The days of video tape are long gone and all consumer video is now fully digital. This has allowed video to become much higher quality and also infinitely more complicated. There are so very many formats and standards of digital video that the correct answer about one type may be very wrong for another type and the top of the range yesterday may be very poor in comparison to today’s latest version.
Luckily, when it comes to drones and the video they shoot, in the consumer marketplace there are only a few types of video at the moment. However, even in those few types, there are different values and structures that affect the quality and use of the video.
Warning! It gets rather technical from here, but if you stick with me you should get a much better understanding of what you’re working with and hopefully you’ll be able to make informed decisions on what you want to work with.
A video file is classified with certain variables: Codec(the compression system) – eg H.264, Prores, DNxHD etc, Wrapper (the file type identifier) – eg .MP4, .MOV, .AVI etc, Bit rateor Data rate(how much data is uses per unit of time) – eg 50 Mbps etc, Bit Depth(number of colours available) – eg 8bit or 10bit, Colour sampling(how the colour is sampled in each pixel) – eg 4:2:0, 4:2:2 or 4:4:4 and Resolution(the size of the picture) – eg HD, 4K, SD etc.
The codec is the specific encoding system used to create the video file. With consumer drones it is usually a type of compression too. Most commonly the system is H.264, which is a really good system for squashing video to play back on almost any device currently available. Nearly all web video uses H.264, as does Blu Ray and most consumer video cameras too. The latest, higher end consumer drones have started using H.265 too. This is a much more advanced compression system but is currently not readable on all devices. There are many other codecs, mostly used in professional applications, but you will see more and more H.265 to supplement the massively popular H.264.
Does exactly what it says on the tin! The wrapper dresses the file up in a simple format that is usually readable across certain devices. AVI is the Windows standard, where MOV is the Apple standard for video files. MP4 is generic and common across the internet especially. A wrapper can contain all manner of codecs (though only one at a time, of course).
Bit Rate/Data Rate
This is where things start to get more complex. This denotes how much data per unit of time is contained in the video file. Generally this means that the higher the bit rate, the better the quality of the video (due to the increased amount of data). Usually this is expressed as Mega bits per second (Mbps or Mb/s). Anything above 50Mbps could be entering the realms of broadcast quality (depending on resolution and some other factors). The latest drones are shooting 4K video at 100Mbps, but high-end Hollywood is often shooting at over 1Gbps!
Since the arrival of the DJI Mavic 2 Pro, there has been a lot of discussion around the value of 10bit versus 8bit video (and what the hell that even means!). Bit depth refers to how much colour information is captured by a camera and thus exists in the video file. Video colour exists in 3 channels (Red, Green and Blue). Each channel in digital video carries a certain number of shades. In 8bit video it is roughly 16 million different shades in total, in 10bit video it is roughly 1 billion shades. Most drones shoot 8bit colour, but the Mavic 2 Pro shoots 10bit. Those extra colours make all the difference between broadcast and non-broadcast quality. If you are not shooting broadcast video or footage for display on a massive screen, you are unlikely to notice much difference.
Colour sampling is how colour is divided between the pixels in digital video. Only video scientists really need to know the details as it’s the sort of thing that ties my brain in knots, but what is important is that there are 3 versions, which are expressed as: 4:2:0, 4:2:2 and 4:4:4. The more 4s the better. Most consumer level video is 4:2:0, TV productions are mostly 4:2:2 and high-end productions tend to use 4:4:4. Similar to bit depths, the higher the numbers, the more colours there are to work with and the more you can do in post-production. For broadcast needs, colour sampling must be at least 4:2:2 for 8bit video, but can be 4:2:0 if it’s 10bit video. All current consumer level drones only shoot 4:2:0.
This is purely the picture size, in terms of pixels. Most common is HD, which is 1920 x 1080 pixels (a lower resolution type of HD is 1280 x 720 pixels). Increasingly, new drones have cameras that shoot in 4K resolution, which is 3840 x 2160 pixels. Whilst the resolution affects the initial clarity of the images, it’s really the other elements that make up the actual video file which determine quality and how easily or not you can actually work with the video.
