Digital Filter Construction Project

[AstroGuy]

Hi all,

I'm currently constructing a digital filter. A digital filter you may ask?

Well I'm getting fed up with swapping from filter to filter, having to remove the eyepiece everytime, paying high prices for expensive filters etc...

I have been experimenting with LCD screens to construct a digital filter which mounts in front of the focal path of most telescopes (in front of the primary lens etc...).

I currently use a 4" Mak (Celestron 4se) and hope to have an 8" Saxon Mak soon up and running on a EQ6 skywatcher mount.

I have found that by dismantling a high resolution LCD screen and removing the backlight, I can create a digital color filter mask. I can turn on desired pixel combinations (through software I have created) and create color windows which allow only programmed wavelengths of visible light to enter the aperature of the telescope.

The system is only in it's infant stages at the moment, but basic accurate colour filtration has beeen accomplished! I don't have to say what a great device/system this could be if it's made available to all amateur astronomers.
This is why I'm putting my findings out there!

I must point out though that this method I have found to filtrate color is not 100% accurate with all LCD panels on the market. High resolution computer monitors work best. The reason being is that some LCD panel inter-pixel distances still have some space in between and allow some of the original light to get through this space, this I have found is not as bad as it seems though, and the benefits still outweigh the downfalls. Basically if some of the original light still gets through, the additional colour filtrated light just sumperimposes and inhances the view. The potential for filtering/enhancing important OIII, H-Alpha and H-Beta lines etc... can be done!!!

I've had ideas of creating a smaller version of the digital filter that fits in an eyepiece. I plan using a small inexpensive keyring LCD photo display gadget. Cost is about $7, LCD size is about 1.5" and basically loading the photos as plain colour, full frame images. By placing the LCD before the eypiece this would work as a basic digital colour filter. If you new the exact colour of OIII wavelength of light, you could create an image file which is just that colour. This would allow mainly OIII light to enter the eyepiece and greatly enhance OIII lines in the view etc...

The eyepiece version could also be used for solar observations using the addition of a Ba'ader White Light Solar filter for visual use. By creating the desired filter image colour such as expensive Calcium K-line filter. Of course emulating complex filters that allow multiple bandpass wavelengths of light to pass through to the eyepiece, would have to be created as individual colour image files that correspond to the colour of the desired wavelength/bandpass you want to allow through for video capture. The captured images would then be combined through software etc... Similar to the way LRGB images are combined to produce the final image, although many channels/bandpass/wavelength captured images may have to be combined rather than only 4.

The colour manipulation and filtering possibilities of such a device are extraordinary compared to fixed treated optical glass versions.

So.......I hope my findings stimulate other budding astro-experimenters out there to start experimenting, and if you would like to discuss any aspect of this project feel free to respond, interact or contribute in any way. If you know of anyone else that has worked on a similar project I would be very interested to get any feedback.

Since I had an idea to do this over 20 years ago, I presume I'm the only one who's experimented with and tried this so far (apart from possibly military, NASA etc....)

I know if this kind of technology fell in the hands of greedy commercialists we would be exploited, so let's try and keep it free for all to enjoy!

[DavidU (Dave)]

A most interesting idea. I will follow this thread and wish you luck.

[AstroGuy]

Originally Posted by DavidU

A most interesting idea. I will follow this thread and wish you luck.

Thanks Dave.

[pjphilli (Peter)]

Hi - Very ingenious Eugenio! - I too look forward to your further reports on this project. Cheers Peter

[mswhin63 (Malcolm)]

I did a simple principle a little while ago by placing coloured filters on the screen to isolate RGB and refocus. This method was not the best but had some merit.

Digital filtering would be much more precise admittedly and would be highly effective low cost filtering method. The only negative could be the sharp roll-off of the colours being filtered and may produce some unknown issues when re-combining (I hope not).

