Magnet + Sony F-828 = Instant Cheap Full Spectrum Fun!! (Images from VideoJohn)

[Andrea B.]

This is totally cool for anybody wanting to inexpensively experiment with IR. The old F-828 (approx $150?) plus a $10 magnet gives you full-spectrum for only about $160 + IR-pass filter cost.

 

Before trying this, do make sure you can still find the batteries for this 20 year-old 2003 camera!

 

The video shows results only for IR photos. I have no idea whether the fixed lens would pass enough UV for good results. 

 

 

[Andrea B.]

Another magnet hack on a different cam -- the Sony DSC-V1.

 https://kiwiinkyushu.wordpress.com/2016/04/13/sony-dsc-v1-infrared-magnet-hack/

 

Now I am wondering how I would go about searching for other cams which have moveable internal UVIR-filtration which can be moved out of the way with a magnet? Some general google queries have not yielded anything useable so far.

 

Note that there are other videos about the magnet hack for the Sony F-828 above -- and also for the Sony F-717.

[VideoJohn]

LOL. I was just about to post my adventures with this camera. Hope it's ok to add to your findings and show off some real world results in 2023.

To answer some questions - you can easily find batteries. I bought a new one in preparation for this camera. Turns out, it came with one, and it still works! Battery life is seriously amazing on this thing. I took it out walking today for over an hour, in temps over 90, and kept it on most of the time. I still have 216 mins of battery life left.

I believe there are 3 models of Sony cameras - f828, 717, and 517, that all have the magnet option, as well as the DSC-V1. I am unaware of any others. :( I might be mistaken, but I believe the 717 caused a minor stir related to seeing people xray nude or something and it kinda pushed the easy access to IR on consumers out the window. At least the warning that comes with that one professional full-spectrum camera would lead me to think as such.

From what I've see online....these are not the best of cameras for this this stuff. I've found my results when trying to push color out of the Hoya R72 to be disappointing. The F828 taps out at the highest megapixel count of 8MB and it's the only one that shoots RAW. So the quality we're used to now a days is not there. Even on 64iso in bright sunlight, it's a noisy image.

Where I have found the camera to shine is as a simple full spectrum camera that I can muck around in post with. The LCD monitor is tiny, focus points are limited (edit: turns out there is a very nice point control focus you can use with the joystick. It's zippy and feels too fragile to not fear you'll break it every time you use it.), memory card situations wonky, the viewer is darn near microscopic, and scopes might be off in full spectrum recording, but gosh darn it, this is a fun camera to use, and even though it's only an 8MB, the images can be very pleasing to look at, at least I think so.

 

Something I'm finding interesting about the F828 is the one of a kind RGBE sensor. I think it's part of the reason there's an underlying yellowish-green pushing through in a lot of my pictures, but that's pure speculation on my part.

If someone wants to know where to snag one for under $100....I'm gonna be a nice guy and tell ya where I know where one is sitting online. Can't guarantee it, though.

[nfoto]

Noise is an aspect that might fade in front of having so much fun ....

 

One of the best, if not the best, IR camera of mine throughout the years is a modified Nikon D40X with internal IR bandpass filter (R72). Even better -- and much more noisy and low resolution -- was the D1. I had a stack of them bought for less than 20$ and ripped out the hot mirror. So infinity was a hit and go, but for some lenses it still did focus well enough.  It might still be floating around hidden deep into a cupboard somewhere, if I can revive a battery or three perhaps a project for a bad day? (it had phenomenal battery appetite especially in cold weather).

[Andrea B.]

John, it's nice to have some local results here on UVP from the old F-828. Thank you for posting them!

 

I'm not surprised by the noise. In 2003 sensor quality was much lower in most cameras, especially in point-n-shoots. But it seems good enough for inexpensive experimentation. And we have better software now which can mitigate noise. The first sky image looks fairly noise free, but I certainly do see the noise problem in other images.

