Video: What’s the difference between 8-bit and 10-bit videography?

Videographers, is it worth spending more for a camera with 10-bit video capabilities? What benefits do you gain over using 8-bit? Jakub, our videographer at MPB Brighton, demonstrates the differences between 10-bit and 8-bit video footage using the 10-bit Sony FX3 and 8-bit Sony A7S III.

Is recording video in 10-bit worth the additional cost, storage space and computing power? These days, more and more cameras allow for recording in 10-bit video codecs, compared to the 8-bit codecs that are still popular. For this test, I'm using two Sony cameras, the Sony FX3 and Sony A7 III. First, let's explain the difference between 8-bit and 10-bit video.

What is 8-bit and 10-bit?

As you may already know, the digital image comprises three separate colour channels—red, green and blue. Therefore, the colour model is called RGB. Combined together, they create the final image. With 8-bit colour depth, each colour channel can capture 256 shades of a particular colour, meaning the final 8-bit image can display 16.7 million colours in total. A 10-bit file contains 1,024 levels per channel, resulting in more than a billion colours in total. This increase in colour resolution significantly impacts the dynamic range and colour accuracy of the 10-bit image. Let's move on to our first test.

Testing 8-bit vs 10-bit

For this shot of a garage, I recorded the same shot in 4K 8-bit on a Sony A7III and in 4K 10-bit on a Sony FX3. I shot it with the picture profile turned off, which means the video file doesn't require any colour grading. As you can see, the straight-out-of-camera files look almost identical. Even after applying a strong colour grade, there isn't much of a visual difference between them.

If we apply a really heavy colour grade, we can notice some artefacts appearing on the 8-bit footage, but they are still rather minor. This demonstrates how the difference during color grading can be subtle between 8-bit and 10-bit colors, especially with minimal post-production work.

Fine details and smooth gradients test

Our next test is quite a demanding scenario for any video codec, as there are lots of fine details and smooth gradients in the shot. It is a shot of tree branches moved by the wind with a clear blue sky in the background.

This shot of a tree was recorded in an S-Log 3 picture profile, which gives me the most flexibility in editing—but requires grading before publishing. This LOG image format is particularly useful for capturing a wide range of tonal values.

Now we are starting to see the advantages of 10-bit recording. Look at the discoloured edges appearing on the 8-bit footage. It shows there is just not enough colour data to maintain the smoother transition between the sky and the leaves. The 10-bit video image preserves more subtle nuances in color and provides a smoother gradient in the sky.

Again, let's see what happens when I apply an extreme colour grading to break the codec. Notice how the banding in the sky gets more visible on the 8-bit footage. This banding is a common issue with 8-bit colour systems, especially when pushing the footage in post-production.

This grade also revealed a weird pattern on a shot recorded in the all-intra codec on a Sony FX3. But, we don't film static shots of cars and trees very often. Let's film some people and see how the files perform in a more real-life scenario.

Portrait test

For this one, I filmed Jamil, our social media manager, in front of a blue wall. To make the test even harder for the cameras, I used a Black Pro Mist filter to soften up the image and make the highlights bloom.

After matching the shots, we can see the differences between the codecs. Around the neon sign, on the 8-bit footage, there are severe artifacts, which are not present on the 10-bit one. Also, the noise and banding on the wall are much worse on the 8-bit shot, the footage lacks the colour data required to render smooth tonal transitions on Jamil's skin. This demonstrates how 10-bit colour depth can provide more accurate colour representation, especially in areas with subtle colour variations.

If we have a look at separate colour channels, it reveals how much more colour data there is on the 10-bit footage, especially in the red and blue channels. Notice the compression artefacts around the neon sign and on the blue wall. This comparison clearly shows the advantage of 10-bit colour in preserving colour fidelity across all colour components.

Green screen test

And now the most demanding test for every video codec: green-screen. This time, I filmed myself in front of a green screen, holding this beautiful Fujifilm X100T. It's not the best-lit green screen, which was intentional to make the test more demanding.

The quality of the key is not that bad on the 8-bit shot, but it was definitely easier to achieve a good key on the 10-bit footage. But we can definitely still see the same issues, for example, on my skin, which we saw in the previous test. There are visible blotches and artefacts where a smooth tonal transition was supposed to be. This test highlights how 10-bit colour depth can be particularly beneficial in professional video production scenarios involving chroma keying.

How does bit depth work with chroma subsampling?

Chroma subsampling is a method of encoding images by reducing the amount of color (chroma) data while still retaining the full brightness (luminance) information. It helps to reduce the overall data rate while still retaining a good perceived image quality. Let's have a look at what difference using 4:2:2 or 4:2:0 chroma subsampling makes when paired with 10-bit bit depth footage filmed with a Sony FX3.

The difference is barely noticeable on a full-colour image unless you zoom in and know what to look for. If you look closely, you will notice more colour fidelity in certain areas, like the neon sign and less noise in the background in the 4:2:2 footage. It proves how efficient this method of compression is while still retaining good perceivable image quality.

The difference is apparent when you check separate RGB channels: there are visible artefacts on all the channels on the 4:2:0 footage. In the red channel, you can clearly see the edges of the logo on the T-shirt, clearly showing that there is much more luminance than colour data.

Exactly like with 10-bit and 8-bit, you will notice the most visible difference when keying or doing heavy colour grading (which usually involves keying anyway). There is just a lot more colour fidelity in the 4:2:2 footage, allowing for creating much smoother masks.

Conclusion

Overall, 10-bit requires significantly more computing power and results in larger file sizes. In most cases—especially if you're not filming in Log or flat profile and trying to achieve the effect directly in the camera—you probably won't need 10 bit. However, it is useful if you do a lot of colour grading or digital effects like chroma keying.

Bit depth and chroma subsampling work in tandem. If your camera gives you that choice, use 4:2:2 when filming in 10-bit for maximum image quality and use 8-bit 4:2:0 when planning on doing only light colour grading and no keying is needed. This approach balances colour quality with file size considerations.

Other camera models that shoot 10-bit video include the Panasonic DC-GH5, Fujifilm X-T3 and Fujifilm X-T4, Blackmagic Pocket Cinema Camera 4K and Pocket Cinema Camera 6k Pro EF, Canon Cinema EOS C200 EF and Sony A7S III.

If you shoot for YouTube using a ready-to-publish picture profile, require a fast turnaround and want to save some storage space, then 8-bit is still a good option for you. All of the modern DSLRs or mirrorless cameras can shoot videos in 8-bit. Find out the difference between the Sony FX3 vs A7S III.

What key advantage does 10-bit video have over 8-bit?

10-bit video captures over 1 billion colours per channel, while 8-bit captures around 16 million. This preserves more colour information and provides a wider range of colour possibilities for post-production work.

When will 10-bit make the most noticeable difference?

During extensive colour grading, smooth gradients, and green screens where banding can occur. The higher bit depth allows for more extreme colour grading without introducing visible artefacts.

Is 10-bit needed if minimal colour grading is done?

No, 8-bit is often sufficient if colour grading is minimal and a ready-to-publish profile is used. However, for professional video production or projects requiring extensive post-production, the additional colour precision of 10-bit can be beneficial.