How do I use BRAW in FCPX?

The good news is, you now can, no thanks to Apple. If you want to skip all the explanatory stuff and just go for the solution, the answer is near the bottom.

What is RAW?

Codecs

Video codecs are the way video and audio footage is encoded into a computer-readable format. There are lots of ways to do this, and the technology evolves over time.

There’s no one-size-fits-all approach with codecs. For example, some codecs can produce great quality at the expense of vast, unwieldy file sizes (we’re looking at you ProRes 4444).

Some produce much smaller file sizes, but need a lot of processing for the computer to output the image. H.264 is an example of a codec which is fine for final output for something that might be viewed on the web, but isn’t brilliant to edit in - the amount of work involved in rendering multiple streams of H.264 will slow down the NLE.

With this in mind, back in 2007 Apple came up with the ProRes series of codecs, which maintain high quality video whilst being optimised for editing. Many manufacturers licensed ProRes to include in their cameras and recording devices. This type of standardisation is really valuable, as it means you can shoot, edit and output in the same codec without having to transcode.

It’s worth mentioning that MOV and MP4 are not codecs - they’re wrappers. They’re the structure in which the codec data is stored. So you can have a MOV wrapper around an ProRes codec, or an MP4 wrapper around an H.264-encoded video stream, for example.

Lossless vs Lossy Compression

One option with codecs is to store every frame of video pretty much as it came off the sensor. This ensures all quality is preserved, but generates enormous, unwieldy files. These files (or more commonly the frames or sequences within them) can be compressed.

Lossless compression involves applying a reversible mathematical process to data, such that it takes up less space but can be completely restored when needed. However, lossless compression can only shrink files so far.

Lossy compression is a bit of a dark art in video, and involves reformulating and removing data representing the visual image in such a way that you can shrink files very significantly, without making too much of a visual, perceptible difference to the end result. By doing so you can reduce file sizes by a factor of ten or more over the original, uncompressed data. MP3 revolutionised portable music by allowing WAV files to be compressed to a tenth of their original size without having a vast impact on the audio quality, although audiophiles who are happy to spend £800 on an ethernet cable would of course dispute this. H.264 worked similar magic with video, and more recently, computers became powerful enough to handle the stronger compression algorithms of H.265 to shrink everything even further whilst maintaining reasonable perceptual quality.

So when you see ProRes 422 LT, ProRes 422, ProRes 422 HQ and ProRes 4444, what you’re seeing is a range of codecs which gradually trade off file size for quality. ProRes 422 LT generates the smallest file sizes, with the biggest loss of quality. ProRes 4444 maintains excellent quality at the expense of enormous file sizes. Of course, ProRes 422 LT may be perfectly adequate for encoding basic-quality video such as generated by a GoPro, but if you use it to encode footage off a 6K High Dynamic Range sensor, then what you end up with may be perceptually quite a long way adrift from what you started with.

The RAW Advantage

As time has gone on, the demand for quality has increased. If you’re mastering for 4K or 8K TV output, or working in High Dynamic Range, where you need to capture a great deal more colour information than Standard Dynamic Range, the traditional codecs aren’t going to be your friends. You want to maximise quality, but you’ll also probably have the budget to pay for somewhat more sizeable storage and powerful editing machines to cope with this.

All the “traditional” codecs involve interpreting the data coming off the camera sensor at a basic level. In particular, sensor data needs to debayered before it can be viewed on an RGB screen; rather than have equal numbers of red, green and blue pixels, camera sensors invariably have a pair of green pixels for every red and blue pixel (the human eye is more sensitive to green light, so this makes for a better image). Debayering or de-mosaicing converts the data from these four sensor pixels into a single RGB data pixel, but this process is information-destructive and non-reversible.

At the same time, raw sensor data has a white balance baked into it by the codec, and other potentially useful chroma and luma data is lost in the encoding process. Many codecs use a lower bit depth than the sensor data, so they’re converting, say, 1024 levels of light for each of red, green and blue into 256 levels of light for each. This may not be visually noticeable, but once you come to tweak and fiddle with the image, you can’t get that lost colour data back. Chroma subsampling involves throwing away some chroma data to reduce the volume of data being processed, which is all well and good until you need that data at some point further down the line.