It was pointed out though that colour camera place filters over each pixel so pixel reduction is noticed with digital filtering. I experimented and confirmed this as in most cameras Green uses twice as many pixels than Red and Blue. My results showed very grainy quality of Red and Blue compared to green. I suppose the final result will be no different than taking a single colour exposure anyway so I suspect it is a mute point.

In saying this the issue of focusing Digital filtering would be very good and I feel extremely economical, but mono imaging with filters will produce the sharpest images and more accurate RGB combination, but at his stage is very expensive.

Initially I could really use digital filtering.

I am only typing with one arm so please excuse the the way i have expressed this reply.

[AstroGuy]

Hi Malcom,

Thanks for the input. I agree in regards to mono imaging sensors, they would yield a better result.

When you made reference to the amount of green sensors/pixels in relation to red and blue, did you refer to the Bayer mask on colour CCD's?

Anybody want to know more? click here: http://en.wikipedia.org/wiki/Bayer_filter

Good thing about using a RGB LCD screen as a Digital Filter to slice up the visual spectrum is that LCD RGB matrix is more uniform than a Bayer mask for a CCD sensor. Therefore, you don't have to compensate for extra green visual data.

Anybody want to know more? click here: http://en.wikipedia.org/wiki/TFT_LCD

Malcom, I understand what you mean in regards to the sharp roll-off.

A more gentle sinusoidal roll-off can be accomplished using the LCD Digital filter if you require it. Other waveforms can be accurately approximated such as, square wave, Triangle or even complex waveforms for the Spectral Bandpass Envelope. Another feature is, you can also shape a Spectral Bandpass Envelope that conforms to any given linear visual spectral range. Although the Spectral Bandpass Envelope is not solid, it should be able to give the same sort of results using long exposures when used for CCD imaging. Take note though, the Digital filter does goble up aperature to varying degrees depending on the Digital Spectral Envelope Filter dataset used so longer than usual/desired exposures are most likely required in some cases. That drawback should not be drastically evident in LRGB imaging using an LCD Digital Filter, therefore longer than usual exposures should not be required.

The Bandpass envelope can be programmed as a dataset in software. The LCD Digital Filter window can then be made to respond according to the values in the dataset. Not only hue/colour values but, neutral density values can also be simulated/created. If using specific R, G or B filtering, then 256 different Neutral Density/shade values are possible per R, G or B. This would be good for moon observations especially. The neutral density values in the dataset shape the height, of any slice of the Spectral Bandpass Envelope.

going off with more jargon

The visual light spectrum takes up about 350nm of bandwidth approx.?

Want to know more? click here: http://en.wikipedia.org/wiki/Visual_spectrum

For complex bandpass filtering, the "Bandpass slices" or, if you like, theoretically the smallest, individual segment of the bandpass filter, capable of being created using a 24bit LCD Digital Filter is:

0.00002086162567138671875nm (nanometres).

or

350nm / (2^24bits)

24bit colour resolution for LCD monitors allow 16777216 possible different hues/colours to be filtered through the LCD matrix. 32bit LCD monitor even more!

Keep in mind though, only one slice of the Spectral Bandpass Envelope can be used per frame that the LCD switches at so, in one second you can display at the eyepiece 60 viewable slices of the bandpass filter visual data.

If you have a more advanced LCD screen as The Digital Filter then, perhaps 120 frames per second or, 120 slices of the bandpass can be displayed in 1 second. The better the LCD used, the faster the framerate the more accurate the bandpass filtering can be. Some LCD screens actually don't turn pixels off unless they change colour, rather than turning off and on 60 times per second. These non-flickering LCD screens would be the best to use as an LCD Digital Filter. Soon all LCD will be that good.

I can see Malcom how digital imaging would yield much better results when using complex Spectral Bandpass Filtering and long exposures of this visual data.

