chart under varying lighting conditions and then use the software to compare to the standard and create a colour profile (essentially a database of colour corrections). I imagine that digital sensors don't produce significant colour shifts over time and production methods give good consistency between cameras as most people use the generic colour profiles that come with digital processing software - as in Adobe Camera Raw used by Lightroom and Photoshop - and don't have any problems colour managing their digitial workflow.
Monitors, however, are the weak link. As mentioned in an earlier post, think of the range of colours produced by TVs in a showroom all tuned to the same channel. All models have unique colour biases, there are big differences between individual examples of the same model (hence generic profiles are not as useful) and the colour bias shifts over time! It's for this reason that if you're serious about colour management (you're regularly sending your shots off to be printed elsewhere and the results are important) you need to profile your monitor - I use the Gretag Macbeth Eye-One Display 2. Let's say your monitor has a colour cast. You make corrections to your shot so that it looks good on your monitor but by doing so you've shifted all the colour data in the file you send off for printing! The profiling software displays known colours on your monitor and a spectrophotometer (see the picture above) measures the actual output. Comparison of the two allows a correcting colour profile to be constructed.
Finally you can do the same for your printer. Print off a colour standard and use a spectrophotometer to measure the output or send it off to someone who can. For home use I'm not so fussy (for weddings I use professional printers) so I stick to the generic colour profiles for printer, paper and ink combinations that came with my Epson printer - that's right these profiles are only accurate for the branded papers and inks. You should have a different profile for every paper type that you can use with your printer.
Using flash
When lighting for portraits, photographers talk specifically about key and fill lights. In the simplest scenario, the key light is the principle light source and determines shadow placement, and the fill light is used to reduce the contrast between highlight and shadow, by 'filling in' the shadow regions with light. A similar situation exists when you have to balance ambient and flash light - your flash can act as either the key or fill light. This is pertinent to the above question, since if you shoot in an automated flash mode, you have to be aware that the camera is making this decision for you, and can get it wrong.90% of my shots are taken in aperture priority mode, with evaluative metering, and I use my flash in E-TTL II mode (the other 10% of the time I use manual mode). This requires me to make both exposure and flash exposure compensations and requires me to think about what I'm doing. In fact, although there is no best method for using your camera, to expose shots correctly in a wide range of lighting conditions will require you to think, whatever the process you use. Cameras are not yet clever enough to do it all for you, although I hope one day they will be. In the meantime, it's all about finding a reliable and efficient method that works for you.
So in the above example I would typically (alas, there are always exceptions) meter the whole scene and then apply exposure compensation to adjust the ambient light to the level I required, followed by flash exposure compensation to control the level of flash light. The ambient light is easiest to get right. Getting the right level of flash can occasionally be tricky, particularly when the camera has been fooled and is trying to use the flash as a key light rather than a fill light - it attempts to flash light the whole scene, resulting in your subject being massively over-exposed (nuked!). In this situation, you can either dial the flash right down (up to -3 stops) or use flash exposure lock on your subject.
Using a StoFen omnibounce
The StoFen omnibounce is a very useful piece of kit but many people don't use it correctly.
I've seen quite a few 'professional' wedding photographers with an omnibounce attached to their flash in the above manner shooting outdoors. I presume they imagine that passing light through the plastic results in it being softened but it doesn't make any difference to the quality of the light. Softness depends on the size of the light source and that is not being affected by the presence of the omnibounce. What it is doing though, is reducing the output of the flash, since a lot of light is being directed elsewhere by the other four sides of the unit. If your shot looks better when you use the omnibounce in this way, it's because your flash has reached the limit of its power output, and you're seeing less light than the camera thinks is required to expose the shot - the same effect as negative flash exposure compensation (FEC). If you take the omnibounce off the flash and dial in negative FEC you should get exactly the same result, whilst also saving your batteries and extending the life expectancy of your flash unit!
The StoFen omnibounce can't work in isolation. It needs surfaces to bounce light off - as illustrated by the above diagram (taken from StoFen's website - note that they recommend using your flash at 45 degrees if less than 15 feet from your subject so that front-lighting doesn't dominate the shot). Omni refers to the fact that light is sent out in all directions (well, almost) and bounce that it then needs to be bounced back. Use it indoors and, provided the room isin't too big, you'll see a huge difference to your shots. The light is now being softened as the effective size of the light source is increased by light being bounced back from all directions. Your flash unit is effectively transformed into a bare bulb - great for off-camera flash.
I also use a Lumiquest ultrabounce which does exactly the same thing, but folds up flat and can be more easily attached to flash units covered in velcro.
