Although it’s true that HDSLRs have opened up the world of cinematic-quality film making to the masses, with jaw-dropping image quality and amazing shallow depth of field, film making is not about the technology. It’s true that technology can help you realise your vision, but it should rarely be the main event. If the script or content is compelling enough, the camera used to shoot it should carry less importance.
The chances are that your audience or client will not notice how fantastic it looks when you get it right, but they will almost certainly know when you get it wrong.
It may be possible to use EQ and dynamics to improve average sound, but trying to use any type of filter to remove unwanted or improve bad sound in post will never work well, if at all, without involving large amounts of money and time. The most common problem new film makers suffer from is wind noise, which there is no way of removing in post because it covers too many frequencies and therefore cannot be EQd without removing all of the mids and lows, thus leaving any dialogue sounding thin and obviously altered. A professional location sound recordist will carry a variety of wind remediation equipment with them, as the only effective way to deal with the problem is at source.
We are used to things sounding a certain way. When they don’t we can easily tell. Bad sound will always be bad sound – get it right in the beginning.
This depends on the scene being shot. Your huge-sensor DSLR, with its ultra-bright lens, gives a great look and feel for a lot of shots as a style choice, but you need to handle people with care – especially in interview situations.
When shooting moving images the problem is exactly that – they are moving. Even in a set-piece, sit-down interview scenario, the subject will move. Although this may only be a few inches in any one direction, this is enough movement for the subject to be out of focus momentarily. You must therefore always ensure that there is a pool of focus about six inches in front of your subject’s nose for them to move into, and the same at the back of their head. When it comes to a ‘two shot’ it’s even more important to get both subjects evenly focused – unless of course you are pulling focus or working soft for effect.
Most of the time, after a period of time practicing as an experienced photographer, you will be able to expose and focus your camera better than your camera can do it for itself. As far as focus goes, autofocus will seek a focal point while shooting moving images. This makes the video shot unusable. The same goes for exposure. Most DSLRs respond quickly to changes in light conditions; great for when you are capturing a moment in time – every still is exposed perfectly – but the change needs to be more gradual for moving images.
You need to be able to make decisions that the camera can’t and decide what your priorities are for exposure, focus and white balance.
With a growing appetite for more pixels, most of us crop our stills to improve composition; after all, we have resolution to spare. For example, when it comes to good-quality magazine print, the optimum resolution would need to be no higher than 350 pixels per inch (PPI) at reproduction (standard) size (SS).
When shooting HD moving images you are always shooting at the optimum resolution for HD output and, despite what all those hi-tech police dramas would lead us to believe, you cannot enlarge an image in your editing application without encountering huge resolution loss.
Professional photographers and film makers get the composition of the shot right from the start. If you are shooting HD then it is possible to use some of your spare resolution if your delivery is SD, but you must then work on an SD timeline or the resolution loss will be downsampled again when you encode the SD files.
For the most part this is true; more pixels mean a more detailed image… in theory. Unfortunately there are other considerations, such as colour space, GOP structure (a group of successive pictures within a coded video stream) and compression artefacts. A detailed explanation of these is a whole other feature, but it can be compared in some ways to shooting stills RAW or JPEG.
Most DSLRs shoot to file types designed to be an output format. They are highly efficient, generally high-quality codecs, but still a highly compressed file that will generally need to be compressed again as you output it from your editing application, depending on your delivery method. You then suffer even more loss of colour and detail.
This is a problem with most HD-dedicated video acquisition formats to a lesser – and sometimes greater – effect. The saving grace is that we may not notice the changes made by the codec; after all, it is designed to change the areas of colour and image that the human eye is not used to recognising and is unlikely to notice.
Okay, so a better mic will almost definitely capture better sound than the camera’s built-in mic and an on-camera mic can be an excellent solution for photo journalists and those working on their own on location. But you have to be aware that as long as it’s mounted to the camera it will also pick up handling noise – even with a cradle/shock mount. Also, to capture good dialogue, the mic really needs to be closer to the subject than you probably want the camera to be. This can be done in a few ways, but you already know the number one rule; closer is better (within reason). Obviously, you don’t want to see the mic in shot, so there is a trade-off.
There is no ‘one size fits all’ microphone; although, with the right choice, you can improve the quality of recording by a huge margin. As a rule I think it’s fair to say that you would have to try hard to buy a mic that sounds worse than the onboard mic of a DSLR. Microphones are important!
Traditionally, high film speeds for stills have a definite look and feel which in some circumstances can add to the appeal of the image. If that look is less than desirable then you can use flash or a slower shutter speed, depending on the circumstances of the shoot.
When shooting video, the gain introduced by using high film speeds is not only ugly, but can create havoc when compressed for distribution. The noise is not ordered or stable. It shows as erratic, distracting noise; mainly in shadow areas that will spoil the viewer’s experience.
With this much noise randomly placed on every frame, the compression methods for delivery which usually rely on only changing the parts of the image that move or need to change between key frames (iFrames) will need to abbreviate the movement of the background and the image can become blocky with heavy artefacts. Unfortunately, the only way to deal with this problem is to keep your film speed down and use more light or a brighter lens. Failing that, you are stuck with some gain, but at least if you are well informed you can make an educated decision. Sometimes needs must!
Most lighting camera people would be horrified at the very thought of correcting any element of exposure, or even colour, in post. The shot should be lit and exposed perfectly on set, or as well as possible on location.
That said, we all grade our video to a lesser or greater degree, but never go too far with any correction and always shoot as if the footage is going to be used unchanged.
Careful handling is also needed before you get your project into the edit. You need to ensure any grading you do isn’t causing more loss. This means re-encoding the DSLR files into a format with a better colour space and without a long GOP structure. ProRes 422 HQ is ideal, but prepare yourself for large files.
When it comes to filters, most of us think of colour correction for tungsten or fluorescent light, or coloured grads, star filters, nets etc – all things we can do with white balance or in post. We don’t often think about the humble neutral density (ND) filter. Whether it’s a grad or a square ND, they all have new life while shooting video.
For video, the shutter speed should be left, as often as possible, at 1/50sec for PAL and 1/60sec for NTSC. This is slow enough to give the film a small amount of inter-frame blur, thus making the motion appear relatively smooth, but not so much as to make the motion appear smeared; much faster than this and the video starts to become slightly jittery, giving an almost aggressive feel. This may even be a look you like, but certainly something to be aware of.
So, to avoid high shutter speeds while shooting at large apertures, you will sometimes need to use ND filters (most pro video cameras have several levels of ND built in). Also, an ND grad may come in handy to hold detail in large areas of highlight while exposing a flesh tone, for example; though this is not an exact art and sometimes doesn’t work depending on composition. The times it does work, however, make it worth the effort.