Budget variable ND filter at DVinfo.net

Nigel Barker · May 16th, 2009, 12:50 AM

I was going to post this as a reply to Dan Brockett's excellent report of using the Singh-Ray variable density filter http://www.dvinfo.net/conf/canon-eos...ml#post1143751 but decided not to clutter up his thread as this is a discussion of the budget rather than the luxury option.

From using ND filters on 5DII I was convinced of the value for regaining aperture control but screw on & drop in filters are not as flexible as some sort of variable device. I was deterred by the price of the Singh-Ray variable ND filter & reluctant to drop $400 without proof that it would do the job for me so decided to investigate the DIY route. From my research it appears that the Singh-Ray gizmo is either two circular polarisers back to back or a linear polariser combined with a circular polariser.

With film cameras you can simply stack two linear polarising filters & rotate them to produce a variable density filter. It's just like the old trick to demonstrate the properties of Polaroid materials by taking two pairs of Polaroid sunglasses & rotating them relative to one another. While linear polarising filters can be used with film cameras apparently modern digital cameras with auto-focus, auto-exposure & whatnot do not work properly with polarised light so a circular polarisation filter must be used which is sort of like a linear polarisation filter except there is an extra layer that 'unscrambles' or 'de-polarises' the light before it enters the camera.

To experiment I bought a selection of cheap used polarising filters from eBay & they just arrived in the last couple of days. Nowadays linear polarising filters are more difficult to find & I only saw used ones but they are dirt cheap. I now have a number of filters some labelled as linear, some as circular & some as PL or polarising. If I combine two I can easily produce my own variable density filter just like with two pairs of sunglasses.

The good news is that it seems not to matter at all whether I use a pair of linear polarising filters or a linear combined with a circular not only do I produce the desired variable density effect but also the camera works as normal i.e. auto-focus is OK & auto-exposure is OK as the the aperture, ISO & shutter speed changes as the filter is lightened & darkened. As has been previously observed by others the changing of aperture, ISO & shutter speed on the 5DII in movie mode varies so e.g. the aperture will open up first then the ISO will rise & the different controls seem to work independently of each other.

With my home made set-up it's just the same as Dan B. has noted with the Singh-Ray. Spin the dial until the aperture, ISO & shutter speed are just what you want & then lock the exposure. It's as simple as that. I haven't had time to experiment much just enough to establish that the principle works. In particular I haven't done much shooting with & without filters to make sure that there is no colour cast or any other quality problem. I did see a pronounced blue tinge when the filters were stopped right down to nearly black which may be quite cool for fake night scenes. Even though I live on the Cote d'Azur it was actually pouring down with rain yesterday so my experimentation was limited but today is bright & sunny & we are off doing a shoot so I will be taking all my filters to try them out.

I don't know how controllable the Singh-Ray filter is. My DIY filters go from fully open to fully closed in a quarter turn. The rotation also seems pretty loose so after being fixed in one position I can see that it might easily slip but they may just be the old used filters that I bought. So far I am really pleased with my tests & actually buy a Singh-Ray if I can get the same effect & have it reliably controllable.

Dan Chung · May 16th, 2009, 01:52 AM

Nigel,

The reason to use 2 circular polarizers vs linear ones is to reduce the polarizing effect of the filters which can otherwise cause problems with changing ND value while shooting a scene with reflections, or when you are shooting water and glass. Otherwise the effect will be similar.

The other thing to watch for with home brew variable ND is that you can get some pretty bad colour shifts with some polarizing filters.

I now have various options for ND, the best for control is 2 4x4 square polarizers in a Matte box allowing to remove the ND completely in a hurry. It is not as portable as a simple screw in Singh-ray however.

Dan

Dan Brockett · May 16th, 2009, 09:49 AM

Hi Nigel:

I have read about others doing the same thing, and from what I have read, it works decently. But there is some voodoo at work with the DIY solutions that will cause random color shifts, which for me, are a PITA to try to fix.

That is what impressed me so far with the Singh Ray footage, there doesn't seem to be any color shift as I rotate the filter. I am sure that there is some sort of optical engineering going on with the Singh Ray or they wouldn't be getting so much money for it.

