Sampling grain at DVinfo.net

Oscar Spierenburg · May 2nd, 2005, 05:26 PM

Although I have a great working rotating ground glass, I came up with an idea which I'd like to test.
Maybe I don't understand grain enough or other people have tried this...
How about sampling the grain of an ordinary static GG, by shooting a white board or something, so you have only the grain strukture of the GG. If you put that sample in negative on top of other footage (in post software) and use it to lighten/darken, ie reduce the grain, maybe even compensate a hotspot. Any thoughts? Maybe grain deforms the image too, so it wouldn't work.

Leo Mandy · May 2nd, 2005, 08:33 PM

I don't think you would have to even shoot the grain on a white board background, I know that Vegas has a Grain FX installed with the NLE that simulates grain for 8, 16 and 35mm colour and B&W film. IS that what you are talking about?

Daves Spi · May 3rd, 2005, 02:09 AM

Very interesting idea... I already though about something like this, especially for scratches removing. As screen is moving, scratches stays static. So its not hard to find them. After you have their positions, you can replace this pixels from pixels aroud by moving. Some very clever logo removers work this way, so you will do not believe there was logo before, if you have a enough movement to determine area under logo. I do not know if somenthing like this exists...

Quote:

Originally Posted by Oscar Spier

Although I have a great working rotating ground glass, I came up with an idea which I'd like to test.
Maybe I don't understand grain enough or other people have tried this...
How about sampling the grain of an ordinary static GG, by shooting a white board or something, so you have only the grain strukture of the GG. If you put that sample in negative on top of other footage (in post software) and use it to lighten/darken, ie reduce the grain, maybe even compensate a hotspot. Any thoughts? Maybe grain deforms the image too, so it wouldn't work.

Radek Svoboda · May 3rd, 2005, 04:56 AM

This is problem with sampling grain:

The grainy screen consists of tiny dark spots of varying darkness. You would have to make screen as dark as darkest spots, to make it uniform. You'd lose too much dynamic range / lattitude, low light capability. The moving screen will average dark and bright spots. It is lot more efficient, especially because there is low percewntage of dark spots.

Daves Spi · May 3rd, 2005, 05:04 AM

I do not know... I think that if you take just mask of the scene (I mean the grain as mask), you are able to get it out. You can do it in several ways. 1) compute dark spots to be brighter, 2) compute bright spots to be darker or 3) make average between. The different is not so big, so I guess u do not lose to much dynamic.

Quote:

Originally Posted by Radek Svoboda

This is problem with sampling grain:

The grainy screen consists of tiny dark spots of varying darkness. You would have to make screen as dark as darkest spots, to make it uniform. You'd lose too much dynamic range / lattitude, low light capability. The moving screen will average dark and bright spots. It is lot more efficient, especially because there is low percewntage of dark spots.

Oscar Spierenburg · May 3rd, 2005, 05:30 AM

I think Daves got my point, but maybe I was not very clear.
This is what I mean:
Make a 35mm adapter with a static GG.
Sample that static grain.
Put that grain footage (just one frame, because the grain is static) in a post prog. in negative on top of other footage and mess around with opacity and layer type until you get the grain to disappear.

So I'm not talking about moving GG or sampling the grain to put into other footage. Like Daves said, it would also reduce scratches. If it works.

Aaron Shaw · May 3rd, 2005, 08:47 AM

Hey Oscar :)

I started a similar thread a while back. Not sure where it is now. Anyway, it was tested and seemed to work out ok. Some people had trouble reproducing good results though. You need to make sure that the adapter is always oriented perfectly or when you try to remove the grain you will just add more.

Oscar Spierenburg · May 3rd, 2005, 01:03 PM

Aaron, that's what I figured. The adapter has to stay on the camcorder and don't move at all or the grain will be displaced and you have to start again.
I'll test this anyhow some time next week. I'll try to find your thread.

May 2nd, 2005, 05:26 PM	#1
Oscar Spierenburg Major Player Join Date: Jan 2005 Location: (The Netherlands - Belgium) Posts: 735	Sampling grain Although I have a great working rotating ground glass, I came up with an idea which I'd like to test. Maybe I don't understand grain enough or other people have tried this... How about sampling the grain of an ordinary static GG, by shooting a white board or something, so you have only the grain strukture of the GG. If you put that sample in negative on top of other footage (in post software) and use it to lighten/darken, ie reduce the grain, maybe even compensate a hotspot. Any thoughts? Maybe grain deforms the image too, so it wouldn't work.

May 2nd, 2005, 08:33 PM	#2
Leo Mandy Major Player Join Date: Dec 2004 Posts: 749	I don't think you would have to even shoot the grain on a white board background, I know that Vegas has a Grain FX installed with the NLE that simulates grain for 8, 16 and 35mm colour and B&W film. IS that what you are talking about?

May 3rd, 2005, 04:56 AM	#4
Radek Svoboda Major Player Join Date: Mar 2005 Location: Prague, Czech Republic Posts: 500	This is problem with sampling grain: The grainy screen consists of tiny dark spots of varying darkness. You would have to make screen as dark as darkest spots, to make it uniform. You'd lose too much dynamic range / lattitude, low light capability. The moving screen will average dark and bright spots. It is lot more efficient, especially because there is low percewntage of dark spots.

May 3rd, 2005, 05:30 AM	#6
Oscar Spierenburg Major Player Join Date: Jan 2005 Location: (The Netherlands - Belgium) Posts: 735	I think Daves got my point, but maybe I was not very clear. This is what I mean: Make a 35mm adapter with a static GG. Sample that static grain. Put that grain footage (just one frame, because the grain is static) in a post prog. in negative on top of other footage and mess around with opacity and layer type until you get the grain to disappear. So I'm not talking about moving GG or sampling the grain to put into other footage. Like Daves said, it would also reduce scratches. If it works.

May 3rd, 2005, 08:47 AM	#7
Aaron Shaw Major Player Join Date: Jul 2004 Posts: 587	Hey Oscar :) I started a similar thread a while back. Not sure where it is now. Anyway, it was tested and seemed to work out ok. Some people had trouble reproducing good results though. You need to make sure that the adapter is always oriented perfectly or when you try to remove the grain you will just add more.

May 3rd, 2005, 01:03 PM	#8
Oscar Spierenburg Major Player Join Date: Jan 2005 Location: (The Netherlands - Belgium) Posts: 735	Aaron, that's what I figured. The adapter has to stay on the camcorder and don't move at all or the grain will be displaced and you have to start again. I'll test this anyhow some time next week. I'll try to find your thread.