Filter for IR contamination

Dean Sensui · February 28th, 2009, 04:00 PM

It's occurred to me that the EX1 may be exhibiting a sensitivity to a type of "metamerism" or the way a pigment or dye changes its appearance under different types of lights.

People who do critical color printing always judge the quality of a print under lights that have the spectral distribution of daylight. These lamps are carefully produced and tested, and they're checked during their lifespan to ensure high quality control. Pigments that look blue under one type of light might appear almost purple under another type of light. So printers who are picky about quality and consistency always use a standardized light source under controlled conditions (a viewing booth) to judge their work.

Because of its extended sensitivity to wavelengths in excess of 650 nanometers, tungsten lights would tend to exaggerate that quality in certain materials. Tungsten lighting's spectral distribution is mostly in the longer wavelengths. That's why what looks "black" to the human eye in tungsten light looks "red" to the EX1.

Under LED lights and fluorescents, those same materials might not exhibit the same degree of redness. And, of course, with a filter that attenuates wavelengths longer than 650 nm, that redness is also reduced.

Graeme Fullick · February 28th, 2009, 04:25 PM

Dean,

I think that you might correct here. I often see what the EX1 sees - a redness in the blacks of some "black fabrics" under some lighting conditions. The EX1 sees it more than me, but I can clearly see it. This does not extend to all fabrics (such as suits), but the synthetics such as nylon are the worst offenders for redness to my eye.

I guess that everyones vision is slightly different.

Leonard Levy · February 28th, 2009, 04:27 PM

Dean - yes,

- the second issue of course is the ND problem , because even though you are under daylight outdoors, , normal ND filters don't cut the IR portion of the spectrum so the IR portion cumulatively and disproportionally increases with increasing strength of ND filtering.

Art Adams only addressed the latter question in his post. I've been trying to contact him privately about it but haven't gotten a response yet.

Bob Grant · February 28th, 2009, 05:40 PM

It's true that our vision doesn't simply stop working at a certain wavelength anymore than our hearing stops at a given wavelength of sound. No doubt it also varies over the population. Certainly some IR LEDs used with security cameras are visible if you look into them, you'll see a dull red glow. I think we retain some sensitivity out to 800nm.

The problem is that such wavelengths are all simply seen as red by our cameras, keep in mind that neither video or film records the wavelength of light, only the relative amounts of the primary colors.

There's one aspect to this not mentioned so far. If the EX did not retain some sensitivity beyond the visible spectrum that would preclude its use for night time wildlife photography. IR illuminators would be useless.
One type of illuminator seems to use 940nm LEDs. These are totally invisible and only work with cameras with no IR filter or one where the IR filter can be removed. This is the best system as silicon is most sensitive at 1000nm.
The other illuminator uses 850nm LEDs for cameras with fixed IR filters. These LEDs can be seen as a dull red glow. I might buy a cheap one of these to see how they perform with the EX.

Piotr Wozniacki · March 1st, 2009, 03:33 AM

Considering that normal ND's cut up up to some 680 nm, using one outdoors should create a "hole" between 680nm and where the internal hot mirror starts to act (700-750) for long reds to leak into the sensor, as the one that exists when shooting indoors, without any ND but under a lighting which is low and shifted towards red when compared with sunlight.

And indeed - today I took a closer look at some black fabrics notorious of becoming red in tungsten, but put on the ND filter and shot it in the sunlight: it's maroon, as expected !

So it seems we understand now what's happening; what's more - it's probably not another Hot Mirror that we need to prevent it (as Art Adams points out in his another article here:

ProVideo Coalition.com: Stunning Good Looks by Art Adams | Cinematography)

To the F35 (and also EX) owners, he says:

"don’t buy Hot Mirror filters for your Genesis/F35 because it’s a waste of money."

But unfortunately, in all those articles the cure, proposed for these cameras, are special kind of ND filters from Tiffen; what we REALLY need is an equally efficient filter without any ND component, as we need it for low-light shooting!

Bob Grant · March 1st, 2009, 07:16 AM

Quote:

Originally Posted by Piotr Wozniacki

Considering that normal ND's cut up up to some 680 nm, using one outdoors should create a similar "hole" between 680nm and where the internal hot mirror operates (750?) for long reds to leak into the sensor, as exists when shooting indoors, without any ND but under a lighting which is low and shifted towards red when compared with sunlight.

And indeed - today I took a closer look at some black fabrics notorious of becoming red in tungsten, but put on the ND2 filter and shot it in the sunlight: it's maroon, as expected !

So it seems we understand now what's happening; what's more - it's probably not another Hot Mirror that we need to prevent it (as Art Adams points out in his another article here:

ProVideo Coalition.com: Stunning Good Looks by Art Adams | Cinematography)

To the F35 (and also EX) owners, he says:

"don’t buy Hot Mirror filters for your Genesis/F35 because it’s a waste of money."

But unfortunately, in all those articles the cure, proposed for these cameras, are special kind of ND filters from Tiffen; what we REALLY need is an equally efficient filter without any ND component, as we need it for low-light shooting!

Take a careful look at page 5 of the article.
The first shots with 0.9 ND filter show extreme red shift in the fabric on the right side of the chart. I'd suggest the ND filter is making the problem worse as it's cutting visible and not far red i.e. it's in effect adding gain to the far red.
Now look at the last two shots. There's still some red shift in the fabric. The extra dye has corrected the ND out to the far red. As Art says the amount of dye has to balance the ND. Probably if it doesn't either the red shift is back or it introduces a shift in the rest of the spectrum. It's tempting to hope that a simple and cheap dye filter will cure our problem but I suspect it isn't that simple.

The 486 is a pretty complex filter with I think over 30 layers of different materials, probably each one provides a narrow notch filter which is why it can achieve such a sharp cutoff just outside the visible spectrum. Maybe our ideal filter doesn't need so many layers as we don't need IR cut however to achieve such a sharp cutoff I think we'll still need this kind of filter and that'll mean some degree of color shift at low angles. Tilting the 486 and watching the rainbow of colors it reflects I'm amazed there isn't more shifts with this filter.

Piotr Wozniacki · March 1st, 2009, 07:26 AM

Quote:

Originally Posted by Bob Grant

Take a careful look at page 5 of the article.
The first shots with 0.9 ND filter show extreme red shift in the fabric on the right side of the chart. I'd suggest the ND filter is making the problem worse as it's cutting visible and not far red i.e. it's in effect adding gain to the far red.

That's exactly what I was trying to say - sorry for my English :)

As to the 486, it's amazing indeed how effective it is - both under tungsten, and sunlight (even with ND on). However, the low angles vignetting is not the only problem that led me to sell it away (cheaply); as I'm often using my 35 mm adapter/lens, I needed a 4x5.65" version (rather than a screw-on) to be put as the first optical element in my matte box.

Piotr Wozniacki · March 1st, 2009, 08:58 AM

From the series of Art's articles it seems that the answer would be an IR filter like Tiffen IR ND (not a hot mirror, i.e. without the dichroic coating), that cuts at some 680 nm. Not only is it the most effective on the red contamination type found on the CineAlta series, but it doesn't introduce any vignetting at wide angles. And, it's much cheaper than hot mirrors!

However, I'd prefer such a filter without any ND component to it, so that I don't lose any light at all - now, does such a beast exist?

The weakest ND one I found is : http://www.bhphotovideo.com/c/produc...y_0_3_ND_.html - just one stop light loss. The catch is that - as Art points out - those filters cannot be combined with other ND's, so we're sort of trapped...

Ronn Kilby · March 1st, 2009, 12:01 PM

The ideal solution would be if Sony (the upgrade kings) offered to replace the built-in ND filters with IR-ND filters. Of course they'd also change the ones on the assembly line, calling them "EX-1B" and "EX-3B" and charging more.

Chuck Fishbein · March 1st, 2009, 09:47 PM

"The ideal solution would be if Sony (the upgrade kings) offered to replace the built-in ND filters with IR-ND filters".

Here, here! I agree.

Anthony McErlean · March 23rd, 2009, 09:14 AM

So I suppose theres nothing for it but to get the B+W 486 filter for my EX3 advised by Brian, (thanks)

I think Brian you supplied this link of us in the UK.. B&W UV-IR Digital (486) 77....

B&W UV-IR Digital 77: Amazon.co.uk: Electronics & Photo

I saw in a post Brian you thought your LED On-board Camera light might have helped the blacks to stay black. Have to tested any more on this.

Thanks.

Brian Cassar · March 23rd, 2009, 09:31 AM

Anthony I found out two things lately:

1)If a potential black fabric which is turning into brown in a tungsten lit room is blasted with light from an LED on board lamp, then the black becomes black again. The reason being that LED light is overpowering the halogen / tungsten light which has a component of IR. That is why I found the Zylight 90 so effective as I can dial in any colour temp that I need so as to match the prevailing light. I did not have time to experiment much on this but I did see this effect. I generally prefer using the 486 and be able to film under the available light rather than switching on the on-board light.

2)Lately I was convinced in buying and using the warm cards. The footage does become much nicer than standard white balancing on a piece of paper. However I've noticed that if one uses the 1/2 warm filter, this will tend to enhance the IR contamination if filming under tungsten or in daylight.

Incidentally for just one instance I thought that I have solved this issue. I white balanced on the grey card that is included in the warm cards pack and I managed to see a black t-shirt as black under halogen light. I just couldn't believe it and I took another white balance and this time although it was the same colour reading of 2300K the t-shirt turned brownish. I've tried several times to get the same effect but couldn't. This one off event led me to re-confirm what others were hinting (I believe Piotr was one of them) that this IR contamination might be somehow linked to white balancing as well. It is not just a question of white balancing but probably a series of other things.

Anthony McErlean · March 23rd, 2009, 09:45 AM

Hi Brian

Quote:

Originally Posted by Brian Cassar

I white balanced on the grey card that is included in the warm cards pack and I managed to see a black t-shirt as black under halogen light. I just couldn't believe it and I took another white balance and this time although it was the same colour reading of 2300K the t-shirt turned brownish....

Thats interesting, you could be on to something there.

I read somewhere that a newspaper, a fews days old, was good to white balance on, I don't know, I never tried it. A polystyrene cup was something else that was suggested.

Just wonder after reading the posts regarding this, have Sony themselves comeback with any new info.

Dave Morrison · March 23rd, 2009, 10:27 AM

It sounds to me like what you are seeing is the side-effect of "fooling" the camera. For example, by using the off-color white balance cards (or the slightly yellowed day-old newspaper), you are skewing the white balance of the camera slightly toward the blue end of the spectrum causing the blacks to "cool down" (along with all the other colors, btw) thereby making the brown-blacks appear cooler and "blacker", no?

Brian Cassar · March 23rd, 2009, 10:58 AM

Oh...I forgot to mention one very important observation. When I managed to restore black to black for just one instance only by taking a WB reading from a grey card, the reds in general were very much muted (obviously). Which leads to more speculation whether it is really a case where the EX cameras are very sensitive to the red colour. So in order to get decent black one will loose the vividness of the red color. Hence that is why probably some people have suggested to work out a picture profile for this issue. However this will be at a cost for the overall warmth of the picture since the reds would be curtailed.

February 28th, 2009, 04:00 PM	#91
Dean Sensui Inner Circle Join Date: Oct 2001 Location: Honolulu, HI Posts: 2,054	It's occurred to me that the EX1 may be exhibiting a sensitivity to a type of "metamerism" or the way a pigment or dye changes its appearance under different types of lights. People who do critical color printing always judge the quality of a print under lights that have the spectral distribution of daylight. These lamps are carefully produced and tested, and they're checked during their lifespan to ensure high quality control. Pigments that look blue under one type of light might appear almost purple under another type of light. So printers who are picky about quality and consistency always use a standardized light source under controlled conditions (a viewing booth) to judge their work. Because of its extended sensitivity to wavelengths in excess of 650 nanometers, tungsten lights would tend to exaggerate that quality in certain materials. Tungsten lighting's spectral distribution is mostly in the longer wavelengths. That's why what looks "black" to the human eye in tungsten light looks "red" to the EX1. Under LED lights and fluorescents, those same materials might not exhibit the same degree of redness. And, of course, with a filter that attenuates wavelengths longer than 650 nm, that redness is also reduced. __________________ Dean Sensui Exec Producer, Hawaii Goes Fishing

February 28th, 2009, 04:25 PM	#92
Graeme Fullick Major Player Join Date: Mar 2005 Location: Newcastle, Australia Posts: 337	Dean, I think that you might correct here. I often see what the EX1 sees - a redness in the blacks of some "black fabrics" under some lighting conditions. The EX1 sees it more than me, but I can clearly see it. This does not extend to all fabrics (such as suits), but the synthetics such as nylon are the worst offenders for redness to my eye. I guess that everyones vision is slightly different. __________________ Graeme

March 1st, 2009, 03:33 AM	#95
Piotr Wozniacki Inner Circle Join Date: Aug 2006 Location: Poland Posts: 4,086	Considering that normal ND's cut up up to some 680 nm, using one outdoors should create a "hole" between 680nm and where the internal hot mirror starts to act (700-750) for long reds to leak into the sensor, as the one that exists when shooting indoors, without any ND but under a lighting which is low and shifted towards red when compared with sunlight. And indeed - today I took a closer look at some black fabrics notorious of becoming red in tungsten, but put on the ND filter and shot it in the sunlight: it's maroon, as expected ! So it seems we understand now what's happening; what's more - it's probably not another Hot Mirror that we need to prevent it (as Art Adams points out in his another article here: ProVideo Coalition.com: Stunning Good Looks by Art Adams \| Cinematography) To the F35 (and also EX) owners, he says: "don’t buy Hot Mirror filters for your Genesis/F35 because it’s a waste of money." But unfortunately, in all those articles the cure, proposed for these cameras, are special kind of ND filters from Tiffen; what we REALLY need is an equally efficient filter without any ND component, as we need it for low-light shooting! __________________ Sony PXW-FS7 \| DaVinci Resolve Studio; Magix Vegas Pro; i7-5960X CPU; 64 GB RAM; 2x GTX 1080 8GB GPU; Decklink 4K Extreme 12G; 4x 3TB WD Black in RAID 0; 1TB M.2 NVMe cache drive Last edited by Piotr Wozniacki; March 1st, 2009 at 08:47 AM.

March 1st, 2009, 08:58 AM	#98
Piotr Wozniacki Inner Circle Join Date: Aug 2006 Location: Poland Posts: 4,086	From the series of Art's articles it seems that the answer would be an IR filter like Tiffen IR ND (not a hot mirror, i.e. without the dichroic coating), that cuts at some 680 nm. Not only is it the most effective on the red contamination type found on the CineAlta series, but it doesn't introduce any vignetting at wide angles. And, it's much cheaper than hot mirrors! However, I'd prefer such a filter without any ND component to it, so that I don't lose any light at all - now, does such a beast exist? The weakest ND one I found is : http://www.bhphotovideo.com/c/produc...y_0_3_ND_.html - just one stop light loss. The catch is that - as Art points out - those filters cannot be combined with other ND's, so we're sort of trapped... __________________ Sony PXW-FS7 \| DaVinci Resolve Studio; Magix Vegas Pro; i7-5960X CPU; 64 GB RAM; 2x GTX 1080 8GB GPU; Decklink 4K Extreme 12G; 4x 3TB WD Black in RAID 0; 1TB M.2 NVMe cache drive

March 23rd, 2009, 10:27 AM	#104
Dave Morrison Major Player Join Date: May 2002 Location: West Central Florida Posts: 762	It sounds to me like what you are seeing is the side-effect of "fooling" the camera. For example, by using the off-color white balance cards (or the slightly yellowed day-old newspaper), you are skewing the white balance of the camera slightly toward the blue end of the spectrum causing the blacks to "cool down" (along with all the other colors, btw) thereby making the brown-blacks appear cooler and "blacker", no? __________________ EX1, Nikon D750 http://www.morrisonphotographics.com http://www.heritagefilm.com

February 28th, 2009, 04:27 PM	#93
Leonard Levy Trustee Join Date: Feb 2006 Location: USA Posts: 1,684	Dean - yes, - the second issue of course is the ND problem , because even though you are under daylight outdoors, , normal ND filters don't cut the IR portion of the spectrum so the IR portion cumulatively and disproportionally increases with increasing strength of ND filtering. Art Adams only addressed the latter question in his post. I've been trying to contact him privately about it but haven't gotten a response yet.

