Unanswerable Question #273

[Andrea B.]

Why do so many filter makers fail to provide detailed transmittance charts?

 

Transmission is always a thing we UVP-ers want to know.

 

By detailed, I mean this: a logarithmic chart which clearly indicates the blocking OD down to about OD 6.

[Bill De Jager]

If wishes were horses... sigh

 

The overwhelming majority of photographers will do fine with a linear chart.  We in the UV field are an exception because we're trying to photograph only a minuscule fraction of the ambient light, with a very high requirement for spectral purity.  I get the impression (correct me if I'm wrong) that the peculiarities of Bayer-induced UV 'color' make our color UV photos unusually sensitive to contamination by small amounts of unwanted wavelengths.  

 

I just looked at the Edmund Optics site, and their shortpass 400nm OD 4 filter has a transmittance chart that's....wait for it.... linear! I'm surprised, because their customers tend to be high-tech industrial or scientific and many will have very exacting requirements.  On the other hand, they do list the OD performance.  The Schneider-Kreuznach site doesn't even list OD for its UV/IR cut, shortpass, and bandpass filters. They do claim "high OD" with no number attached.

 

I wonder whether this behavior is because a linear graph shows what looks like better performance. OTOH, for industrial and scientific uses you'd expect customers to be savvy enough to not be swayed by such a graph if they have more exacting requirements.

[lukaszgryglicki]

It is just to avoid claims IMHO. They only provide minimum, so you can't say later that they promised OD4 or anything.

 

[dabateman]

Its a cost issue. 

You would need to contact Hoya or Schott, buy a batch of glass. Get a person to run a couple of samples from that batch and then provide you with the data. Or if lucky the exact data for your exact section of glass.

As Jonathan found out the minimum specification allows for ug5 to equal ug11.

So clearly sample to sample will have totally different spectra to OD6.

 

Its too bad I missed unanswerable question 272:

Why did I walk into the kitchen again? 

[JMC]

On 12/10/2021 at 4:24 PM, Andrea B. said:

Why do so many filter makers fail to provide detailed transmittance charts?

 

Transmission is always a thing we UVP-ers want to know.

 

By detailed, I mean this: a logarithmic chart which clearly indicates the blocking OD down to about OD 6.

Various reasons. Variability - I've seen far too many filters which weren't even close to the linear charts the supplier provided. Linear charts can hide a multitude of sins, that would be far too obvious in a logarithmic chart. Batch to batch variability of the glass and any coatings can be a big contributor to this. Also even the collimation of the light (and off axis light) can influence the measured transmission especially for coated filters. The OD issue - it's easy to say 'OD4 between x and y nm' however this can be an average across that range and not a best case value. Some people use averages and others use it as an absolute. Again, easy to hide the difference between OD3 and OD4 in a linear chart and not so easy in a logarithmic chart. Lack of hardware - not everyone has access to OD6 testing capabilities, although it really would be handy for people selling filters especially coated ones.

 

The coated bandpass filters I have from Edmund and Thorlabs are a good example of the potential issues. Yes they have good blocking as long as the light is hitting them at 90 degrees. Any off axis light however can become a big issue as the blocking coatings become less effective at angles other than 90 degrees.

[Andy Perrin]

Quote

The coated bandpass filters I have from Edmund and Thorlabs are a good example of the potential issues. Yes they have good blocking as long as the light is hitting them at 90 degrees. Any off axis light however can become a big issue as the blocking coatings become less effective at angles other than 90 degrees.

Strictly speaking, this is all correct, and they do show the potential issues. However, for people reading this, before being too hard on those companies (especially on Thorlabs), I want to point out some things.

 

1) The filters are mainly meant for laser optics, where the beam really is coming in perpendicular. If we use them for imaging, it's an "off the label" use case, and it's hard to blame the companies when we are doing something they did not intend.

 

2) Thorlabs sells two lines of filters, standard and premium, with different tolerances on blocking.

 

3) Thorlabs actually DOES give full OD data and even downloadable excel files with the transmission and blocking:

 

So if anything that company should be singled out for exceptional service. I wish Edmund did the same!

 

[JMC]

True Andy, and I'll edit my post when I am next on the computer. My experience with Thorlabs has certainly been better than Edmund. However both do have issues with off axis light. Not a problem for laser work, but does need consideration when used for imaging work.

[ulf]

7 minutes ago, JMC said:

However both do have issues with off axis light. Not a problem for laser work, but does need consideration when used for imaging work.

That is the case for all dichroic filters. The Baader U is designed for almost parallel light beams, too to be used with telescopes.

When using it on long lenses like the UV-Nikkor, that is not an issue, as the FOV is narrow enough.

[Bill De Jager]

When I started using the Canon EF 40mm f/2.8 STM lens for UV a few years ago, I forgot about the issue with dichroic filters.  A 40mm lens should have about a 48-degree horizontal field of view on a full-frame camera, or 24 degrees off-axis. (The diagonal field of view is larger, but in most UV photos limited depth of field will blur one or both opposing corners.)