RAW codecs, in the strictest sense, preserve as much sensor data as they can, and leave the debayering and other processing until render time. This means you that when you want to process the image - altering exposure, colour balance, or revealing details in shadows - you can take full advantage of every bit of colour and luminosity data the sensor captured.

You’d think RAW codecs would use lossless compression, but in fact both ProRes RAW and BlackMagic RAW (BRAW) allow the user to trade off file size against fidelity. It’s fair to say however that the compression in this case is optimised towards retaining as much of the hidden information that you’re likely to need in post-processing as possible, rather than what’s needed for immediate visual rendering fidelity; for example, Blackmagic RAW does perform partial debayering during encoding, although ProRes RAW only debayers at render time. Both these codecs can use a target bit rate: that is to say, rather than allocate a fixed amount of data for each frame, they use adaptive bitrates to aim for a fixed visual standard of encoding quality, regardless of the complexity of the scene.

RAW codecs were originally developed by camera manufacturers, notably RED, but as users’ technical demands increased, Apple developed their own ProRes RAW format in 2018 in conjunction with monitor/recorder manufacturer Atomos, and Blackmagic released Blackmagic RAW (BRAW) in 2019.

Comparison of bit rates for different codecs, at 4K/50p. The manufacturers don't publish much useful data about bit rates, so we had to work backwards from various publicly published information. Consequently, E&OE.

Why You Don’t Want To Transcode Too Much

It used to be a lot of hassle dealing with lots of different codecs, but typically software such as Final Cut Pro deals with most of this behind the scenes. If you import an H.264 video into Final Cut Pro, for instance, FCP gives you the option of either keeping the H.264 file intact, or transcoding it to ProRes (“optimising” the file). Optimising will give you better playback performance, at the expense of creating more data. Additionally, an optimised ProRes file might take up a whole lot more storage space than the H.264 file it was transcoded from.

Because each codec has different ways of compressing data, when you transcode from one to the other, you’re almost inevitably going to be losing some picture data. This is not normally a problem, but if you have to transcode several times, you may experience noticeable degradation in the image quality, or compression artefacts.

As an example, let’s say you shoot using a mirrorless camera using a proprietary codec; you then transcode into ProRes for editing. Then when you output, it’s for YouTube so you transcode into H.264. (YouTube will ingest ProRes happily, but the file size will be much larger to upload.) Not only are you risking losing visible image quality and colour data which was present in the original camera file, but you’re also dealing with more storage space than you really needed to.

Consequently there’s a lot to be said for shooting and editing in the same format, even if the final delivery format is different.

Why is dealing with RAW so complicated?

There are two issues you’ll come across when dealing with RAW. One is compatibility when recording, and the other is compatibility when editing. Both aren’t technical hurdles; rather, they’re a result of commercial organisations who choose not to play nicely together, to the inconvenience of their paying customers.

RED have the RAW recording market in somewhat of a stranglehold, thanks to their patents. If you want to put RAW recording capabilities into a camera, you’re at the mercy of RED letting you, unless your engineers and legal department can develop some sort of technical loophole which gets you around this. If you’ve wondered why some high-end cameras only allow RAW recording using an external recorder, here’s your answer.

RED don’t care so much about external recorders, so they’re happier to license RAW recording technologies to companies such as Atomos, who don’t make cameras. But RED’s lack of flexibility in allowing other manufacturers to include RAW recording in cameras has really not done the whole market any favours. Take for example the DJI Ronin 4D, which was announced with ProRes RAW support, which was subsequently dropped, along with the price to compensate, upon release. (A firmware glitch meant French users could indeed use ProRes RAW for a time, which shows it was a commercial rather than technical hurdle which led to this.)

The second issue you’ll encounter is Blackmagic RAW vs ProRes RAW. Both companies are keen to establish their technology as the market leader. Both companies have their own NLE (Da Vinci Resolve, and Final Cut Pro X.) So, guess what? Blackmagic don’t support ProRes RAW in Da Vinci Resolve, and Apple don’t support Blackmagic RAW in Final Cut Pro.

In a further twist of irony, Blackmagic’s cameras will happily record in ProRes, even the higher quality variants, and Resolve will cheerfully edit ProRes footage. Just not ProRes RAW, because Blackmagic want you to use Blackmagic RAW instead.