 

Would you please upload a raw file from this camera so that I could play around with it? I would like to run it through Raw Digger and see what actually was recorded. Thanks! 

Note that An uploaded raw file will not appear as an image, only as a downloadable file.

 

On a side note:  Adobe Camera Raw has never worked too well for UV or IR images. I see some of the Adobe characteristic patterning in the sky in the last image. 

 

Also, I'm not sure where all that color is coming from given that you used a Hoya R72 IR-pass filter? It might be because the F-828 had a CCD type sensor? Another reason why I would love to experiment with a raw file when you have a moment to upload one.

[Fedia]

I really like that youtube channel snappines, very soothing and interesting. This video was among the first I watched.

 

19 hours ago, VideoJohn said:

the F828 is the one of a kind RGBE sensor. I think it's part of the reason there's an underlying yellowish-green pushing through in a lot of my pictures

 

That was what I asked myself immediately. How would it perform in full spectrum given this particularity of the sensor. The E from RGBE stands for Emerald. Instead of having two green photosite for each blue and and red photosite like on a bayer array, it has an emerald photosite in place of one of the greens. Thats why they called it this way :

 

 

I wonder how does the emerald photosite output its data, does it go 40% to blue channel and 60% to the green ?  It could be anything. I think the camera still produces an RGB output (I remember seeing a video implying this).

 

I'm trying digital aerochrome at the moment, and isolating IR in the blue channel could be difficult with this kind of sensor since some green light could leek in through the emerald channel.

 

I would be interested to know more about this yellow-green push you are talking about.

[VideoJohn]

On 7/14/2023 at 11:22 AM, Andrea B. said:

On a side note:  Adobe Camera Raw has never worked too well for UV or IR images. I see some of the Adobe characteristic patterning in the sky in the last image. 

 

Also, I'm not sure where all that color is coming from given that you used a Hoya R72 IR-pass filter? It might be because the F-828 had a CCD type sensor? Another reason why I would love to experiment with a raw file when you have a moment to upload one.

 

 

 

 

 

 

 

What do you recommend instead of ACR? I would love better results.

So what I'm referring to with color is what I'm used to from the GFX and R72 is pushing the image and getting out some wild colors. I couldn't get decent enough results with the same methodology and the f828. I assume it's the sensor quality difference. Really wish I had the funds to convert the GFX. Only the blue bike and tree pic is with the R72. It's blue color is all from adding it in the shadows. The other images are just full spectrum and playing around in ACR. If you want more raw files, let me know.
 

DSC00041.SRF DSC00053.SRF DSC00127.SRF

DSC00151.SRF

[VideoJohn]

On 7/14/2023 at 4:26 PM, Fedia said:

I wonder how does the emerald photosite output its data, does it go 40% to blue channel and 60% to the green ?  It could be anything. I think the camera still produces an RGB output (I remember seeing a video implying this).

 

I'm trying digital aerochrome at the moment, and isolating IR in the blue channel could be difficult with this kind of sensor since some green light could leek in through the emerald channel.

 

I would be interested to know more about this yellow-green push you are talking about.

 

I've posted some raw files so you can push colors around and see for yourself. If you look at the bird on the wire picture and the first picture with the big fluffy cloud, you can see it still kinda creeping in the clouds a tad.

I'm starting to wonder based on what Andrea said that maybe it's just an effect of the software I'm using, or perhaps it's a user error issue. I get something similar on the edges of other cameras from time to time when pushing color around, but in this camera, it seems to show up more. 

[nfoto]

RawTherapee opens these RAW files without any hiccups. However, I must add that image quality apparently isn't great. Lots of detail smearing and noise. I surmise the quality can be "improved" by judiciously applying noise removal software, but for my own I won't bother.

[Fedia]

Very nice colours in the DSC00127.SRF once you white balance on the road an saturate the colours. The foliage is blue. Shooting in full spectrum with no filter on my canons would give reddish plants.