Flash Gels
Flash light is balanced for noon daylight (5600K), so adding fill flash on a sunny day requires no correction. When shooting indoors, however, the ambient light you are most likely to encounter will be due to incandescent or fluorescent light bulbs. Incandescent bulbs generate light by passing an electrical current through a tungsten filament which then emits light as it becomes hot - the colour temperature of the light is related to the temperature of the filament, it is a black-body radiator. These bulbs contain either an inert gas (as in typical household bulbs with a colour temperature of 2500-2800K) or a halogen gas (as in halogen lamps, 3000-3200K). To correct for this you would typically use a colour temperature orange (CTO) flash gel - see one fitted to a Speedlite (http://www.davidfenwick.co.uk/blog/2009/01/lastolite-micro-apollo-plus-flash-gel.html). They come in various strengths, but in my experience, the most useful is a full CTO gel.Fluorescence is an entirely different process for light generation and it makes no sense to talk about colour temperature. The key thing for a photographer to know is that it gives your images a green tint - hence you attach a green gel to your flash to match this. In post-production you can then remove this green colour cast from the image.
Whether you balance the colour temperature of your light sources is a creative decision that you're free to make - there are no rules! A couple of examples:
In the shot above you can clearly see that the flash light is much cooler than the ambient lighting - no colour temperature correcting gel was used. It was a Xmas wedding, it was freezing outside and so I decided to give the shots a warm atmosphere - a 'cosy feel'.
In the shot above the flash light has been modified to match the ambient halogen lights - notice the blueness of the evening light coming through the sky light.
Lastolite Micro Apollo XL
David,Your flash use is lovely and soft. Which size of the Lastolite Apollo do you use?
Your blog states the XL version, but at 16 x 10 inches I don't see how this would NOT block the Speedlite IR and metering sensors. Am I missing something - do you tilt it up?
I want to get one but feel it must be the smaller 10 X 7 unit.
Thanks!
Tony B
As the cliche says, a picture is worth a thousand words!
Any light-modifying attachment for a flash that was not designed to work with autofocus assist, or an external metering sensor, would have a very limited appeal. You can rest assured that the Lastolite Micro Apollo XL is not one of these. I used it at about 30 weddings last year with no problems at all - in particular, no issues focusing in low light. In the image above you can see why - the majority of the unit sits above the flash head.
The bottom of the softbox is also slightly V-shaped to ensure it does not limit any functionality. In the shot above I set the 580EX II as a slave. In this mode the AF-assist light is used to indicate the flash is ready to fire and is clearly visible above.
A couple of useful tips.
The shot above is from the first wedding at which I used the softbox. From the reflection in the subject's glasses you can see that the light is not evenly distributed - there is a hot spot. To get around this, pull down the built-in wide panel of your Speedlite. This sets flash coverage for a focal length of 14mm and strongly diffuses the light. This does not cause any loss of light, as the inside of the unit is reflective, but does give even lighting.
The unit is attached to the Speedlite using velcro. After a few weddings I found the velcro starting to detach from the flash body when I removed the softbox. I therefore replaced the large rectangular strips of velcro with small black velcro dots. These are still firmly attached six months later.
Hope this reassures you about the XL version of the Lastolite micro Apollo. A bigger softbox equals softer light (and the closer you get to the subject the softer the light becomes).
First and second-curtain sync
It's worth reading the previous two posts before tackling this one. Except for when using FP mode on your Speedlite, the shutter is fully open when the flash fires. The shutter is fully open when the first curtain has come to rest, and before the second curtain has started moving - the time that the shutter is open is at least 10x longer than the flash duration. When does the flash fire? By default, at the first possible opportunity - when the first curtain has come to rest and the shutter has just opened. This is termed first-curtain sync. More sophisticated flash units also give you the opportunity to fire the flash just before the second curtain starts moving and the shutter starts to close - second-curtain sync. This is accessible on the 580EX II Speedlite by the button just below the red box shown in the previous post.