On another note, I am going to be staying in Mougins in your neck of the woods in late June, I would welcome any suggestions from you about must see/must do experiences. I am shooting a doc, following a subject around the South of France and any visually beautiful location recommendations would be welcomed.

Thanks so much,

Dan

Dan Chung · May 16th, 2009, 07:01 PM

Dan,

Out of interest which vari-ND did you go for and what the widest lens you can use it with without cut off in the corners?

I might be asking you to buy me one and stick it in the post to China!

Dan

Nigel Barker · May 16th, 2009, 11:16 PM

Quote:

Originally Posted by Dan Chung

The reason to use 2 circular polarizers vs linear ones is to reduce the polarizing effect of the filters which can otherwise cause problems with changing ND value while shooting a scene with reflections, or when you are shooting water and glass. Otherwise the effect will be similar.

The interesting thing is that according to all that I have read is that a linear polariser interferes with the auto-focus & auto-exposure of a modern camera (film or digital). That is why it is almost entirely circular polarisers that are on the market today. However it seems to make no difference when I put a linear polarising filter on as the 5DII works just fine focusing & adjusting exposure

Quote:

Originally Posted by Dan Chung

The other thing to watch for with home brew variable ND is that you can get some pretty bad colour shifts with some polarizing filters.

I haven't seen that so perhaps I have just been lucky with the filters that I acquired. I have found one pair of two filters that displays vignetting but the combo does end up rather thick.

Nigel Barker · May 16th, 2009, 11:21 PM

Quote:

Originally Posted by Dan Brockett

On another note, I am going to be staying in Mougins in your neck of the woods in late June, I would welcome any suggestions from you about must see/must do experiences. I am shooting a doc, following a subject around the South of France and any visually beautiful location recommendations would be welcomed.

I live about 10 minutes drive away from Mougins. There are many, many beautiful locations around here & I would love to help out. Let's take this off-line. I have sent a PM with my email address.

Cheers

Nigel

Christopher Lovenguth · May 22nd, 2009, 10:02 AM

I picked up a linear PL yesterday and just finished a quick test to post.

DIY Vari ND filter on 5D II on Vimeo

Using 50mm set at F1.8 using lens twist. First filter on lens is Hoya Circular PL size 77 (using step-up ring). I set the circular PL where it's not affecting any glare. I then taped this filter so it couldn't move. I then screwed in front of circular PL a Hoya PL(linear). I locked exposure at 40 and iso 100 finding the right exposure by twisting the linear filter. Without the filters the exposure was 2000 at iso 100 with F1.8. Just used auto WB for this quick test. During the recording I rotated the linear filter to change the exposure.

I'm shooting through dusty warehouse windows except for the upper right corner with the tree, this area has no glass.

I had no issues with 5D II finding exposure with the linear filter on lens in live view. I'm using a 77 size filter only because that is the largest size lens I have in my arsenal and will be using step up rings for all other size lenses. I didn't notice much color shifting with the Hoya's. It's strange however, if you place the circular in front of the linear all that happens is a strange color shifting and very little density change. I don't have two circular or two linear filters to test these options.

The blue color cast when the image is almost black is most likely due to the blue/UV area of the spectrum of light is the brightest and most sensitive to human eyes (that's why blue light alarm clocks are so bright when really dim as opposed to red light).

Robert Esmonde · May 23rd, 2009, 07:29 AM

Thanks Christopher for that really interesting clip and info.

Can you say how you mounted the linear filter in front of the circular CPL one? Were you able to just screw it into the CPL filter or did you have to reverse the LP filter and use some kind of reversing ring to connect it to the CPL filter?

Christopher Lovenguth · May 23rd, 2009, 07:02 PM

Hi Robert,

I just screwed the linear on the front of the circular.

Robert Esmonde · May 24th, 2009, 02:57 AM

Thanks Christopher.

I understood that the front LP filter needed to be reversed, so it's great news to hear that it works fine just screwed into the CPL one.

Nigel Barker · May 24th, 2009, 04:07 AM

I just double checked on my set-up. Like Christopher's it only works properly when the linear polariser is screwed on the front of the circular polariser.