 

I just went back and looked at some UV photos I took with that lens and a Baader-U for my big UV landscape effort in 2019.  It seems that, outside of depth of field issues in many cases, the edges of photos did fairly well in monochrome photography, perhaps with a bit of sharpness degradation at the pixel level that could have also been due to the lens being wide open in some cases.  The appearance of landscapes and vegetation didn't change across the images, suggesting a lack of severe contamination with unwanted wavelengths towards the edges.  Apertures ranged from f/2.8 to f/5.6.  I must stress that this was still relatively casual photography without strict technical requirements.

 

Unfortunately, for some reason I had the camera set to JPEG-only and didn't notice that until I got home, so I can't investigate those photos further.  It's entirely possible that properly processed color images would have shown shifts towards the edges.

 

Nevertheless, I'm going to broaden my UV filter selection.  I'm looking forward to trying out the Nikon 24mm f/2.8 AIS in the field.

[Andrea B.]

Everyone knows that lenses have light fall-off from center to edge. But I haven't seen much discussion of the fact that fall-off alone can sometimes cause edge discoloration under non-dichroic UV-pass filters or non-dichroic IR-pass filters. Of course, having mentioned this, it would have been useful for me to post an example. I know I have some somewhere. My recollection is that this kind of edge discoloration is less noticeable than dichroic discoloration.

 

*****

 

Andy makes an excellent point above that we employ these filters outside their intended use. Worth keeping in mind. 

(Actually we are employing other gear items outside their intended use: converted cameras, non-UV-dedicated lenses.)

 

*****

 

Good to know that about Thorlabs supplying full transmission data.

 

[Andy Perrin]

Quote

But I haven't seen much discussion of the fact that fall-off alone can sometimes cause edge discoloration under non-dichroic UV-pass filters or non-dichroic IR-pass filters. 

Andrea, are you sure the fall-off is causing the discoloration or is this just correlation without causation? I say that because I can't imagine how the light intensity could change the color, but when you have a beam hitting the filter at a steeper angle, it would pass through a larger thickness of filter glass than if it hits perpendicular, and that could easily cause such an effect.

[ulf]

9 hours ago, Andrea B. said:

Everyone knows that lenses have light fall-off from center to edge. But I haven't seen much discussion of the fact that fall-off alone can sometimes cause edge discoloration under non-dichroic UV-pass filters or non-dichroic IR-pass filters. Of course, having mentioned this, it would have been useful for me to post an example. I know I have some somewhere. My recollection is that this kind of edge discoloration is less noticeable than dichroic discoloration.

 

 

I wrote this below before reading Andy's post above.

 

There is a logical explanation for that.

The light beams passing at an angle, pass the filter glass in a longer distance than the beams for the center of the image.

The longer distance is the equivalent of a thicker filter.

A thicker filter has a different shape of the transmission curve with slightly shifting cutoff wavelengths, not just an overall lower transmission.

However that effect is not as strong as the wavelength shift in a dichroic filter where also a high OD can collapse.

Another potential problem with dichroic filters is the shiny surfaces that can cause different reflection problems, together with some lenses and even the sensor's surface.

[Nate]

On 12/11/2021 at 11:25 AM, Andy Perrin said:

Thorlabs sells two lines of filters, standard and premium, with different tolerances on blocking.

Andy, have you or anyone else bought one of these for imaging? I see that the FBH343-10 wouldn't work so well with a 365 torch, but just curious on a bright sunny day.

[Andy Perrin]

Nate, I don’t own ThorLabs’ UV filters but I have many of their SWIR (that’s shortwave IR) filters between 1100 and 1600nm for my TriWave. The issues with OD are similar on the other end of the spectrum, near the end of the camera’s range (~1600nm for the germanium sensor). However I’ve mostly used them with longer effective focal length lenses, not wide angles, because the TriWave has a crop factor of 5.6, so nearly everything behaves like a long focal length. Anyway, I’ve been extremely happy with them with respect to blocking — they have performed as described. 
 

My current avatar is a sunflower between 1500-1600nm shot with a Thorlabs Premium Longpass. 

[JMC]

3 hours ago, Nate said:

Andy, have you or anyone else bought one of these for imaging? I see that the FBH343-10 wouldn't work so well with a 365 torch, but just curious on a bright sunny day.

Nate, I have some of their Standard ones (340-390nm). I use them for imaging with a xenon lamp which has a fairly narrow output, mainly in the 280 to 440nm range, and I use them in a dark room with no other light. They work ok for that, but I haven't tried them in sunlight where there are a lot of unwanted wavelengths to deal with.

 

The filters I have are only coated on one side and some of them show the degradation on the uncoated portion that is often seen with Schott UG glass. So they need cleaning to remove that. I do this every few months and store them with desiccant.

 

I've not tried their Premium ones.

[Nate]

Thanks @Andy Perrin and @JMC I might get one to play around with.

[Andy Perrin]

Nate, you should also look at Omega’s filters too. They also show OD. 
https://www.omegafilters.com/product/4399

[Nate]

Thanks Andy