Blackmagic’s RAW format is available to other camera manufacturers to use themselves for internal RAW recording, but none have yet taken up the option. This said, Blackmagic do have a lot of penetration at the low- to mid-range video camera market, and if you want an inexpensive camera with the potential to deliver great visual results, Blackmagic is likely to be high on your shopping list.

ProRes RAW is also available to other manufacturers, but products using it are niche; the only mainstream hardware which encodes ProRes RAW comes from their launch partner Atomos. So you can record in ProRes RAW from any camera, providing it outputs the correct unprocessed signal over SDI and you don’t mind using an external recorder, along with its additional weight, power supply etc.

Adobe don’t have their own RAW format to push, so Adobe and Blackmagic are happy for Premiere Pro to support BRAW. So if you’re the masochistic type who enjoys working with unnecessarily expensive, slow, clunky software with a poor user interface, then you can always use PPro.

So for anybody who wants to shoot RAW with a mid-range camera - which includes anybody working with HDR workflows, or working at 4K or higher resolutions - you’re pretty much looking at shooting with Blackmagic and editing with Resolve or Premiere. This suits Blackmagic down to the ground, but doesn’t help the many professional production companies and editors who use Final Cut Pro X, and don’t have the time or inclination to use a more traditional, less agile NLE such as Resolve or Premiere Pro.

Blackmagic RAW in Final Cut Pro X

So if you’re one a pretty large community of professionals who end up wanting to shoot in Blackmagic RAW and edit in Final Cut Pro, you have two options.

One is to transcode from Blackmagic RAW to ProRes. (Not ProRes RAW; it’s a recording format, and as such it’s expecting sensor data as an input, not a file.) Color Finale Transcoder and EditReady will allow you to do this, as well as dealing with other niche formats that FCPX doesn’t handle. But there’s a catch. Because you’re then editing the transcoded footage, not the RAW footage, you’ve just lost a lot of your flexibility in grading and post-production, not to mention using about double the storage space because you now have a BRAW and a ProRes version of each file.

The second option is a tool from a small Australian production house, LateNite Films, called BRAW Toolbox. Chris Hocking and his small team have managed to come up with creative ways around Apple’s very haphazard support for third-party integrations, allowing you to import BRAW into FCPX, without transcoding.

To import your BRAW clips, you go through the BRAW Toolbox, as FCPX won’t recognise them natively. You can pre-process all the parameters you might want to: exposure, gamma, colour temperature, LUTs etc. Processing is very quick, because you’re not generating new video data, you’re simply updating the metadata the playback engine is using to render the RAW video.

Once processed, the BRAW clips can be dragged straight into FCPX. As far as Final Cut is concerned, it’s just playing out clips like any other. But behind the scenes, BRAW Toolbox is doing something quite clever. Final Cut thinks it’s playing out a simple colour generator for each clip, but BRAW Toolbox uses an Effect on each of these clips to hijack the playout and insert its own playout of the BRAW file, using the Blackmagic official libraries, totally seamlessly.

In use you wouldn’t particularly notice anything different from editing any other FCPX media, although you now have full access at any time to the BRAW Toolbox processing tools. All parameters can be changed dynamically, and even keyframed. Taking exposure as an example, if exposure changes mid-shot, you might be used to using the Colour Board to change shadows, midtones and highlights (aka lift, gamma, gain). With BRAW Toolbox, you can directly change (and keyframe) the exposure interpretation of the original sensor data, giving you access to every last bit of colour data the sensor captured to maintain the quality of your processed image.

So Chris and his small team have managed to achieve something that the world’s largest computer company apparently isn’t sufficiently motivated to do; allow you to edit RAW footage from the market leader in mid-range RAW cameras, using what’s supposed to be a state-of-the-art NLE.

Of course, in an ideal world, Apple and Blackmagic would get together round a table and license each others’ technologies, rather than make their customers’ lives harder by being anticompetitive. Don’t hold your breath.

On a more positive note, Chris Hocking also develops the excellent CommandPost for Final Cut Pro X, which adds in all those features that you wish Apple would, such as tight integration with a wide range of control surfaces. We can only hope Apple hires Chris to be the FCPX Product Manager, as he seems to be doing far more to keep FCPX relevant than Apple's development team.