Robert Esmonde · May 24th, 2009, 04:35 AM

Thanks Nigel. I was not referring to the stacking order of the filters. I had read somewhere else (maybe this thread?) that when stacked with the CPL nearest to the lens, the outer Linear Polariser filter had to be turned backwards and needed a reversing ring to connect to the CPL filter.

From what Christopher has said, it seems that just screwing the Linear Polariser into the CPL filter works fine, and there's no need for the complexity and extra depth of a reversing ring.

Nigel Barker · May 24th, 2009, 06:07 AM

Robert, sorry my response must have been confusing. I was trying to confirm that a reversing ring is not needed. I too had read that was necessary but it appears not to be the case.

I also read that the reason that we have circular polarisers nowadays instead of the older linear polarisers is that auto-focus on modern cameras does not work properly with polarised light so a circular polariser actually has an extra layer that effectively scrambles up the light so that it is not polarised when it enters the camera lens. The 5DII auto-focus seems to work OK whatever type of polariser or combination of polarisers I have on the lens.

Robert Esmonde · May 24th, 2009, 06:41 AM

Thanks Nigel.

That clears things up for me. Thanks too for your detailed opening post. Really useful.

Christopher Lovenguth · May 24th, 2009, 11:45 AM

I would guess the only "benefit" to stacking the glass as close as possible to each other with using a reversing ring would be to to get the glass without any gaps so there is less likely any possibility for issues of distortion, dust, etc. But in adding the reversing ring, it's most likely the same gap. I'm such a purest that I hate putting any filters on my lenses but ND is a must for shooting outside with a low depth of field and can't afford $400 for the Singh-Ray when I still need a field audio recorder, etc. I already had the circular PL from the original 5D kit I bought years ago off someone and never used and it's pristine. So my vari-ND cost me about $60 with all the step-up rings and the 77mm Hoya Linear I bought new from Adorama. I am in no way dismissing what is most likely a great product from Singh-Ray and with there not being a huge market for something like this I can understand the price. But if you work with the limitations of DIY it can be done for less then half the cost by picking up a CPL and PL or maybe even just two PL's.

May 16th, 2009, 12:50 AM	#1
Nigel Barker Inner Circle Join Date: Feb 2009 Location: Norwich, Norfolk, UK Posts: 3,531	Budget variable ND filter I was going to post this as a reply to Dan Brockett's excellent report of using the Singh-Ray variable density filter http://www.dvinfo.net/conf/canon-eos...ml#post1143751 but decided not to clutter up his thread as this is a discussion of the budget rather than the luxury option. From using ND filters on 5DII I was convinced of the value for regaining aperture control but screw on & drop in filters are not as flexible as some sort of variable device. I was deterred by the price of the Singh-Ray variable ND filter & reluctant to drop $400 without proof that it would do the job for me so decided to investigate the DIY route. From my research it appears that the Singh-Ray gizmo is either two circular polarisers back to back or a linear polariser combined with a circular polariser. With film cameras you can simply stack two linear polarising filters & rotate them to produce a variable density filter. It's just like the old trick to demonstrate the properties of Polaroid materials by taking two pairs of Polaroid sunglasses & rotating them relative to one another. While linear polarising filters can be used with film cameras apparently modern digital cameras with auto-focus, auto-exposure & whatnot do not work properly with polarised light so a circular polarisation filter must be used which is sort of like a linear polarisation filter except there is an extra layer that 'unscrambles' or 'de-polarises' the light before it enters the camera. To experiment I bought a selection of cheap used polarising filters from eBay & they just arrived in the last couple of days. Nowadays linear polarising filters are more difficult to find & I only saw used ones but they are dirt cheap. I now have a number of filters some labelled as linear, some as circular & some as PL or polarising. If I combine two I can easily produce my own variable density filter just like with two pairs of sunglasses. The good news is that it seems not to matter at all whether I use a pair of linear polarising filters or a linear combined with a circular not only do I produce the desired variable density effect but also the camera works as normal i.e. auto-focus is OK & auto-exposure is OK as the the aperture, ISO & shutter speed changes as the filter is lightened & darkened. As has been previously observed by others the changing of aperture, ISO & shutter speed on the 5DII in movie mode varies so e.g. the aperture will open up first then the ISO will rise & the different controls seem to work independently of each other. With my home made set-up it's just the same as Dan B. has noted with the Singh-Ray. Spin the dial until the aperture, ISO & shutter speed are just what you want & then lock the exposure. It's as simple as that. I haven't had time to experiment much just enough to establish that the principle works. In particular I haven't done much shooting with & without filters to make sure that there is no colour cast or any other quality problem. I did see a pronounced blue tinge when the filters were stopped right down to nearly black which may be quite cool for fake night scenes. Even though I live on the Cote d'Azur it was actually pouring down with rain yesterday so my experimentation was limited but today is bright & sunny & we are off doing a shoot so I will be taking all my filters to try them out. I don't know how controllable the Singh-Ray filter is. My DIY filters go from fully open to fully closed in a quarter turn. The rotation also seems pretty loose so after being fixed in one position I can see that it might easily slip but they may just be the old used filters that I bought. So far I am really pleased with my tests & actually buy a Singh-Ray if I can get the same effect & have it reliably controllable.

