Why is 60/50p not better than 24/25p in the RX100 videos?

Started 4 months ago | Discussions thread
ForumParentFirstPreviousNextNext unread
Flat view
Interceptor121
Regular MemberPosts: 360Gear list
Like?
Why is 60/50p not better than 24/25p in the RX100 videos?
4 months ago

I have in the past run some tests between 24/25/50/60p modes to determine which one was giving me the highest quality and most fluid. I watch my clips streaming from my AppleTv to a HDTV.

When I had the RX100 I started off with 50p (I am in a PAL area) I considered that for compatibility reason it was better to use this and considering the bitrate was the same with 60p I thought I would get a higher quality, however when I did some tests I could not tell any difference. I would then produce two clips one at 50p and one at 25p for watching on computer and appletv.

I would go very excited to the tv with the laptop to show how more smooth the 50p footage was however people in the audience kept telling me they saw no difference.

I then started looking at AVCHD specifications and how Sony was implementing those.

This is an interesting read if you have no knowledge http://en.wikipedia.org/wiki/Group_of_pictures

I then analyzed the clips produced by the camera in the 60PS and 24FX formats using mediainfo.

I found out that both had two reference frames per second and that the video bitrate was 22 Mbps for the FX format and 26 Mbps for the PS the other 2 Mbps were for the audio.

At the end of the mediainfo stream analysis I found the following field Bits/(Pixel*Frame)and noticed that the value for 24FX was higher than 28PS more specifically 0.407 vs 0.198. I started digging more into it.

When I looked at the GOP structure I found out that the 60PS footage has the following

M=1 N=30 that means the group of pictures is 30 frames long and there is one I frame in each GOP

The 24FX footage had instead

M=1 N=12

Which means the group of pictures is 12 and there is one full frame at the beginning.

So in one second both formats have two full frames and the rest are predicted frames.

A full frame is like a JPEG and a predicted frame is a forecast of the future frame based on the past frames already recorded.

So in one second of footage there are the same number of full frames for both and the rest are prediction so the 60P format has only more motion compensated frames but actually no full information compared to the 24FX format.

It is considered that a P frame is half or less of a full frame so with this in mind is like saying that in one second of 60P there are two full frames and 58 P frames or normalised to 2+29=31 frames

In one second of 24p there are two full frames and 22 P frames or normalised 2+11=23 frames

Dividing this for the bitrate I have 26/31=0.84 mbits/frame and 22/23=0.95 mbits/frame

This shows that on average a single frame of 24FX file contains more information than a frame of 28PS file. As the additional frames are all predictions this also explains why there is actually no real benefit in the 60p format compared to 24p as the only additional frames are prediction and not real data.

Some users report however that their 60p is more fluid than 24p if they watch this on a Tv, why?

This is because TVs have issues with 24p as there is a process called telecine to make sure that 24 frames fit into a normal 60Hz system and also because the shutter speed selected by the camera of 1/125 does not fit round with the frame rate so the footage looks 'cinematic'.

A corresponding PAL user where the frame rate are exact multiple of the TV refresh rate can't tell the difference between a 25pFX and 50PS file on a normal TV set because actually there is none.

In fact in scenes of high dynamic range the 24/25p format performs better because of the additional data per frame.

So where did all of this come from? Essentially marketing, in order to have a visible improvement in quality you need higher bitrates, so a difference between 22 and 26 Mbps is not sufficient to increase the number of reference frames and therefore this is all a bit of a marketing hype.

Recently EOSHD have done some test on the Sony A6000 that shares the same codec of the RX100 and compared the two modes 24/60 and concluded AVCHD 24FX has more detail, this is the reason why.

There are still benefits to shoot double frame rate if you plan to slow down the footage to half speed as the frames interpolated by the camera are better than how a normal non linear editor would process the same information however other than this there is pretty much no benefit to shoot the RX100 in what the manual defines 'highest quality setting'.

I see many posts of users here that even want to shoot 60p in a PAL system thinking that the higher frame rate will help but that is not the case. The most fluid format coming out the RX100 is 25p at 1/50th shutter speed this has no rolling shutter issue that the 24p has and is fluid along the scale.

Hope you guys find this useful and save yourself the learning process I had to undertake (let alone a bit of disk space and having to buy new programs to deal with AVCHD 2.0)

-- hide signature --
 Interceptor121's gear list:Interceptor121's gear list
Sony RX100 II Nikon D7100 Sigma 70-200mm F2.8 EX DG Macro HSM II Sigma 17-70mm F2.8-4 DC Macro OS HSM Sigma 2x EX DG Tele Converter +1 more
Sony a6000 Sony RX100
If you believe there are incorrect tags, please send us this post using our feedback form.
ForumParentFirstPreviousNextNext unread
Flat view
Post (hide subjects)Posted by
ForumParentFirstPreviousNextNext unread
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow