Strategies for Dealing With Moire?

[I Bloom]

I figure this is worth a thread.

Let's say your shooting a bunch of shots involving LCD screens. (Seems more and more likely that you'll be in that situation these days.)

In some situations the RGB pattern on the LCD will interfere with the grid pattern of the sensor and create a moire pattern. Sometimes this is impossible to avoid by keeping the image slightly out of focus, say for example you are racking focus across the surface of a screen.

I'm just wondering if others have dealt with this issue and what boundaries you set or innovative solutions you've come up with. Especially when you don't have a choice what type of screen you are shooting.

Cheers,
IBloom

 

[Roberto Lequeux]

Wouldn't a softening filter be the only solution? Perhaps something of the sort on the LCD itself if you don't want to soften the frame?

 

[hunterrichards]

I just shoot film and make a contact print, tell your clients that if their customers really want to see their product, they will have to sit in a theater.

Or replace the lcd in post or dont worry about it, 100% of people dont give a shit about miore on a computer/tv screen. Well maybe some do, you can always get an infront of the lens AA filter from caprock if needed. Ok now I sound like an ass...

 

[Jordan Livingston]

The responses above so far are pretty out of touch... I've shot numerous commercials involving LCD screens as products and it is ALWAYS a huge issue. The client doesn't want to comp out the screen content in many cases, especially if it is a touch screen product with human interaction. Roto work is time consuming and expensive.

My "solution" has always been to slightly de-focus the shot (as in, make the shot subtly out of focus); however, the bigger problem is the monitoring. The 720p output from the RED doesn't accurately portray what's going on. The moire on the HD monitors typically looks much worse than it really is in the recording. We have ultimately just had to shoot tests, load them into the computer, and made guess-and-check decisions from there.

I would love to hear about a more professional solution. Unfortunately, I've seen many RED commercials go to the F35 because of no other reason than the DP was concerned about not being able to properly monitor the amount of moire of an LCD product shot.

- Jordan

 

[David Mullen ASC]

I had this problem on Day One of this pilot I'm shooting on the Red -- after telling the costume person that the Red One has less moire problems than the Genesis/F35, I noticed that the shirt the lead actor was wearing -- which was a solid blue color Oxford but had a course weave to it -- was aliasing like crazy. I spent most of the day telling myself that it HAD to be the 720P half-rez downconvert to the monitor and it wasn't on the originals. At the end of the day, I looked at some 4K TIFF frames on my computer and confirmed that it wasn't in the original, just in the monitor output.

Now I could have done that earlier in the day obviously (grabbed a still frame and looked at it) but we were so slammed to keep shooting that I just made a guess that it wasn't a problem. At the end of the day, I could tell the costume person not to worry about it.

But in general, my technique with monitor screens is to shoot them slightly out of focus until the moire disappears. However, if I could instantly tell what was really getting recorded in 4K, I could back-off from the amount of defocusing probably.

 

[I Bloom]

I had this problem on Day One of this pilot I'm shooting on the Red -- after telling the costume person that the Red One has less moire problems than the Genesis/F35, I noticed that the shirt the lead actor was wearing -- which was a solid blue color Oxford but had a course weave to it -- was aliasing like crazy. I spent most of the day telling myself that it HAD to be the 720P half-rez downconvert to the monitor and it wasn't on the originals. At the end of the day, I looked at some 4K TIFF frames on my computer and confirmed that it wasn't in the original, just in the monitor output.

Now I could have done that earlier in the day obviously (grabbed a still frame and looked at it) but we were so slammed to keep shooting that I just made a guess that it wasn't a problem. At the end of the day, I could tell the costume person not to worry about it.

But in general, my technique with monitor screens is to shoot them slightly out of focus until the moire disappears. However, if I could instantly tell what was really getting recorded in 4K, I could back-off from the amount of defocusing probably.

I've never seen moire on clothing from a Red. I think LCD's are more of an issue. I'm actually doing a full debayer and 1080p scale of a bunch of test footage now to see if what was very clear in the eyepiece comes through the debayer.

IBloom

 

[Martin Weiss]

Don't have a camera to check right now, but couldn't you use the 1:1 zoom to check for moiré?

 

[Greg M]

Don't have a camera to check right now, but couldn't you use the 1:1 zoom to check for moiré?

excellent idea...we just wrapped a commercial shoot that presented several occasions where moire forced us to slightly change our shot or wardrobe. I wish I selected 1:1 to verify first. BTW- shooting on M-X.

But, consider, just looking at 1:1 may not be a solution, as ultimately a TVC will be down resed to 1080 (or less) and the moire could come back. I guess extensive tests are in order.

 

[Roberto Lequeux]

The responses above so far are pretty out of touch...

Any time! Though you must have missed the "?" at the end of the sentences, as you missed that he was asking for things other than defocusing.

