How to Get the Most Out of ETC Mode

Ralph B

Well-known member
As most of you know, ETC 1:1 mode on the GH2 adds a significant amount of noise to the image. So much so, that I really don't care to use it. However, today I made an interesting discovery - the noise is far worse at 24P than it is at 60i. In fact, it's darn good at 60i.

Attached are two frame grabs from ETC mode. They are absolutely identical in all settings, except one is 24P and one is 60i. Both used a shutter of 1/60, ISO 1250, f5.8, NR -2.

As you can see, at 60i you give up a tiny bit of vertical resolution, but you gain a tremendously cleaner picture. The difference is even more dramatic when playing back at speed.

EDIT:
After posting these pictures, I discovered that the forum reduced them to 960 x 540. The files I uploaded were 1920 x 1080.
As such, the difference in noise is not as apparent as it really is. I encourage everyone to run their own tests and see for yourself.

FURTHER EDIT:
Go here to download the full res pictures:

http://img130.imageshack.us/g/etc60i.jpg/
 

Attachments

  • ETC 60i.jpg
    ETC 60i.jpg
    93.1 KB · Views: 0
  • ETC 24p.jpg
    ETC 24p.jpg
    97.8 KB · Views: 0
Last edited:
They are absolutely identical in all settings, except one is 24P and one is 60i.

Sorry. One just seems brighter than the other. Why would you want to compare two different capture rates at the same setting? Surely getting the right picture is more important than equal settings.

Confused.
 
Sorry. One just seems brighter than the other. Why would you want to compare two different capture rates at the same setting? Surely getting the right picture is more important than equal settings.

Confused.

Please see the edit to the original post.
 
I would expect interpolating the vertical resolution back in to decrease the noise. Have you tried the slow motion mode with the ETC yet? That can yield 1080p @ 30 FPS if you speed it back up in post processing. It doesn’t have sound but it is very sharp with smooth motion.
 
I'd love to see the difference on ISO 800, could you post some examples? You could also experiment raising NR bit by bit.

Thanks in advance.
 
I downloaded the full size images. Noise was almost identical. Resolution was not. Why would you trade resolution and bit rate for a tiny bit of noise reduction?

This is always a problem when posting still images as a substitute for video. If you saw these clips playing at speed, the difference in noise is obvious.
That's why I encourage people who have the camera to run their own tests, and decide for themselves whether this is something of importance to them.
 
I can clearly see the difference in noise those shots, but I wonder why there would be a difference between the modes ?
 
I can clearly see the difference in noise those shots, but I wonder why there would be a difference between the modes ?

To me, it looks like the result of dual row summation, which is how an interlaced picture is scanned.
In this case, dual row summation is your friend.
 
I'd love to see the difference on ISO 800, could you post some examples? You could also experiment raising NR bit by bit.

Thanks in advance.

Here you go!

http://img87.imageshack.us/g/etc60iiso800.jpg/

By the way, the best way to compare two images like this is to open them in Photoshop, copy one, and paste it over the other. Then you can use Control Z to toggle between them.

Pietro, regarding the NR, I've already done experiments with it and reached the conclusion that I don't like it. I leave it permanently at -2. The problem is it really doesn't reduce the noise - it only smears it. And it smears out fine detail along with it. Far better to apply more sophisticated noise reduction in post, like Neat Video.
 
Back
Top