To see the impact of curtain sync, check out the images above. The camera was in aperture priority mode (the flash behaves differently depending upon which mode you're in - a subject for a future post) and to expose for the ambient light conditions a shutter speed of 1/2 sec was required - using flash with long exposure times to record ambient light is known as 'dragging the shutter'. In the first example I set the toy rolling (right to left) and used a remote shutter release to trigger the camera, mounted on a tripod, without firing the flash - the toy travelled for the entire duration of the shot and hence is completely blurred. In the second example I switched on the flash. It fired as soon as the first curtain had come to rest, freezing the motion of the toy, and then the shutter stayed open (to record the ambient light) for the remainder of the 1/2 sec exposure time - thus a sharp image of the toy was recorded (when the flash fired at the beginning of the exposure) followed by a blur - the result of which makes the toy appear as if it is travelling backwards. In the third example I set the Speedlite to second-curtain sync. Now a blurred image of the toy was recorded until the flash fired just at the last moment (before the second curtain started to move and close the shutter) to record a sharp image - in this case, the toy appears to be travelling forwards and gives a visually more appealing sense of movement.
High-speed sync (FP) flash
As I mentioned in my previous post, the EX-series of Speedlites have an option that allows you to use flash even when your shot requires a shutter speed faster than the X-sync (flash sync) speed. It's termed high speed sync mode (technically known as FP, or focal plane, mode for historical reasons, though this abbreviation has been referred to as fast pulse, which more accurately describes what the flash unit is doing) and is highlighted by a red box in the picture above. A quick remider - at shutter speeds faster than the X-sync speed only part of the sensor will be exposed to the scene at any one time, as one curtain starts to move before the other has come to rest. To overcome this, the unit pulses the light very quickly (50 KHz), effectively creating constant light for the duration of the shot - this is the elegant solution that FP mode offers. When I'm shooting outside I always have the flash unit set to FP. The price you pay is that the flash output is significantly reduced (to about a third), but by shooting at a wide aperture and getting close to your subject, less light output will be required as fill to balance the ambient, or key, light.
X-sync (flash sync) speed
The shutter in your camera consists of two metal plates (known as the first- and second-curtains) which travel sequentially (the time lag being determined by the shutter speed you set) to expose the digital sensor to light. They travel vertically, as the distance is shorter and allows faster shutter speeds. The sequence is illustrated in the above pictures. Once the shutter speed reaches a certain point (known as the X-sync speed), the second curtain is on the move before the first curtain has come to rest. This means that the sensor is only ever exposed to a narrow band of the scene at any moment. Have a good think about it - once you've got to grips with the concept we can explain flash sync and second-curtain sync.
A burst of flash lasts only milliseconds, and since you want all of the area in your shot to be illuminated equally, the sensor must be fully exposed to the scene when the flash fires. It's no good if only a narrow band is exposed as only this area will be lit by flash - the rest will be dark. Therefore the camera needs to synchronise the firing of the flash with the movement of the curtains. The fastest shutter speed at which this can occur is known as the X-sync, or flash sync, speed. The faster the curtains can physically move, the faster the X-sync on the camera will be. On a Canon EOS-5D it's 1/200 sec - not that impressive.
So what, you might ask. Unfortunately, this does create limitations (particularly for the long-suffering wedding photographer). Imagine you're outside on a very bright sunny day - it's midday and there's no shade available. You want to take a picture of the bride and you're aware that the lighting will create high contrast shadows under her brow (panda eyes) and nose. Since you're a competent photographer you decide to add a touch of fill-flash (maybe with some positive flash exposure compensation). You've stopped your lens down to f/22 and set your ISO speed as low as it will go (to get the slowest shutter speed you can) but the camera still only needs 1/1000 sec to expose the scene correctly. You switch the flash on to take the shot and you see the X-sync shutter speed flashing in your viewfinder. This is the camera telling you that if you use the flash you're going to get this shutter speed and consequently over-expose the shot. You could add a neutral density filter and polariser to reduce the amount of light reaching the sensor, but at a wedding you're unlikely to have the time. If you have an EX-series Speedlite there is an alternative escape route - I'll tell you about it in the next post.
Flash colour temperature
Light from flash units is typically balanced for noon daylight (5600K). Summer sunlight from 9am to 3pm is in the range 5400-5700K, so using your flash as a fill-in light during this time leads to a good mix between ambient and flash light. When you use your flash indoors and capture ambient light produced by tungsten lightbulbs (2800K), however, it's impossible to adjust the white balance of your shot correctly for both light sources. If you set the white balance for the flash light, the ambient light will appear orange, setting correctly for the ambient light means the flash light will appear blue. To overcome this you can adjust the colour temperature of your flash using filters - I purchased a great selection from FlashGels.co.uk. The picture above shows three types of filter: colour temperature orange (CTO - for cooling the flash light to balance with the light from tungsten lamps), colour temperature blue (CTB - to warm the flash light) and plus green (adds a green cast to balance with fluorescent lighting). These can be attached to the flash unit using velcro, but having watched Blue Peter during my formative years, I'm naturally a dab-hand with cardboard, double-sided sticky tape and scissors (provided an adult is available to supervise me) and so constructed a filter mount.