I have found that cheap low contrast LCD screens used as LCD Digital Filters give less results although, they can still enhance and contribute to the viewing experience when used for Moon and planetary observations. Believe it or not, I have not tried imaging through a LCD Digital Filter yet My main focus is on the software for the LCD Digital Filter and getting my 8" Saxon Mak up and running for testing it, not to mention other projects on the drawing board. I also have a Canon 550D DSLR on it's way to me which I will definately try out some imaging with.

I really need more aperature for advanced testing of The LCD Digital Filter. From what I have calculated, large aperature optical systems will benefit the most from a LCD Digital Filter.

Anywhooooo, sorry if I tend to rattle off a bit. I get like that sometimes when I have to much coffee

BTW, I've ordered one of those little 1.5" LCD photo frames, and hope to test it out to see if it makes a decent LCD Digital Filter for eyepieces. If it works...could be a good, portable economical solution. I will post on this later when i've done some testing when it arrives and i get the time

Will catch up later, have a good one! and take care everybody!

[AstroGuy]

I was talking with some others in a different thread yesterday. I got some ideas, and then.... while jamming with those other guys I figured that the LCD Digital Filter could also be made to perform as a fairly advanced Digital Spectroscopic Analyser. I will try and incorporate this feature into the software . Thanks guys!

Once I have a fairly stable operational software application ready , I will release the software to anyone who wants to use it (free of course) . I will also try to make a Youtube video on how a LCD screen can be modified to perform as the LCD Digital Filter. Interface to the LCD is by standard VGA cable from the computer.

By the way, The software I'm writing for this project is in Visual Basic 6 and will only run on a windows based computer. MAC people will have to run windows emulation software or convert , sorry.

Take care y'all and good luck searching the heavens!

[Astroman (Andrew Wall)]

This sounds really interesting, glad I decided to stop and have a look at this thread (was actually looking for digital focuser, lol ) Now to try and work out how to put a 19" LCD monitor in my 8" OTA j/k

Will keep looking in from time to time to see how this is progressing.

[bojan]

Hmm..
To my understanding, LCD screen consist of small segments (pixels), which are activated/deactivated and scaled/dimmed to achieve the desired colour saturation and luminance on the area much larger than one segment (pixel).
Also, the filtering characteristics of the LCD pixels are not according to standard, used in stellar colorimetry.
Compared to combination of b/w CCD camera and individual RGB filters (on the rotating wheel) the LCD in the light path (with it's low transmission and pixelisation that will affect the resolution of the image) must be inferior.
Personally, I can't see how this idea could be useful.
Could you give us some more detailed theoretical background? For example, how do you intend to calibrate your measurements?

[tonybarry (Tony)]

Hi Bojan,

The idea has merit (I think). If the filter is placed in the light path where a full-diameter solar filter would be placed, you avoid the pixelation issues.

You should also avoid diffraction issues as long as the filter has the same "colour" across its surface.

The LCD spectroscope idea is interesting to me personally. I am trying to work out in my head how it would work for emission spectroscopy. Might not have enough resolution ... although we do claim "millions of colours" for a display. That ought to be narrow band enough for the most discerning user.

I will think more on this.

Regards,
Tony Barry

[bojan]

Tony,
The problem (apart from diffraction, which can't be avoided - it is the same thing as placing the fishnet in front of of the lens - no matter how thin it is, it WILL create diffraction patterns and destroy the resolution) is also in colour response - LCD screen "generates" colours by appropriate dosage of individual primers (RGB) to create the *perception* of colour in our brains.
This perception has nothing to do with colour index, because the same or very similar perception can be achieved with different primers... so we have no standard colorimetry here.

Transmission is another problem - it is generally very low for most LCD's (bright back light takes care of this.. but in astronomy we have to try to catch every photon available whenever possible).

Spectroscopic analyser is even less feasible - because spectroscope does not deal with colours (which are our physiological response to stimuli and very often the totally different stimulus will result in the same or very similar response)... It deals with wavelengths, which are physical properties of light (EM radiation)
Filters used in LCD are very wide, passing comparatively huge bandwidth, while spectroscope bandwidth resolution is and must be very narrow - couple on nm and less.
Any Wikipedia article on the subject will clarify those points for you.
This idea simply won't work.