May 16th, 2009, 01:52 AM	#2
Dan Chung Major Player Join Date: Apr 2008 Location: Beijing Posts: 665	Nigel, The reason to use 2 circular polarizers vs linear ones is to reduce the polarizing effect of the filters which can otherwise cause problems with changing ND value while shooting a scene with reflections, or when you are shooting water and glass. Otherwise the effect will be similar. The other thing to watch for with home brew variable ND is that you can get some pretty bad colour shifts with some polarizing filters. I now have various options for ND, the best for control is 2 4x4 square polarizers in a Matte box allowing to remove the ND completely in a hurry. It is not as portable as a simple screw in Singh-ray however. Dan Last edited by Dan Chung; May 16th, 2009 at 06:59 PM.

May 22nd, 2009, 10:02 AM	#7
Christopher Lovenguth Regular Crew Join Date: Mar 2009 Location: Brooklyn, NY Posts: 85	Diy vd I picked up a linear PL yesterday and just finished a quick test to post. DIY Vari ND filter on 5D II on Vimeo Using 50mm set at F1.8 using lens twist. First filter on lens is Hoya Circular PL size 77 (using step-up ring). I set the circular PL where it's not affecting any glare. I then taped this filter so it couldn't move. I then screwed in front of circular PL a Hoya PL(linear). I locked exposure at 40 and iso 100 finding the right exposure by twisting the linear filter. Without the filters the exposure was 2000 at iso 100 with F1.8. Just used auto WB for this quick test. During the recording I rotated the linear filter to change the exposure. I'm shooting through dusty warehouse windows except for the upper right corner with the tree, this area has no glass. I had no issues with 5D II finding exposure with the linear filter on lens in live view. I'm using a 77 size filter only because that is the largest size lens I have in my arsenal and will be using step up rings for all other size lenses. I didn't notice much color shifting with the Hoya's. It's strange however, if you place the circular in front of the linear all that happens is a strange color shifting and very little density change. I don't have two circular or two linear filters to test these options. The blue color cast when the image is almost black is most likely due to the blue/UV area of the spectrum of light is the brightest and most sensitive to human eyes (that's why blue light alarm clocks are so bright when really dim as opposed to red light).

May 16th, 2009, 09:49 AM	#3
Dan Brockett Inner Circle Join Date: Feb 2006 Location: Los Angeles, California Posts: 2,109	Hi Nigel: I have read about others doing the same thing, and from what I have read, it works decently. But there is some voodoo at work with the DIY solutions that will cause random color shifts, which for me, are a PITA to try to fix. That is what impressed me so far with the Singh Ray footage, there doesn't seem to be any color shift as I rotate the filter. I am sure that there is some sort of optical engineering going on with the Singh Ray or they wouldn't be getting so much money for it. On another note, I am going to be staying in Mougins in your neck of the woods in late June, I would welcome any suggestions from you about must see/must do experiences. I am shooting a doc, following a subject around the South of France and any visually beautiful location recommendations would be welcomed. Thanks so much, Dan