 

This camera probably wouldn't perform very well with the IRchrome but on the other hand would do great if you try to achieve blue trees.

 

Is it common for sony cameras to be able to produce such blues without any filters ?

[VideoJohn]

20 hours ago, nfoto said:

RawTherapee opens these RAW files without any hiccups. However, I must add that image quality apparently isn't great. Lots of detail smearing and noise. I surmise the quality can be "improved" by judiciously applying noise removal software, but for my own I won't bother.

 

I played around in R3P once or twice and wasn't really impressed. Saying the IQ isn't great is not exactly the most ringing of endorsement to change workflows. Still thanks for the suggestion. 

[VideoJohn]

18 hours ago, Fedia said:

Very nice colours in the DSC00127.SRF once you white balance on the road an saturate the colours. The foliage is blue. Shooting in full spectrum with no filter on my canons would give reddish plants.

 

This camera probably wouldn't perform very well with the IRchrome but on the other hand would do great if you try to achieve blue trees.

 

Is it common for sony cameras to be able to produce such blues without any filters ?

 

I'm still really confused about the white balancing on this thing. Maybe it's me, maybe it's tech limitations. I haven't played around deep enough, nor has it really mattered much once I hit post. Still, I'd love to see better colors as a starting point, but from what I recall, changing my WB in camera only seemed to produce purple or red/pinkish when using the pre-built settings. I have to relearn how to do the custom WB, but I suspect it won't make much difference. I seem to recall it was still pinkish the one time I used it on gravel asphalt. For anyone who knows, is this somehow related to the needing to make custom white balances on DNG files because of some tech reason I'm blanking on recalling at the moment?

What blue are you referring to? Much of the blue in those photos is artificially introduced or added. I believe the pink trees and mushrooms blues would be as close to "natural" in any of these photos, but they've still had colors pushed around to some degree.

[Fedia]

3 hours ago, VideoJohn said:

What blue are you referring to?

 

I just took the raw file you uploaded and did general adjustements. It has nice tones.

 

[Andrea B.]

Just now got to your raw images, John. I'll post my results soon!

[Andrea B.]

These SRF files from the Sony F-828 are really strange! It has a 4-color CCD sensor which means that it is recording in Red, Green, Blue, and "Emerald". This causes the file exported from Raw Digger to be inverted. So I had to run it thru PS to present here as the Raw Composite.

 

First, here is the tree photo JPG extracted from the SRF. John used a Hoya R72 IR-pass filter. I have no information about what in-camera settings were used to produce this dark plum color in the JPG.

 

 

Here is the inverted raw composite from Raw Digger. This represents the file as shot before any white balance is applied. There is almost no raw color, only kind of a pale pinkish tone.

 

 

Here is the raw data histogram. You can see why there is almost no color in the raw version. The 4 histograms are almost identical. But look at the right-hand toe of the red histogram. It is just slightly leading the other 3 histograms. 

 

 

Here is a leveled version of the file. It looks quite nice in B&W even though John was wanting to get some color from this file. (Can't do that with the raw composite.) I would probably add some carefully brushed-on local contrast to the tree leaves to finish it.

[Andrea B.]

Continuing the raw analysis of John's SRF files.

 

Here is the sky/clouds photo as a JPG extracted from the raw SRF. (Extraction done in Photo Mechanic.)

 

 

For this file, the raw composite shows more raw color than did the preceding file, particularly in the sky which appears greenish. Remember that raw composites always look a bit strange because no white balance has been applied to the raw data. We are only trying to see what was recorded and not what we might eventually get from the file.

 

 

Here is the raw data in histogram form. The tall spikes there represent the midtones, mostly the sky in this case. You can see that the first green spike leads the others, thus explaining the greenish sky. The blue spike slightly leads the red one which might make that green more of a cyan?

