Upping the Anti at DVinfo.net

James Miller · September 18th, 2009, 02:16 PM

Testing of an internal polarizing Anti Aliasing filter.

Have been testing this afternoon, marked improvement over things like tiled roofs etc.

Will post more details when I have them, ran out of light today. Also shot a circular zone test card for comparison, will conclude under more controlled conditions.

Up The Anti - Testing of the AR Filter on Vimeo

Of course there is always a trade off, induces slight blooming (not the Philip kind), lightens contrast, loss of light.

I do like the warm softness the shots have about them.

James

Sean Parker · September 18th, 2009, 05:37 PM

Wow, I thought that looked fantastic! I'm a little confused about what this is though — is this a lens filter or a software filter?

Thanks for sharing!

Jon Fairhurst · September 18th, 2009, 08:30 PM

I agree. It looks really smooth and creamy. It's a nice combo - less aliasing and a very nice look.

Dan Chung · September 18th, 2009, 08:42 PM

James,

Looks very interesting. Can you give us more details on this? I've been using a Schneider Digicon filter to try do the same sort of thing MPTV 4"x4" Digicon - Schneider Optics

Dan

(edit - interesting read on aliasing filters here http://luminous-landscape.com/forum/...howtopic=30758)

Evan Donn · September 18th, 2009, 11:20 PM

I don't know, to me it brings back the look I associate with lens adapters, which isn't a good thing...

James Miller · September 19th, 2009, 12:44 AM

Ok it's early here, so after I have made the wife a cup of tea i'll get on with testing.

Basically it's something I've been working on. An inexpensive set of polymer polarizing behind the lens filters. (Cost around UK £30)

Polarizing to: 1/ polarize if needed. 2/ fixed variable ND sandwich 3/ tinted color soft look.

Wet etched (still trying to perfect this) helps reduce the aliasing artifacts. Automated now, have ability for different strengths.

The filters are fiber paper edge bound that install on the camera body between it and the lens. Hence the need for a polymer based solution.

Prototypes hopefully within the next 4 weeks. I'll PM the 5D regular crew for their address if they would like a test a set.

James

Chris Barcellos · September 19th, 2009, 02:04 AM

So I downloaded the film, and I did a little bit level adjustment because it seemed gain was a bit high, and I think it looks pretty good. Would love to hear more about this process.

Richard Gooderick · September 19th, 2009, 03:06 AM

It would be useful to have a with and without.
A lot of it looks like what comes out of my 5Dmk2 in the garden, but with a different colour setting and softer-looking.
I tend to get aliasing with edges eg shopfront shutters under streetlights. Not in shots like this.
Still, if Dan Chung is excited about it I have obviously missed something!

Rick Hill · September 19th, 2009, 04:00 PM

Interesting. Can you make a video that really puts your solution to the test?

I find the most offensive aliasing occurs when filming any objects with lines (i.e. brick walls, house siding, automobile grills, etc.) with moving pans (vertical pan for horizontal lines and vice versa). This particular film has all the high frequency content stationary. I can get similar results with no modifications if I de-focus a bit.

Long wide pans of architectural scenes (with a high f-stops for large depth of field) always kill me.

This is very exciting if you have a good solution here.

Jon Fairhurst · September 19th, 2009, 04:57 PM

The other classic aliasing shot has been a body of water in the wind. The little whitecaps alias consistently with annoying color artifacts.

James Miller · September 20th, 2009, 12:10 PM

Here is a dirty little test leaning out my front door.

MobileMe - iDisk

There is only one folder "DVinfo Test"

Click the little grey down arrow at the end to download. I'll put this rubbish shot on Vimeo, but you get the idea.

2 x .mov h264 files (with / without) and custom picture profile used. Half resolution

Richard Gooderick · September 20th, 2009, 01:45 PM

Thanks James
That's very impressive.

Randy Panado · September 24th, 2009, 04:40 AM

Quote:

Originally Posted by James Miller

Ok it's early here, so after I have made the wife a cup of tea i'll get on with testing.

Basically it's something I've been working on. An inexpensive set of polymer polarizing behind the lens filters. (Cost around UK £30)

Polarizing to: 1/ polarize if needed. 2/ fixed variable ND sandwich 3/ tinted color soft look.

Wet etched (still trying to perfect this) helps reduce the aliasing artifacts. Automated now, have ability for different strengths.

The filters are fiber paper edge bound that install on the camera body between it and the lens. Hence the need for a polymer based solution.

Prototypes hopefully within the next 4 weeks. I'll PM the 5D regular crew for their address if they would like a test a set.

James

Can be used on 7D as well? If so, hook it up please :). I'll pay for shipping to test out.

James Miller · September 25th, 2009, 08:01 AM

Ok, Here are a couple of grabs taken from handheld, using the 50mm 1:4 @ f4.

I have found different lens need different strengths of filter.

One is without filter one with, both with identical camera settings.

I would normally run this through compressor with a +15 Sharpness, but this is the raw file opened in Quick Time and saved as a pict.

It's soft but workable.

Jon Fairhurst · September 25th, 2009, 10:30 AM

That's a bit too soft for my tastes. It could be just right on a wider angle lens though. From my experience, filters have very different looks on different lenses.