So, as colour (astro)photography is concerned, stick to the standard colour sensors with Bayer matrix - the principle is the same as this LCD screen idea, but it has no diffraction problems (because filter is in front of the sensor), transmission is high, it's been done already by both industry and individuals and it's available off the shelf and it is known how well/bad it works.
For any serious and scientifically usable work, the combination of b/w CCD and standard colour wheel is un-avoidable.
For spectroscopy, the diffraction grating and/or prism are the only practical/economical ways (at the moment.. maybe in the future there will be optically active materials available to do the job) to achieve usable/required resolution, while keeping absorption (photon efficiency) at acceptable levels.

[AstroGuy]

Good to see you here Tony! How's life been treating you?
The Saxon Mak is sleeping like a baby at the moment, but should be up and about in a couple of weeks!


Hi Bojan,

Glad you could join us.

Firstly I would like to point out that Manufacturer's are already using LCD's as Neutral Density filters in optics such as cameras and sunglasses. I myself built a neutral density filter over 20 years ago, and the quality was good even back then. Manufacturers are also using LCD in the all to popular 3D glasses for television. LCD does not have that bad light transmission properties, then again don't expect the world from them. A good LCD will have good optical quality, maybe not as good as BK7 optical glass for the same size, but still good. I do admit there is light loss with the LCD Digital Filter. I did mention that larger aperature scopes will benefit the most. And if you are doing RGB seperation then take into account that each R, G and B is approximately 33% of the aperature of the LCD so in effect, only 33% effecient for the amount of LCD aperature window.

My goal at this stage is to create a workable prototype. Resolution, quality defractive limitation can be resolved later. So, with all respect Bojan, unless you have done some research and practical hands on work with LCD's I can't understand fully where you are comming from with your point of view. Don't get me wrong Bojan, I do value your input and have taken into account some interesting ideas/perceptions that you have put forth. I have not yet thought about calibration etc..I just want to test it. No loss except for a little time and very little money, no harm in experimentation. Really it's mainly all in fun in the persuit of scientific study. If something usable develops then great! nothing is certain, if it flops, I'll be the one to take the egg on the face, no problemo.

Got some work so I will catch up later, take care everyone.

[bojan]

Hi Eugenio,
My point of view is very simple, I don't see the benefit in using LCD as a filter, especially in front of the lens (objective). The optical quality is not a small issue here - the only reason why solar filter does not introduce (a lot of) optical distortion is because it is VERY thin... but colour LCD is a multi-layer assembly, very different from neutral filter (with a lot of transparent parallel conductive paths in it), and it is not designed nor manufactured to be used as a filter in front of the objective lens of the telescope (I am guessing the price would be much, much higher in that case). The fact that user does not see a problem with LCD glasses is simply because the magnification is not involved... with the telescope however, it is a different matter. So, your test may be able to prove the concept in terms of colour filtering... but I doubt you will be able to progress with this much further than that. Mind you, if you do, I will try to be among the first to congratulate you

As you pointed out, the LCD filter will take out at least 66% of incoming light (because only one third of the incoming area is available, the rest is blocked with other colour pixels) but the similar happens with Bayers martix as well (the proportion is different as the array is GGRB).
Conventional filter, on the contrary, has (almost) 100% transmission for the central wavelength, no blockage or reduction of aperture. Of course, there is a need for another 2 exposures to get the full spectrum, but the overall result is still 3x better in terms of total photon collection.

So.... my point of view is not to attack you personally but to try to direct you to think realistically, and to point out the obvious problems and issues with the idea as I see them. You have to be able to answer my concerns with compelling counter-arguments and/or repeatable test result, right?

