infrared conversion

[Jeremy Roussak]

I'm thinking of having my old Canon 20D converted for infrared, since I don't use it much now I have a 5D2. I'd value some help: what are the pros and cons of different filters (i've seen mention of 720nm to 830nm as possibilities)? Can anyone recommend a place in the UK which will do it? The only one I've found is Advance Camera Services, in Norwich.

Any other tips gratefully received.

Thanks

Jeremy

[Geoff Wittig]

I'm thinking of having my old Canon 20D converted for infrared, since I don't use it much now I have a 5D2. I'd value some help: what are the pros and cons of different filters (i've seen mention of 720nm to 830nm as possibilities)? Can anyone recommend a place in the UK which will do it? The only one I've found is Advance Camera Services, in Norwich.

Any other tips gratefully received.

Thanks

Jeremy

Can't speak to conversion in U.K.; I had my old Eos-1Ds converted by Lifepixel in Washington State, U.S. They did an excellent job; not a single spot of dust on the sensor when it was returned.

Which filter you want depends on your goals. If you like bizarre color shifts and the 'color infrared' look, go for a filter in the 590 to 665 nm range. You can get a decent BW infrared look from this with some Photoshop post-production work. On the other hand, if you just want the 'deep infrared BW' look, go for the 830 nm conversion, which is what I did. This gives you the traditional frosty white vegetation/deep black skies look.

If you have a redundant camera and can afford the conversion, by all means do it. I find that the IR converted camera gives me something to shoot at high noon on sunny days, when otherwise I'd be drumming my fingers on the table impatiently waiting for the harsh contrast to go away. Even the harshest sunlight/shadow contrast generally falls within the available dynamic range of the sensor with the deep IR conversion.

[ddk]

I'm thinking of having my old Canon 20D converted for infrared, since I don't use it much now I have a 5D2. I'd value some help: what are the pros and cons of different filters (i've seen mention of 720nm to 830nm as possibilities)? Can anyone recommend a place in the UK which will do it? The only one I've found is Advance Camera Services, in Norwich.

Any other tips gratefully received.

Thanks

Jeremy

I'd go with the 720nm conversion, you get the same bw results in post as you'd get with the 830nm without loosing a stop and if you wanted there's still enough color left to process the blue and pink skies.

[Robert DeCandido PhD]

Have a look at the NikonGear forum on IR (and UV):

http://nikongear.com/smf/index.php?board=57.0

Look at their fixed thread on the subject:

http://nikongear.com/smf/index.php?topic=22929.0

Some fine (if not the best) IR shooters post there regularly and they welcome questions, input, photos etc.

Folks generally recommend cameras with CCD sensors rather than CMOS sensors for IR (and conversion).

rdc/nyc

[Yakim Peled]

Folks generally recommend cameras with CCD sensors rather than CMOS sensors for IR (and conversion).

If a camera is converted, does it still matter if it's CCD or CMOS? If so, how is it seen in the pictures?

[t6b9p]

Folks generally recommend cameras with CCD sensors rather than CMOS sensors for IR (and conversion).

This is a little misleading and you have to take it in context. Nikongear UVIR is primarily working with Nikon DSLR models which until recently have only been CCD. The comment in the Nikongear IR section suggests that CMOS (i.e. Nikon) sensors show promise for IR but need to be verified is because not too many people are using IR modded CMOS Nikons yet. On the other hand there are plenty of CMOS Canons modified for IR and they appear to work fine.

The misunderstanding may also originate from the fact that modified CCD sensors tend to have a better UV response than modified CMOS sensors and this seems to be holding true for both Nikon and Canon.

[t6b9p]

720nm to 830nm as possibilities

There are pros and cons to each of these options. If the 720 could clearly perform as well, and the same as, the 830nm then why would people want the option. The 720 leaks Red as well as passing IR whereas the 830 only passes IR, this results in a different spectral response to the subject. Also at 830nm the Bayer dyes become basically transparent to the IR light and therefore the nearly equal tranmission of IR light through the R, G and B dyes results in a raw file in which the R, G and B channels are better balanced.

The other option is for a wide band conversion UV through IR and then use filters on the front. There are pros and cons to this as well however this option doesn't lock you into a specific wavelength range.

I suggest you search the forums specializing in IR, or UV and IR, for more background on these two options before commiting your decision.

[Robert DeCandido PhD]

"The misunderstanding may also originate from the fact that modified CCD sensors tend to have a better UV response than modified CMOS sensors and this seems to be holding true for both Nikon and Canon."