February 28th, 2009, 05:40 PM	#94
Bob Grant Trustee Join Date: Nov 2005 Location: Sydney Australia Posts: 1,570	It's true that our vision doesn't simply stop working at a certain wavelength anymore than our hearing stops at a given wavelength of sound. No doubt it also varies over the population. Certainly some IR LEDs used with security cameras are visible if you look into them, you'll see a dull red glow. I think we retain some sensitivity out to 800nm. The problem is that such wavelengths are all simply seen as red by our cameras, keep in mind that neither video or film records the wavelength of light, only the relative amounts of the primary colors. There's one aspect to this not mentioned so far. If the EX did not retain some sensitivity beyond the visible spectrum that would preclude its use for night time wildlife photography. IR illuminators would be useless. One type of illuminator seems to use 940nm LEDs. These are totally invisible and only work with cameras with no IR filter or one where the IR filter can be removed. This is the best system as silicon is most sensitive at 1000nm. The other illuminator uses 850nm LEDs for cameras with fixed IR filters. These LEDs can be seen as a dull red glow. I might buy a cheap one of these to see how they perform with the EX.

March 1st, 2009, 12:01 PM	#99
Ronn Kilby Major Player Join Date: Apr 2007 Location: Fletcher Hills, CA Posts: 211	The ideal solution would be if Sony (the upgrade kings) offered to replace the built-in ND filters with IR-ND filters. Of course they'd also change the ones on the assembly line, calling them "EX-1B" and "EX-3B" and charging more.

March 1st, 2009, 09:47 PM	#100
Chuck Fishbein Major Player Join Date: Jun 2008 Location: Weehawken New Jersey Posts: 210	"The ideal solution would be if Sony (the upgrade kings) offered to replace the built-in ND filters with IR-ND filters". Here, here! I agree.

March 23rd, 2009, 09:14 AM	#101
Anthony McErlean Trustee Join Date: Oct 2008 Location: Northern Ireland Posts: 1,082	So I suppose theres nothing for it but to get the B+W 486 filter for my EX3 advised by Brian, (thanks) I think Brian you supplied this link of us in the UK.. B&W UV-IR Digital (486) 77.... B&W UV-IR Digital 77: Amazon.co.uk: Electronics & Photo I saw in a post Brian you thought your LED On-board Camera light might have helped the blacks to stay black. Have to tested any more on this. Thanks.

March 23rd, 2009, 09:31 AM	#102
Brian Cassar Major Player Join Date: Jan 2007 Location: Malta Posts: 306	Anthony I found out two things lately: 1)If a potential black fabric which is turning into brown in a tungsten lit room is blasted with light from an LED on board lamp, then the black becomes black again. The reason being that LED light is overpowering the halogen / tungsten light which has a component of IR. That is why I found the Zylight 90 so effective as I can dial in any colour temp that I need so as to match the prevailing light. I did not have time to experiment much on this but I did see this effect. I generally prefer using the 486 and be able to film under the available light rather than switching on the on-board light. 2)Lately I was convinced in buying and using the warm cards. The footage does become much nicer than standard white balancing on a piece of paper. However I've noticed that if one uses the 1/2 warm filter, this will tend to enhance the IR contamination if filming under tungsten or in daylight. Incidentally for just one instance I thought that I have solved this issue. I white balanced on the grey card that is included in the warm cards pack and I managed to see a black t-shirt as black under halogen light. I just couldn't believe it and I took another white balance and this time although it was the same colour reading of 2300K the t-shirt turned brownish. I've tried several times to get the same effect but couldn't. This one off event led me to re-confirm what others were hinting (I believe Piotr was one of them) that this IR contamination might be somehow linked to white balancing as well. It is not just a question of white balancing but probably a series of other things.

March 23rd, 2009, 10:58 AM	#105
Brian Cassar Major Player Join Date: Jan 2007 Location: Malta Posts: 306	Oh...I forgot to mention one very important observation. When I managed to restore black to black for just one instance only by taking a WB reading from a grey card, the reds in general were very much muted (obviously). Which leads to more speculation whether it is really a case where the EX cameras are very sensitive to the red colour. So in order to get decent black one will loose the vividness of the red color. Hence that is why probably some people have suggested to work out a picture profile for this issue. However this will be at a cost for the overall warmth of the picture since the reds would be curtailed.