Also, you may wish to consider why this idea is already not being used - maybe there is a good reason for this being the case?
In my school days, I dreamed about using mylar foil and vacuum chamber behind it to produce a cheap parabolic mirror. Or a mirror cast from plastic and aluminised.. I even made a prototypes and they worked (in a sense that they focussed the light in a small area) buy that was all... Much later I found out that many others before and after tried the same or similar, even unbelievably more complicated contraptions were made in attempt to achieve the minimal acceptable optical quality and stability .. and we are still using solid glass mirrors.

[AstroGuy]

High Y'all


It seems that a lot of people doubt that this project will work. Well have a quick read of this:

Spanish techno scientists discovered a way of using a LCD matrix to route light onto a single photodetector to take video/stills with it. They are calling it the single pixel camera. It should revolutionise the CCD market by cutting costs on expensive extremely low noise CCD's, it's easier and cheaper to build 1 pixel than a whole matrix!

http://www.sciencedaily.com/releases...1220094601.htm

Gotta get back to work...

[bojan]

http://physicsbuzz.physicscentral.co...el-camera.html

[AstroGuy]

Originally Posted by bojan

Hi Eugenio,
My point of view is very simple, I don't see the benefit in using LCD as a filter, especially in front of the lens (objective). The optical quality is not a small issue here - the only reason why solar filter does not introduce (a lot of) optical distortion is because it is VERY thin... but colour LCD is a multi-layer assembly, very different from neutral filter (with a lot of transparent wires in it), and it is not designed or manufactured to be used as a filter in front of the objective lens of the telescope (I am guessing the price would be much, much higher in that case). The fact that user does not see a problem with LCD glasses is simply because the magnification is not involved... with the telescope however, it is a different matter. So, your test may be able to prove the concept in terms of colour filtering... but I doubt you will be able to progress with this much further than that. Mind you, if you do, I will try to be among the first to congratulate you

As you pointed out, the LCD filter will take out at least 66% of incoming light (because only one third of the incoming area is available, the rest is blocked with other colour pixels) but the same happens with Bayers martix as well.
Conventional filter, on the contrary, has (almost) 100% transmission for the central wavelength, no blockage or reduction of aperture. Of course, there is a need for another 2 exposures to get the full spectrum, but the overall result is still 3x better in terms of total photon collection.

So.... my point of view is not to attack you but to try to direct you to think realistically, and to point out the obvious problems and issues with the idea. You have to be able to answer my concerns with compelling counter-arguments and/or repeatable test result, right?

Also, you may wish to consider why this idea is already not being used - maybe there is a good reason for this being the case?
In my school days, I dreamed about using mylar foil and vacuum chamber behind it to produce a parabolic mirror. Or a mirror cast from plastic and aluminised.. I even made a prototypes and they worked (in a sense that they focussed the light in a small area) buy that was all... Many others before and after tried the same or similar, even unbelievably more complicated assemblies were made to achieve the minimal acceptable optical quality and stability .. and we are still using solid glass mirrors.

Hi Bojan,


So you don't believe that even if the LCD matrix used has fairly good optical transparency, colour saturation and optical resolution, it would not be useful for solar observations when used with such a filter as the Ba'ader solar film? Particularly when programmed to bandpass good detectable energy levels of the Calcium K Line and Hydrogen Alpha wavelengths of the visible light spectrum? In my opinion, if the LCD Digital Filter could be used for only that, it would be worthwhile just in the $$$$ savings. Then again, that is just my opinion.

I've used Ba'ader solar film before, and there is quite a lot of white light that can be tapped at the eyepiece. And if you eat a lot of carrots/sardines (phosphorous and Vitamin A) and your eyes have high cornea light sensitivity, it's even more pronounced and can actually be uncomfortable.

Bojan, I admit I sound a little convoluted at times, that is because I'm also trying to explain it all in a language that non-technical people can also grasp. It might take you a couple of times reading what I have written, but in many ways I have covered a lot of what you are already pointing out.
I can see you are an extremely intelligent individual, and I hope you will stay with us here for a while and keep sharing your scientific and intuitive insight and knowledge!