Yes CCD sensors (in the Nikon D200) respond better to (capture more of) IR light than CMOS sensors - nothing mis-leading in that statement. I have seen some fine work done by folks using converted Canons (such as the 5D with its CMOS sensor) - but the folks in the know prefer CCD sensors - that Nikon used in many of their camera bodies of the last several years including the D2X, the D40 and D70. There was a good discussion of this over at FM forums a few months ago - do a search and it can be found.

Misleading and misunderstanding....Oh well - thanks for editorializing. But what I think you mean to say is that CCD sensors have better IR response than CMOS sensors - since the question was about IR and not UV. The two are different. If you want to talk about response to UV light (CCD vs CMOS) then go right ahead - but I thought it best to lead the discussion down the same track as the posed questions.
==================

This is a little misleading and you have to take it in context. Nikongear UVIR is primarily working with Nikon DSLR models which until recently have only been CCD. The comment in the Nikongear IR section suggests that CMOS (i.e. Nikon) sensors show promise for IR but need to be verified is because not too many people are using IR modded CMOS Nikons yet. On the other hand there are plenty of CMOS Canons modified for IR and they appear to work fine.

The misunderstanding may also originate from the fact that modified CCD sensors tend to have a better UV response than modified CMOS sensors and this seems to be holding true for both Nikon and Canon.

[t6b9p]

Robert

Knowing your work with UV, I was suprised to see you make the statement that CCD are better than CMOS for IR work.

But what I think you mean to say is that CCD sensors have better IR response than CMOS sensors

Actually, I said exactly what I meant to say. CCD and CMOS sensors both have inherently good IR response which is why both manufacturers go to great lengths to eliminate it with the internal cut filter. I only mention UV because a lot of people read that CCD sensors are better for UV than CMOS and then think the same applies for IR and it doesn't.

I have been a regular contributor to Nikongear UVIR forum for many years and helped edit the IR sticky that you refer to. The forum is of course biased towards Nikon DSLRs most of which have, up until recently, been CCD. These models have great IR response when modified, so the only current question in that forum is how the new Nikon CMOS sensors will respond to IR when modified. They should respond just as well as any of the Canon CMOS sensors based on their inherent IR sensitivity BUT no one contributing on that forum has a converted Nikon CMOS to provide conclusive proof. It does not state in this sticky that CCD is better than CMOS for IR.

In a more generalized IR forum such as IRphoto.com, IR modified Canon CMOS seem to be more popular than Nikon CCD.

I have owned and used a D70, D40, D200 and S3 Pro UVIR for UV, IR and luminescence photography in scientific applications and I am familiar with the relative sensitivities of the CCD sensor in both the IR and UV regions. I am stuck on CCD sensors (Nikon) because CMOS sensors are not practical for my UV work. However, I would have no problems using a converted CMOS or CCD DSLR for IR work. I am currently considering converting a D700 (CMOS) for IR work if the funds become available.

Robert, I would be interested in the FM link you refer to, I was unable to find it.

[Robert DeCandido PhD]

Well I have done much more work with IR - almost none with UV...Also forgot to mention that much more work has been done testing Nikon lenses for compatibility with IR converted bodies (some lenses produce "hot spots" at all or certain apertures). Very little testing has been done with Canon lenses on Canon bodies....

Here are a couple of links:

http://www.fredmiranda.com/forum/topic/860568/0#8069173

http://www.fredmiranda.com/forum/topic/893223/0#8416283

Here's a nice quote that might help from late January 2010:

From Poster A: "A few years ago I had my 10D converted by MaxMax and liked their work so much that I'm planning to send my 50D to them for the 830nm conversion. If that sensor is unable to produce better IR images than I've gotten from my 10D then I'll want to rethink my plan."

Response from Poster B: "I am making large prints from the files and find the detail to be lacking. Even though I was careful not to stop down more than f8 on the 50mm f1.8 there seems to be some diffraction. A 720nm conversion may work better. Another thing is that the 830nm filter gets the IR into the less sensitive range of the sensor so the exposures are significantly longer. The smaller sensor of the 50D means it needs to be extra steady so shutter bounce even on a strong tripod can be a problem. I went to Utah and Arizona with it and even on bright days I found shutter speeds too slow unless I upped the ISO. That is another problem. The noise in the 50D with the IR filter was unacceptable over ISO 320 in my experience. Even at 200 it started giving sky noise. I had a 30D with a 720 IR filter at one time and I it had no such problems. It produced better images than the 50D + 830. Don't let the megapixel count of the 50D throw you off. The 30D with the 720 filter made at least as detailed if not more detailed prints than the 50D with a 830 filter.