This refers to the speed with which the camera takes pictures (frames), expressed a Frames Per Second (fps). There is a little bit of history here, but not overly complex, so bare with me. When motion pictures started and right up until the advent of video tape, film had to be run through a camera at some speed, exposing each individual frame of celluloid, to make the picture move. The standard speed for film was/is 24fps. This rate makes movement look natural to the human eye, so has remained popular. When TVs came along things got more complicated, due to the different alternating current speeds around the world. Countries with 50Hz electricity adopted 25fps (Europe, Africa, Australasia and large parts of Asia) as their standard for video and those with 60Hz (Americas and parts of Asia) adopted 30fps (more accurately it’s 29.97fps, but usually referred to as 30fps). This was to stop the screens flickering as they refreshed. Video takes after TV, so those frame rates have stuck. More recently, new standards have started to become available in many cameras, with 50fps and 60fps providing an even smoother result. Not everyone loves the faster rates, but they are increasing in popularity. There are also faster and slower frame rates, which are used for specific purposes (slo-mo or time-lapse etc). It is worth exploring what sort of a screen will show your final footage, as the refresh rate of that screen will have some sort of effect on the footage, even if only very slight. Most computer monitors operate at 60Hz, no matter where they are in the world, whereas most TV screens use their country’s standard rate. As far as image quality is concerned, the frame rate will affect the amount of light getting to the sensor, but makes no other material difference.
With all these constituent variables in every digital video file, it’s easy to understand why there is so much confusion.
Why it all matters
Quality and perceived quality – that’s what this is all about. Different uses of video have differing quality requirements, but there are two main levels: Broadcast and non-broadcast. In the vast majority of cases, this is not remotely relevant and so the differences between a Mavic Air, Autel Evo and a Mavic 2 Pro become a matter of personal choice and we all know that beauty is about as subjective as it gets! However, for users who intend their footage to go to broadcast, or hope that one day it will (especially indie film makers here), then there are set standards which must be met.
I have explained most of the factors which go to making footage broadcast acceptable, but there are a few others that are to do with the camera itself and how it is set that are critical, in addition to the basic video settings.
Cameras all have differing sizes of sensor. The sensor is the bit that actually captures the image, the digital equivalent of a frame of celluloid film. In the digital world, the larger the sensor, the easier it is for it to make the image and the clearer the image will be with less available light. That means that the larger the sensor, the wider the range of lighting situations in which the camera can be used and still get a clear image. For broadcast purposes, the image needs to be as clear as possible, with as little grain as possible (grain sneaks in at lower light levels and starts to make the picture unusable when it is being compressed for broadcast). Further to that, larger sensors allow for more aperture settings and an all-round more pleasing aesthetic to the video. As broadcasters, more than any other thing, want a video that is properly exposed and shot, they require a minimum standard of camera to achieve it. For different broadcast standards (eg UHD/4K, 8K and HD), there are different camera requirements for different percentages of a programme shot on any type of camera, but the most important factor is the size of the camera sensor. (https://tech.ebu.ch/publications/r118).
In the case of consumer drones, only the Mavic 2 Pro meets the full requirements for shooting Full HD television (whilst it has 10bit 4K at 100Mbps on a 1” sensor, it does not reach the full 4K standard, but is close enough to provide a small percentage of a UHD programme). Not even the Phantom 4 Pro V2 is qualified for Full HD, as it lacks the bit depth of the M2P (M2P is allowed to shoot with 4:2:0 colour sampling as it has a 10bit camera).
That really is it! If you want to do a lot of colour correction and grading work, or effects, then 10bit will make a difference, as will sensor size and bit rate, but how the camera is flown will be the really key factor!
In most cases these relative quality levels are likely to be irrelevant, as the video will still look amazing on YouTube, but if you intend to use it for other things, where the technical quality is as important as the visual quality, you would be well advised to remember this post.