Autofocus assist beam
Autofocus relies upon the contrast between dark and light areas - this becomes more difficult to assess as light levels drop. My venerable Canon EOS 5 has a built-in AF-assist light which automatically switches on and beams a near infra-red light pattern from an LED to help the AF system. My Canon EOS 5D, however, must rely upon either a Speedlite or the ST-E2 Speedlite transmitter for AF-assist. I imagine the reason for this is that an in-built AF-assist light would struggle to cover the 9-point AF system of the 5D.
When I'm shooting a wedding it's generally around the first dance that I see the reassuring three vertical red lines of the AF-assist beam on my subjects. I can be shooting at f/2.8, or lower, and yet I rarely get a shot that is out of focus. Impressive stuff. Tips for focusing: I only use the more sensitive central focus point, to take advantage of both the f/5.6-sensitive cross-type sensors and f/2.8-sensitive vertical line sensors, and I look for areas of good contrast, such as the line between a white shirt and a dark jacket, to focus on. If you're going for a portrait shot, try and focus on eyelids or eyebrows.
In the 580EX II Speedlite manual it states that the AF-assist beam is compatible with lenses of focal length 28mm and longer, but it seems to work perfectly well at 24mm with my 24-70mm f/2.8 L zoom lens. The effective range of the light is 0.6 - 10 m at the central AF point, and 0.6 - 5 m at the outer points.
When shooting in low light the AF-assist beam fires twice - initially to calculate a focusing distance for the lens and then, secondly, to confirm that the subject is in focus before taking the shot. If the subject is moving too quickly, the focus distance will have altered between these two readings and focus lock cannot be achieved. For this reason the AF-assist beam is not compatible with AI servo - your camera must be in one-shot mode. For the tricky shot of the bride and groom walking down the aisle I have the camera in AI Servo mode only if the church is well-lit. In low-light I switch to one-shot mode to take advantage of the AF-assist beam and take plenty of shots to maximise my chances of getting one that is sharp (I always ask the couple to walk slowly beforehand).
As mentioned in an earlier post I use an Ee-S focusing screen to aid manual focus. This has no impact on the autofocus system but does affect the metering system (this is set with custom function 00). With lenses of aperture f/2.8, or lower, the viewfinder remains bright.
E-TTL II
These are some brief notes on E-TTL II. For an in-depth read on Canon's flash systems read NK Guy's magnum opus at Photonotes.Evaluative through-the-lens II (E-TTL II for short) is the latest incarnation of Canon's automatic flash exposure control system. I currently use a Canon EOS 5D with 580 EX II Speedlite which is based upon this system - it's a match made-in-heaven for flash photography. Previously I used a Canon EOS 5 with 540 EZ Speedlite (based upon A-TTL, where A = advanced) which was far less reliable and required me to constantly adjust flash exposure compensation (FEC).
The system is a 'black box' in that Canon publish only limited information about the algorithms that control flash exposure. Their attitude is switch it on and use it. I would urge you to do the same. Get an empirical feel for how the flash system works and then make adjustments as you wish.
"The output of the fill-in flash depends on the shooting conditions. With lower light levels (below about EV 10), you get a flash output just as if you were shooting a subject at the same distance indoors. Above EV 10, the flash output is gradually reduced, to a maximum of -1.5 stops (-2 stops with E-TTL autoflash) at EV 13 and above. This auto flash output reduction helps to create a better balance between the daylight and the flash illumination in bright sunlight."
When shooting indoors I still tend to use some flash exposure compensation (at least 2/3 stop) to keep the light contribution from the flash very subtle but the shot will still look good without.
With E-TTL II, the camera transmits the lens focal length, exposure control mode, aperture and image sensor size. With more recent lenses focusing distance data is also communicated - even if the Speedlite is attached via an off-camera shoe cord (although for this to work accurately your flash unit must be a similar distance from the subject as if it were mounted directly on the camera - a flash-bracket is fine). Distance data is not used if flash is bounced or used wirelessly.
Metering for flash uses the same system as for ambient light metering and is linked to the current auto-focus point. The camera fires a pre-flash which allows it to meter the scene for flash exposure. To see this, put the flash in second-curtain sync and then set a long exposure time.
Flash exposure locking is available for situations in which the camera is fooled and you're not confident using FEC. These are the same situations that can cause problems metering for exposure - when the overall tone of the shot is not medium-grey.
Find me on Google+