XAVC S vs AVCHD vs MP4 – which to use?

The Sony A7s records to three codecs:

  • XAVC S – this is the consumer version of XAVC. Maximum data rate is 50 Mbps
  • AVCHD  – this is also a consumer codec, maximum data rate is 28 Mbps at 50/60p, 24 Mbps at 24/25/30p
  • MP4 – 1440×1080 (1.33 pixel aspect ratio corrects to 1920×1080) 12 Mbps

Here are still frames of all the codecs (click to enlarge). Order:

  • XAVC S 25p 50 Mbps
  • AVCHD 50p 28 Mbps
  • AVCHD 25p 24 Mbps
  • MP4 1440 25p 12 Mbps

Cine1 Gamma Cinema Color Space

XAVC25P

AVCHD50P

AVCHD25P

MP41440

What do you see?

S-Log2 Gamma S-Gamut Color Space

XAVCSLOG2

AVCHD50PSLOG2

AVCHD25PSLOG2

MP4SLOG2

What do you see?

What I think

I would avoid MP4 at all costs. Not only is the pixel aspect ratio a headache, but you also have very poor data rates for any kind of professional work. For the same reason, I don’t recommend dual recording either.

AVCHD looks similar to XAVC S, and if you just want to upload the video straight away it might be okay. There are three disadvantages to AVCHD:

  • Poor audio codec
  • Harder to edit
  • Doesn’t grade very well

There are two ‘advantages’:

  • It’s half the data rate, therefore half the storage.
  • You don’t need an SDXC card for it, you can use SDHC cards (32GB and lower).

You can see quite clearly how the colors change in S-Log2 mode with AVCHD.

To my mind, quite clearly, XAVC S is the only codec I would use. A 1 TB drive can hold 45 hours of footage! Done deal.

Click on the link below to the next lesson or head over to the main menu (above). If you need help with something, feel free to send me an email. I’ll try to get back to you as soon as possible.

FAQs

These are important topics raised by subscribers that shed more light on this lesson.

Q. Why do you say MP4 and AVCHD are bad? Don’t many people use them?

As far is MP4, is concerned, there is a loss in resolution, not only due to the stretch, but also due to the heavier compression.

I’ll try to explain the difference. MP4, H.264, AVCHD and XAVC are all of the same compression family (If you’d like to learn more: http://wolfcrow.com/blog/understanding-mpeg-2-mpeg-4-h-264-avchd-and-h-265/). In short, they are all interframe codecs. Secondly, they are all ‘delivery codecs’ – which means, they were created for final display, and not for in-camera acquisition. They are ‘consumer-grade’.

‘Proper’ video cameras use XAVC I (The A7s is XAVC S, the consumer version), Prores, XDCAM, Canon MXF, etc. Acquisition demands a far more ‘robust’ codec, one that can withstand not only the rigors of post production but also the heavy recompression of broadcast. Even in the case of Youtube, you must first render your timeline to H.264 to upload to Youtube. For 1080p, this is a data rate of about 8-10 Mbps. Youtube will further compress it one more time for playback.

The point is, even when you shoot with consumer codecs, it has to undergo at least 2 compression stages to final viewing, and these codecs were never designed for that. You will see heavy banding, compression artifacts, loss of color and tonality, loss of resolution, etc., if you look close enough. It’s there even with XAVC S, or even if you shot on the Alexa or Red Epic – but it’ll be far worse if you shot on AVCHD, and even more so on MP4. This is also why Vimeo allows you to download videos. This guide has fully downloadable videos so you can study the footage after only one stage of compression, rather than two that you see on playback.

It is paramount to use the maximum data rate when shooting with a consumer codec. So if you choose to shoot AVCHD or MP4, you must be convinced the generation loss on subsequent recompression(s) does not affect your work.