Comparing sensors. How the "native" color of infrared varies depending on the camera.

[Fedia]

As you know, all cameras react differently to IR light once they are converted to full spectrum. I don't know if this is really well understood so I wanted to start a topic and have your opinion about it.

I just got the Sony f828 (the camera that can be converted to full spectrum with just a magnet) and I can now compare its fullspectrum colors to my Canon.

 

 

Canon 1200D, white balanced full spectrum colors, no additional filters :

 

Sony f828, white balanced full spectrum colors, no additional filter :

 

The pictures were white balanced and saturated from the RAW files in Lightroom.

 

The two cameras are pretty far apart as we can see from these pictures. They are also pretty far apart in terms of technology : the first is an 18Mpx CMOS and the second is a 20 years old 8Mpx CCD one of a kind RGBE sensor.

 

Sony camera are known to not perform as good as Canon with the IRchrome filter. The images above indeed show that this Canon camera has a predisposition to record IR in the red channel compared to the Sony. Now do more recent Sony cameras produce blue SOOC like this one does ? I don't know... Maybe the sony users here can help me from their experience.

 

 

[colinbm]

The WB should be critical here.
I think a custom WB on the same target should be used here, whether it is a grey card or a white PTFE or black asphalt.

[Fandyus]

That's interesting. My Sony a6000 looks very similar to the Canon, almost identical in fact. Kind of wish it looked like the other Sony. I really need to get more full spectrum cameras with different sensor technologies to be able to experiment more.

[Fedia]

1 hour ago, colinbm said:

The WB should be critical here.

Prim Holstein cows were the target for both image. From experience I can say they are trusty target.

I understand the a normalized target would formalise the experiment but it is impossible that the color difference observed here is due to an imperfect white balance.

[Fedia]

1 hour ago, Fandyus said:

My Sony a6000 looks very similar to the Canon, almost identical in fact

Really ? Interesting, so maybe the brand is not really a good indicator of what kind of color you get.

 

So is it the RGBE color array that gives this blue color to the full spectrum image ? It wouldn't be surprising if it changed something.

 

What about the fact that this particular Sony is a CCD. Would it affect the colors ?

[Fandyus]

27 minutes ago, Fedia said:

Really ? Interesting, so maybe the brand is not really a good indicator of what kind of color you get.

 

So is it the RGBE color array that gives this blue color to the full spectrum image ? It wouldn't be surprising if it changed something.

 

What about the fact that this particular Sony is a CCD. Would it affect the colors ?

I believe the response curve of a CCD photocell is very similar to that of a CMOS photocell so most likely not. It is definitely due to the RGBE color array. But there probably are subtle differences between full spectrum Canon and Sony cameras, since both of those companies make their own sensors and presumably use slightly different manufacturing methods to achieve the desired transmission for their RGB filters. But having owned both a full spectrum Canon and a full spctrum Sony, I have not noticed much difference between the way colors rendered on either. Mind you I only owned those two together for a short time until I sold the Canon, and I never did side by side comparisons, so I'm not the best authority to speak on this.

[lukaszgryglicki]

So Nikon and Fuji are also very similar in recording full-spectrum photos...

 

[Andrea B.]

To compare how two different cameras actually record IR full spectrum sunlight, it is necessary to look at the raw files before white balance is applied. White balance of alternate light (UV or IR) can vary depending on what converter is used. 

 

That said, we know, as you mentioned,  that the Sony F828 has an unusual Bayer array of Red, Green, Blue and "Emerald". So it likely does record full spec somewhat differently than the Canon.

[Fedia]

3 hours ago, Fandyus said:

I believe the response curve of a CCD photocell is very similar to that of a CMOS photocell so most likely not

I believe the difference beetween CMOS and CCD is in the way the electrons are collected. Apart from that they do the same thing and they have the same color RGB filter on their photocells. So I don't see either why this would influence de color rendition. That being said, there is a hype on the internet right now about CCD sensors supposedly producing an image that looks more like film. While there is no reason to believe it, I give it the benefit of the doubt.

 

Christoph on this forum also repeatedly said that CCD sensors were better for the type of colorfull full spectrum images I enjoy.

[Stefano]

Another potentially interesting thing is how UV false colors look with that camera. I bet they look unusual.

[Fedia]

1 hour ago, Andrea B. said:

To compare how two different cameras actually record IR full spectrum sunlight, it is necessary to look at the raw files before white balance is applied

That would be intresting to do. I can upload the RAW files if you want to give them a scan in Raw digger.

[Fedia]

21 minutes ago, Stefano said:

Another potentially interesting thing is how UV false colors look with that camera

I will try, I have the filters for it.