September 18th, 2009, 02:16 PM	#1
James Miller Major Player Join Date: Mar 2007 Location: Sussex, UK Posts: 317	Upping the Anti Testing of an internal polarizing Anti Aliasing filter. Have been testing this afternoon, marked improvement over things like tiled roofs etc. Will post more details when I have them, ran out of light today. Also shot a circular zone test card for comparison, will conclude under more controlled conditions. Up The Anti - Testing of the AR Filter on Vimeo Of course there is always a trade off, induces slight blooming (not the Philip kind), lightens contrast, loss of light. I do like the warm softness the shots have about them. James

September 18th, 2009, 08:30 PM	#3
Jon Fairhurst Inner Circle Join Date: May 2006 Location: Camas, WA, USA Posts: 5,513	I agree. It looks really smooth and creamy. It's a nice combo - less aliasing and a very nice look. __________________ Jon Fairhurst

September 18th, 2009, 08:42 PM	#4
Dan Chung Major Player Join Date: Apr 2008 Location: Beijing Posts: 665	James, Looks very interesting. Can you give us more details on this? I've been using a Schneider Digicon filter to try do the same sort of thing MPTV 4"x4" Digicon - Schneider Optics Dan (edit - interesting read on aliasing filters here http://luminous-landscape.com/forum/...howtopic=30758) Last edited by Dan Chung; September 18th, 2009 at 10:26 PM.

September 18th, 2009, 11:20 PM	#5
Evan Donn Major Player Join Date: Feb 2006 Location: Philadelphia Posts: 795	I don't know, to me it brings back the look I associate with lens adapters, which isn't a good thing... __________________ My latest short documentary: "Four Pauls: Bring the Hat Back!"

September 19th, 2009, 02:04 AM	#7
Chris Barcellos Inner Circle Join Date: Nov 2005 Location: Elk Grove CA Posts: 6,838	So I downloaded the film, and I did a little bit level adjustment because it seemed gain was a bit high, and I think it looks pretty good. Would love to hear more about this process. __________________ Chris J. Barcellos

September 18th, 2009, 05:37 PM	#2
Sean Parker Regular Crew Join Date: Aug 2009 Location: Portland, OR Posts: 36	Wow, I thought that looked fantastic! I'm a little confused about what this is though — is this a lens filter or a software filter? Thanks for sharing!

September 19th, 2009, 12:44 AM	#6
James Miller Major Player Join Date: Mar 2007 Location: Sussex, UK Posts: 317	Ok it's early here, so after I have made the wife a cup of tea i'll get on with testing. Basically it's something I've been working on. An inexpensive set of polymer polarizing behind the lens filters. (Cost around UK £30) Polarizing to: 1/ polarize if needed. 2/ fixed variable ND sandwich 3/ tinted color soft look. Wet etched (still trying to perfect this) helps reduce the aliasing artifacts. Automated now, have ability for different strengths. The filters are fiber paper edge bound that install on the camera body between it and the lens. Hence the need for a polymer based solution. Prototypes hopefully within the next 4 weeks. I'll PM the 5D regular crew for their address if they would like a test a set. James

September 19th, 2009, 03:06 AM	#8
Richard Gooderick Major Player Join Date: Oct 2007 Location: London, UK Posts: 795	It would be useful to have a with and without. A lot of it looks like what comes out of my 5Dmk2 in the garden, but with a different colour setting and softer-looking. I tend to get aliasing with edges eg shopfront shutters under streetlights. Not in shots like this. Still, if Dan Chung is excited about it I have obviously missed something! __________________ http://www.gooderick.com

September 19th, 2009, 04:00 PM	#9
Rick Hill Regular Crew Join Date: Aug 2008 Location: McHenry, Illinois Posts: 55	Interesting. Can you make a video that really puts your solution to the test? I find the most offensive aliasing occurs when filming any objects with lines (i.e. brick walls, house siding, automobile grills, etc.) with moving pans (vertical pan for horizontal lines and vice versa). This particular film has all the high frequency content stationary. I can get similar results with no modifications if I de-focus a bit. Long wide pans of architectural scenes (with a high f-stops for large depth of field) always kill me. This is very exciting if you have a good solution here.

September 19th, 2009, 04:57 PM	#10
Jon Fairhurst Inner Circle Join Date: May 2006 Location: Camas, WA, USA Posts: 5,513	The other classic aliasing shot has been a body of water in the wind. The little whitecaps alias consistently with annoying color artifacts. __________________ Jon Fairhurst

September 20th, 2009, 12:10 PM	#11
James Miller Major Player Join Date: Mar 2007 Location: Sussex, UK Posts: 317	Here is a dirty little test leaning out my front door. MobileMe - iDisk There is only one folder "DVinfo Test" Click the little grey down arrow at the end to download. I'll put this rubbish shot on Vimeo, but you get the idea. 2 x .mov h264 files (with / without) and custom picture profile used. Half resolution

September 20th, 2009, 01:45 PM	#12
Richard Gooderick Major Player Join Date: Oct 2007 Location: London, UK Posts: 795	Thanks James That's very impressive. __________________ http://www.gooderick.com

September 25th, 2009, 08:01 AM	#14
James Miller Major Player Join Date: Mar 2007 Location: Sussex, UK Posts: 317	Ok, Here are a couple of grabs taken from handheld, using the 50mm 1:4 @ f4. I have found different lens need different strengths of filter. One is without filter one with, both with identical camera settings. I would normally run this through compressor with a +15 Sharpness, but this is the raw file opened in Quick Time and saved as a pict. It's soft but workable. Attached Thumbnails

September 25th, 2009, 10:30 AM	#15
Jon Fairhurst Inner Circle Join Date: May 2006 Location: Camas, WA, USA Posts: 5,513	That's a bit too soft for my tastes. It could be just right on a wider angle lens though. From my experience, filters have very different looks on different lenses. __________________ Jon Fairhurst