Edit: Bloom has a good point. Were you doing a full debayer when you got moire from an R1?

ultimately a TVC will be down resed to 1080 (or less) and the moire could come back. I guess extensive tests are in order.

But in this case the culprit isn't the footage, but the compression. Though I guess your point is that sometimes you have no choice for compression so moire prone patterns are something you need to develop an eye for and get to know your deliverables. Right?

 

[Philip Allister Anderson]

A good trick is to put diffusion gel on screens that are in the background. Even something like 253 (hampshire) diffuses enough to get rid of the moire. Other then that green screen the screen and replace it later.

 

[Roberto Lequeux]

A good trick is to put diffusion gel on screens that are in the background. Even something like 253 (hampshire) diffuses enough to get rid of the moire. Other then that green screen the screen and replace it later.

That is what I thought. We are just coming from different sides. I wasn't thinking of a screen as the subject, but set dressing.

:grouphug:

I was just guessing on the whole putting a filter on the actual screen. Can you still make out what's on the screen?

 

[Sam Winzar]

Something I do to check if something is causing moiré or aliasing issues in the RAW 4K is to use the 1:1 Zoom. If you see the problem there, then it's really there. Saves having to offload footage to check.

 

[Graeme Nattress]

I would love to hear about a more professional solution. Unfortunately, I've seen many RED commercials go to the F35 because of no other reason than the DP was concerned about not being able to properly monitor the amount of moire of an LCD product shot.
- Jordan

That's bonkers. The F35 has demonstrably more aliasing than the RED, especially vertically where it's very noticeable. I keep mentioning last week's Dr Who which was on F35 and there were a good number of shots throughout the episode showing nasty aliasing.

Graeme

 

[Liam Hall]

But in this case the culprit isn't the footage, but the compression

You mean downsampling.

 

[Tim Whitcomb]

That's bonkers. The F35 has demonstrably more aliasing than the RED, especially vertically where it's very noticeable. I keep mentioning last week's Dr Who which was on F35 and there were a good number of shots throughout the episode showing nasty aliasing.

Graeme

I like the word BONKERS . Especially when it comes from Graeme!

 

[Sanjin Jukic]

No Moire with R1-M shooting Panasonic Plasma LCD (this is even with old model TH-PWD6):

From "The Dark Knight" Blu-Ray: Shot handheld on RED1 with Nikkor 17-35mm f/2.8 zoom @ 17mm wide open at f/2.8, ISO 250, 1/50 shutter.

Download 4K JPG>>>

From "The Dark Knight" Blu-Ray: Shot handheld on RED1 with Nikkor 17-35mm f/2.8 zoom @ 17mm wide open at f/2.8, ISO 250, 1/50 shutter.

Download 4K JPG>>>

Those shots were done at February 14, 2009.

 

[Graeme Nattress]

F35 is almost alias by design. Can't wait to get these new crop of digital cinema cameras on the charts and see how they perform as now we're seeing a pretty good correspondence between what we see on the charts and see in real world productions.

Graeme

 

[Roberto Lequeux]

You mean downsampling.

I see. That makes sense. Thank you Liam!!

SJ, that 4k is "So fresh and so clean, clean!". I do see quite bit of aliasing on the scaled down version you posted of that extreme close up of your TV. Then I downloaded the 4k jpg and scaled it down by 50% in Photoshop in 32-bit mode and got zero moire.

So this "proves" that R1 doesn't show moire in situations where many others would be going BONKERS.

On the other hand, will we be able to finish properly? We gotta keep it fresh and clean down the pipe. I suppose that is up to each project, and their post workflow.

1. Are there situations when you would be forced to downsample to a size that would make it harder to keep your material clean? If so, what acquisition and finishing resolutions are those?

2. And what software (or modes) should we stay away from?

Or just pay one of the many professionals who know how to finish RED properly... so many options, so many more coming. So many RT options affordable enough for me to build a station!

 

[Sanjin Jukic]

I see. That makes sense. Thank you Liam!!

SJ, that 4k is "so fresh and so clean, clean!". I do see quite bit of aliasing on the scaled down version you posted of that extreme close up of your TV. Then I downloaded the 4k jpg and scaled it down by 50% in Photoshop in 32-bit mode and got zero moire.

So this "proves" that R1 doesn't shot moire in situations where many others would be going BONKERS. However... will we be able to finish properly? I suppose that is up to each project, and their post workflow.

1. Are there situations when you would be forced to downsample to a size that would make it harder to keep your material clean? If so, what acquisition and finishing resolutions are those?

2. And what software (or modes) should we stay away from?

Roberto,

it's so clean visible the LCD pattern structure without any moire artifacts and it looks that I shot almost like "on canvas" :hurray:

 

[Greg M]

Unfortunately, I've seen many RED commercials go to the F35 because of no other reason than the DP was concerned about not being able to properly monitor the amount of moire of an LCD product shot.

- Jordan

For no other reason??? come on now.