[Fandyus]

19 hours ago, Fedia said:

I believe the difference beetween CMOS and CCD is in the way the electrons are collected. Apart from that they do the same thing and they have the same color RGB filter on their photocells. So I don't see either why this would influence de color rendition. That being said, there is a hype on the internet right now about CCD sensors supposedly producing an image that looks more like film. While there is no reason to believe it, I give it the benefit of the doubt.

 

Christoph on this forum also repeatedly said that CCD sensors were better for the type of colorfull full spectrum images I enjoy.

I think it's a combination of two factors. My first "weird" camera was a CCD canon powershot converted to 850nm. It had a pretty pleasing noise character, looked somewhat less abrasive. Another factor is that CCD is an old technology, it was used when the RGB filters were relatively narrowband, leading to better color rendition at the cost of noise performance. Nowdays manufacturers have loosened up the transmission of RGB filters on the sensor and it improved the noise performance but a lot of new cameras end up producing colors that look somewhat sterile and boring, especially unedited.

Edit: here's an example of CCD noise. Demosaicing was bypassed. I did not resize the image so that one could zoom in to get the idea.

[Fedia]

5 minutes ago, Fandyus said:

I think it's a combination of two factors. My first "weird" camera was a CCD canon powershot converted to 850nm. It had a pretty pleasing noise character, looked somewhat less abrasive. Another factor is that CCD is an old technology, it was used when the RGB filters were relatively narrowband, leading to better color rendition at the cost of noise performance. Nowdays manufacturers have loosened up the transmission of RGB filters on the sensor and it improved the noise performance but a lot of new cameras end up producing colors that look somewhat sterile and boring, especially unedited.

That's precious information thank you ! Where did you learn that ?

[VideoJohn]

10 hours ago, Fandyus said:

But having owned both a full spectrum Canon and a full spctrum Sony, I have not noticed much difference between the way colors rendered on either.

 

I know this is not exactly the fairest of comparisons, but between my F828 and EOSM full spectrum, the F828 makes blues in ways I haven't seen my Canon do. I'm learning on the cause being that RGBE sensor as the F828 blues seems to have this "emerald" tone to them that my Canon doesn't. The Canon feels more like a true blue. Did your Sony's blue seem to have any other color influence in it?

 

I am speaking way out of my depths here, simply parroting information I've read, but there's a similar sorta debate with certain Hasselblad backs regarding file sizes and the technology of the times, that sounds a lot like your explanation bout why these cameras produce something different from their more modern predecessors.

[Fedia]

18 hours ago, Fandyus said:

It is definitely due to the RGBE color array

I'm using the "Dirty white" gel + DB850 combo on the f828 right now, which a combo @Christoph from Hidden Realms uses for his "blue trees" technique. The result I have look very much the same he has on old Nikons DSLR with CCD sensors which have regular RGB arrays. So it is not proved yet that the RGBE color array is responsible for particular colors of this camera.

[Fedia]

So this comes out soooo well. The filters are the combo mentioned above.

 

 

 

 

 

[FL_96]

Nice rendering ! Interesting to see how the blue foliage seems natural on your images.

[Christoph]

Fantastic, Fedia! Great find with the RGBE sensor camera. The colors look very cool, especially in the last two with the DB850 combo. 

 

From my experience with various cameras, primarily Nikon, I can say that the full spectrum look is most of the time very similar to your first image in the first post and lies somewhere between pale yellow/orange and pale violet. The big exception is of course the Foveon sensor.

[Fandyus]

18 hours ago, Fedia said:

That's precious information thank you ! Where did you learn that ?

Well, shamefully enough, I don't actually have any hard proof of this. But I've seen many photo nerds claim this and it makes sense to me. Old cameras tended to be more saturated right away but more noise. Cameras today have very flat rendering and very low noise levels.

[Fandyus]

@Fedia I must say, this camera you have fascinates me. I think I will buy one myself. How did you get it converted to full spectrum? I already checked the deals on local online second hand stores.

[Fedia]

40 minutes ago, Fandyus said:

How did you get it converted to full spectrum?

Oh I think you missed a topic from two weeks ago. Let me put on a link for you :

[Fandyus]

Thanks, @Fedia

[Fedia]

20 hours ago, Fandyus said:

Edit: here's an example of CCD noise

I think this is great even though the precise noise patern is altered due to Jpeg compression ! It looks like when you add film grain in Lightroom (which I did on the pictures above).

 

1 hour ago, Fandyus said:

But I've seen many photo nerds claim this and it makes sense to me

I trust the nerds online.

[Fandyus]

Update: I am now also getting this camera. I have already negotiated with a seller. Unfortunately, there are some hiccups shipping wise, but I hope all goes well.

I didn't know I'd end up buying another camera this soon, but this one is just too cool not to get, especially since it can be had for so cheap.