 

 

 

The other two raw files really don't provide much more information that what the two preceding analyses show. This camera does not record a lot of raw color under the Hoya R72. So pushing the color with various tools is the only way to bring it out. Given the amount of noise from this sensor, one would have to be careful with extreme edits. But I do see that there are dramatic, colorful versions possible from the preceding conversions shown by John and Fedia.

[Andrea B.]

I'm going to show the other tree foto in raw form because it shows a bit of raw pink. And I should mention for general information that none of the raw composites have any kind of automatic sharpening or detail enhancement. So they can look a bit soft just out of Raw Digger.

 

Here is the JPG extracted from the raw SRF file.

 

 

 

In the raw composite you can see the red channel has quite a bit more color than in the first file. Those tree blossoms are probably a nice strong, saturated visible color. But I couldn't say whether the red/pink in this photo is high visible red or some beginning IR which lands in the red channel. With a 720 IR you can get either I think.

 

The raw data histogram. The midtone peaks are matching mostly, hence the gray-ish areas in the raw composite. In the brighter areas, the red and green is producing some pale raw yellows and yellow-oranges, and of course that pink.

[Andrea B.]

Here is a version from the preceding raw composite. The processing gives the scene a look which I have in the past labeled "Candyland". I have an art series of such suburban scenes with unusual colors and this kind of a grainy look.

 

 

[Fedia]

1 hour ago, Andrea B. said:

These SRF files from the Sony F-828 are really strange! It has a 4-color CCD sensor which means that it is recording in Red, Green, Blue, and "Emerald". This causes the file exported from Raw Digger to be inverted. So I had to run it thru PS to present here as the Raw Composite.

Did you keep the emerald channel to make the composite images, and if so how did you distribute it in the other channels ?

[lukaszgryglicki]

Really sorry for OT but you just reminded me about one of my "dream CFA" - have a color filter array made of: UV, Vis, NIR and BB/FS.

Or a 9 element matrix UV-C, UV-B, UV-A, R, G, G, B, NIR, FS.

[photoni]

2 hours ago, lukaszgryglicki said:

Really sorry for OT but you just reminded me about one of my "dream CFA" - have a color filter array made of: UV, Vis, NIR and BB/FS.

Or a 9 element matrix UV-C, UV-B, UV-A, R, G, G, B, NIR, FS.

 

Really sorry for OT...

Mono is better...

 

The recent images from the James Webb Space Telescope, he photographed structures about 48,000 light years wide;

the Pandora cluster (also known as Abell 2744) is made up of 50,000 light sources that shine in the near infrared
MIRI instrument (Mid-Infrared Instrument) uses
F770W, F1000W, F1130 and F2100W filters
(770 nm - 1000 nm - 1130 nm - 2100 nm )
assigned to blue - green - green - red channels

ESA has provided a 170 MB file here

 

 

while the James Webb Space Telescope employed several filters F115W, F150W, F200W, F277W, F356W and F444W

for the wavelengths of 1.15 μm, 1.5 μm, 2.0 μm, 2.77 μm, 3, 56 μm and 4.44 μm.

These have been assigned the colors blue, green and red respectively.

[KaJashey]

Really interesting with the emerald photo sites and magenta foliage.

 

Is the slowest exposure limited to 1/60? Some of these old Sony big lens superzooms are video cameras at heart and don't go slower than 1/60.

[Andrea B.]

Did you keep the emerald channel to make the composite images, and if so how did you distribute it in the other channels ?

 

I am not sure how Raw Digger handled the SRF file. When I use Raw Digger on a file from a typical RGGB sensor, I always select the option "average the two Green channels".  But that selection was not available for the SRF files. Clearly a Green channel is not the same as an "Emerald" channel, so averaging them would not work well. I'll go experiment in Raw Digger and try to determine how the SRF file was handled. Be back later.

•  

[Andrea B.]

My take on how Raw Digger handled the SRF files is in this new topic:  LINK

[GaryR]

Didn't know about the magnet hack, back when I had the F828, but NightShot the laser emitter were crazy cool.

Excellent thread on the F828!