I value everyone's input on this project and so do others. So please feel free to contribute as little or as much as you feel comfortable with. We really enjoy the interaction.

Back to work now.....:comput er:

[bojan]

Originally Posted by AstroGuy

.. Particularly when programmed to bandpass good detectable energy levels of the Calcium K Line and Hydrogen Alpha wavelengths of the visible light spectrum?

No, you can't program the LCD filter to pass only one spectral line, being it a Calcium or Hydrogen or whatever else..
For something like that you have to have very narrow band filters (0.7 Angstroms or narrower) - they are on the market of course (they are not even that much expensive, definitely within the financial reach of many amateurs).
And they are "programmable" or , the better expression would be "tunable". This is achieved by changing the temperature of the filter assembly bu couple of degrees Celsius (it is actually the sandwich made of many individual interferometric filters).

Examples are here:
http://www.solarscope.co.uk/
http://www.meade.com/product_pages/c...o/coronado.php

[AstroGuy]

Originally Posted by bojan

http://physicsbuzz.physicscentral.co...el-camera.html

This image of the single pixel camera in my opinion is a fake.

If you bring it into photoshop you can see that the colour channels are not even RGB, instead they are indexed. Complete fake used to discredit the single pixel camera technology in my opinion. Quite dubious And to top it off it shows a current Hi-resolution image (probably taken with modern DSLR) against an image that was taken with the single pixel camera technology from 1996.

Even if this example was accurate, don't forget that in 1996 we still used DSTN (256 colour) LCD not TFT LCD (close to true colour).

That's like comparing the original ford car of the 1900's with a porsche of 2010. Complete fake and bias example of this technology. I think the author of that article had a grievance with the current developers of this technology.



Regards...

[AstroGuy]

Originally Posted by bojan

No, you can't program the LCD filter to pass only one spectral line, being it a Calcium or Hydrogen or whatever else..
For something like that you have to have very narrow band filters (0.7 Angstroms or narrower) - they are on the market of course (they are not even that much expensive, definitely within the financial reach of many amateurs).
And they are "programmable" or , the better expression would be "tunable". This is achieved by changing the temperature of the filter assembly bu couple of degrees Celsius (it is actually the sandwich made of many individual interferometric filters).

Examples are here:
http://www.solarscope.co.uk/
http://www.meade.com/product_pages/c...o/coronado.php

In theory, a TFT LCD has a resolution of 0.00002086162567138671875nm (nanometres).

or

350nm / (2^24bits)

24bit colour resolution for LCD monitors allow 16777216 possible different hues/colours to be filtered through the LCD matrix. 32bit LCD monitor even more!

[bojan]

Originally Posted by AstroGuy

In theory, a TFT LCD has a resolution of 0.00002086162567138671875nm (nanometres).

or

350nm / (2^24bits)

24bit colour resolution for LCD monitors allow 16777216 possible different hues/colours to be filtered through the LCD matrix. 32bit LCD monitor even more!

You are mixing here two totally separate concepts.

350nm is the bandwidth of visible, white light, yes... but all the LCD filter does is, it is subtracting certain amount of band from while light to create for us the PERCEPTION of specific colour.
Number of available colours (or colour resolution) is NOT the same thing as wavelength resolution.


If you want to isolate various single lines from spectrum, you have to have a tuneable, narrow band filter. Period. No LCD filter can do this.
The single spectral lines can't be "synthesised" or made from primers.. their approximate colour maybe (if their colours are inside the colour space, determined by standard primers).. but I can tell you right away that for H-alpha this is not the case - the colour of this light is well outside any standard colour space and it cant be reproduced accurately).
And, once again, this is about colour (perception) we are talking about.
The wavelength of light (spectral line) is totally different thing.

More information on the subject is available here:
http://en.wikipedia.org/wiki/Colorimetry