"If I was going to make another IR camera it would be a 5D with a 720 filter. 5D's are super cheap now on ebay. In my opinion maxmax should not continue to offer the 830 on the 50D. It sucked to put it mildly. I lucked out when lifepixel offered the $280 year end conversion special. I also was fortunate to have asked maxmax to send me the IR block/anti aliasing filter with piezo dust remover back to me after the conversion. I sent it to lifepixel to rescue the camera from the 830nm IR filter. They will not send you the filter they took out unless you ask them for it. I had to send the 50D back to maxmax after the conversion since he forgot to adjust the auto focus to IR. When I first got it back it was so far off that even with the micro adjust to the limit it would not come close to hitting auto focus correctly. Back when I sent in the 30D it came back dead on.

"My advice is a full frame such as a 5D - very very cheap at this time. If you need to go lower in cost a 40D will beat the 50D due to lower noise if the ISO needs to go past 320. The hot spots in the lenses is one problem. Diffraction is another. I found the 135L to be the best IR lens as far as resolving power. Of course its a super lens in visible spectrum too so not too big a surprise. And no hotspots. I never stop any lens down past f/8 to f?13 when using IR. Diffraction caused by small apertures in visible light is more pronounced in IR."

From rdc again here: So it may well be that Canon (and Nikon) CMOS sensors are fine for IR photography and virtually as good as CCD sensors. Some testing and comparison needs to be done but as you indicate, few people are converting Canon bodies to IR. It may be that all of us were influenced by Bjorn's thinking that CCD sensors are better (in theory if not practice) - so we all see the world through that lens...

Anyway, good to know that according to the NikonGear "sticky" thread, there is no significant difference between IR results from CMOS vs CCD sensors.

rdc/nyc

[Yakim Peled]

In a more generalized IR forum such as IRphoto.com, IR modified Canon CMOS seem to be more popular than Nikon CCD.

Good to know. Thank you.

[imagico]

Concerning CCD vs. CMOS - it should be noted that live view can be very useful since focussing is a general issue for IR work.

Concerning 720nm vs. 830nm - if you do a smaller wavelength conversion you can always try a larger wavelength by using a screw-in filter (loosing viewfinder, metering etc. of course).  You of course loose some light through multiple filters but as others have said you need fairly long exposure for the far infrared anyway so a tripod is necessary in most cases.

Greetings,

Christoph

[t6b9p]

Just to be clear on a couple of points:

but as you indicate, few people are converting Canon bodies to IR

I indicated that in a general IR forum, Canon seemed to be the dominant IR camera.

Anyway, good to know that according to the NikonGear "sticky" thread, there is no significant difference between IR results from CMOS vs CCD sensors.

It actually states "examples so far look promising" refering to Nikon CMOS, however very few people have converted Nikon CMOS for IR at this time.

Well I have done much more work with IR - almost none with UV

I was looking forward to your UV wildflower and bird project back in 2007 but it sounds like you didn't proceed with that?

[Robert DeCandido PhD]

Dear Klaus,

The UV bird work took a "back seat" when the person heading the ornithology dept at WCS (aka Bronx Zoo) was replaced - and access to the collection became more difficult. I also began working as a consultant to the Wind Power Industry...In addition, I have been working for the government of Thailand doing bird (mostly raptor) migration surveys - this will run through 2011 at least. Finally, Deborah and I are doing the only study in North America on urban nesting kestrels (a small falcon) - and we use our IR (and standard) photography there. Here is a link to info in German about our research:

http://www.battaly.com/nehw/AmericanKestrel/German.pdf

and in English:

http://www.battaly.com/nehw/AmericanKestrel/news/

Anyway, good to know that Canon cameras with their CMOS sensors are likely just as fine for IR photography as Nikon's CCD based cameras (such as the D200) and the newer CMOS ones (D700).

Regards,

rdc/nyc

[t6b9p]

Actually it's Shane not Klaus.

[Ligament]

Bumping this very old thread. Taking a trip to Iceland soon, looking at bringing a D800 or D810 IR converted, need opinions on 720nm vs. 850nm for this situation. Will likely be windy...will extra stop with 850nm be a liability?

[BrianVS]

CMOS sensors have much less sensitivity in IR than CCD's- 720nm vs 850nm: 720nm is going to give a big gain for sensitivity.

[Camboman]

Bumping this very old thread. Taking a trip to Iceland soon, looking at bringing a D800 or D810 IR converted, need opinions on 720nm vs. 850nm for this situation. Will likely be windy...will extra stop with 850nm be a liability?

I had my d600 converted to 850nm through Kolari Vision and I am down 3 stops compared to visible. It definitely necessitates a different mindset for shooting. Not a typo - 3 stops!

Still, I love the bright foliage and dark skies I get with this filter.