May 16th, 2009, 07:01 PM	#4
Dan Chung Major Player Join Date: Apr 2008 Location: Beijing Posts: 665	Dan, Out of interest which vari-ND did you go for and what the widest lens you can use it with without cut off in the corners? I might be asking you to buy me one and stick it in the post to China! Dan

May 23rd, 2009, 07:29 AM	#8
Robert Esmonde Regular Crew Join Date: Nov 2006 Location: Dublin, Ireland Posts: 51	Thanks Christopher for that really interesting clip and info. Can you say how you mounted the linear filter in front of the circular CPL one? Were you able to just screw it into the CPL filter or did you have to reverse the LP filter and use some kind of reversing ring to connect it to the CPL filter?

May 23rd, 2009, 07:02 PM	#9
Christopher Lovenguth Regular Crew Join Date: Mar 2009 Location: Brooklyn, NY Posts: 85	Hi Robert, I just screwed the linear on the front of the circular.

May 24th, 2009, 02:57 AM	#10
Robert Esmonde Regular Crew Join Date: Nov 2006 Location: Dublin, Ireland Posts: 51	Thanks Christopher. I understood that the front LP filter needed to be reversed, so it's great news to hear that it works fine just screwed into the CPL one.

May 24th, 2009, 04:07 AM	#11
Nigel Barker Inner Circle Join Date: Feb 2009 Location: Norwich, Norfolk, UK Posts: 3,531	I just double checked on my set-up. Like Christopher's it only works properly when the linear polariser is screwed on the front of the circular polariser.

May 24th, 2009, 04:35 AM	#12
Robert Esmonde Regular Crew Join Date: Nov 2006 Location: Dublin, Ireland Posts: 51	Thanks Nigel. I was not referring to the stacking order of the filters. I had read somewhere else (maybe this thread?) that when stacked with the CPL nearest to the lens, the outer Linear Polariser filter had to be turned backwards and needed a reversing ring to connect to the CPL filter. From what Christopher has said, it seems that just screwing the Linear Polariser into the CPL filter works fine, and there's no need for the complexity and extra depth of a reversing ring.

May 24th, 2009, 06:07 AM	#13
Nigel Barker Inner Circle Join Date: Feb 2009 Location: Norwich, Norfolk, UK Posts: 3,531	Robert, sorry my response must have been confusing. I was trying to confirm that a reversing ring is not needed. I too had read that was necessary but it appears not to be the case. I also read that the reason that we have circular polarisers nowadays instead of the older linear polarisers is that auto-focus on modern cameras does not work properly with polarised light so a circular polariser actually has an extra layer that effectively scrambles up the light so that it is not polarised when it enters the camera lens. The 5DII auto-focus seems to work OK whatever type of polariser or combination of polarisers I have on the lens.

May 24th, 2009, 06:41 AM	#14
Robert Esmonde Regular Crew Join Date: Nov 2006 Location: Dublin, Ireland Posts: 51	Thanks Nigel. That clears things up for me. Thanks too for your detailed opening post. Really useful.

May 24th, 2009, 11:45 AM	#15
Christopher Lovenguth Regular Crew Join Date: Mar 2009 Location: Brooklyn, NY Posts: 85	I would guess the only "benefit" to stacking the glass as close as possible to each other with using a reversing ring would be to to get the glass without any gaps so there is less likely any possibility for issues of distortion, dust, etc. But in adding the reversing ring, it's most likely the same gap. I'm such a purest that I hate putting any filters on my lenses but ND is a must for shooting outside with a low depth of field and can't afford $400 for the Singh-Ray when I still need a field audio recorder, etc. I already had the circular PL from the original 5D kit I bought years ago off someone and never used and it's pristine. So my vari-ND cost me about $60 with all the step-up rings and the 77mm Hoya Linear I bought new from Adorama. I am in no way dismissing what is most likely a great product from Singh-Ray and with there not being a huge market for something like this I can understand the price. But if you work with the limitations of DIY it can be done for less then half the cost by picking up a CPL and PL or maybe even just two PL's.