Colour temp of LED panel

[pubquiz]

I noticed an LED panel to use for flats mentioned in another thread and it looks good value for 12" x 12" for £15 inc driver

here https://tinyurl.com/ybtasxun

It is available in Natural White 4500k, warm white 3500k and cool white 6500k ...which of these is better for flats or is there any difference?

I currently use a small electroluminescent panel for my Wave 102mm refractor which was much more expensive but plan on getting a C.95/C11 soon so wanted something bigger

 

I recently bought this from ebay https://tinyurl.com/y9ku7v8q

It seems quite good but I haven't actually used it yet and the wife has her eyes on it since it arrived for her art work tracing etc

So if the £15 one would do I could get that and give her the one I have....just wondered if the colour temp made any difference re flats?

 

Tom

[Oddsocks]

Hi Tom.

In answer to your question, does colour temperature make a difference to flats? the answer is yes, especially with OSC sensors where a mismatch between the night skies spectral balance and the flat panel spectral balance can leave ugly and difficult to deal with colour gradients in the individual RGB colour channels of the calibrated image.

There are also differences to the way purpose made astronomical OSC cameras respond to light from LED panels, depending whether the camera has a built in UV/IR blocker filter or not, and with DSLR cameras depending of whether they are unmodified, partially modified, fully modified, using in-camera white balance settings or custom white balance settings and/or used with clip in LP filters etc.

In an ideal world, using a well written image calibration program this should not matter, as the flat frame should be chrominance normalised prior to being used to calibrate the light frame, i.e, each 2x2 bayer pixel group should have the chrominance value separated from the luminance value, then the chrominance value averaged across all four pixels so that overall the flat frame becomes monochrome before being divided into the light frame.

If this were done there would be no remaining per-channel colour gradients in the calibrated light frame other than those due to light pollution and instrument response. It is clear from the number of posters on this and other astro imaging forums regarding colour gradients after flat application affecting OSC and DSLR cameras, particularly in the vignetted field, that this is not the case for many of the astro imaging applications available.

I have a suspicion this can be attributable to whether the calibration program applies flat field calibration to the lights prior to debayering, where chrominance normalisation would be impossible unless the program had LUT's (look-up-tables) for each sensor on the market so that the right amount of compensation would be applied while the flat is still in bayered form, or the program applies the flat field after debayering where chrominance normalisation is easy.

The first option, to calibrate lights with flats prior to debayering is light on computer power and fast to execute but is I suspect inaccurate for OSC images where a significant colour temperature difference exists between light and flat. The second option, to calibrate lights with flats after debayering uses more computer power, has to deal with bigger files, is slower to execute, but more accurate. I should say that this is only my personal opinion, I have no hard evidence to back it up as very little is published by the program writers of astronomical image processing programs to say how their programs handle OSC calibration files. 

Given the unknowns it is best to go for a flats light panel that produces an output spectrum as near to natural daylight as possible so that when acquiring flats the three RGB colour channel histogram peaks of the debayered flats are close together and ideally overlapping.

The white LED's used in these panels emit all their light originally as UV and a phosphor coating inside the LED's lens fluoresces under the UV to produce the white light. Depending on the quality of the phosphors used there can be gaps in the emitted colour spectrum, though not so noticeable with OSC cameras such as yours, it can sometimes be difficult to get good narrow band flats with monochrome cameras and LED panels if the narrow band filters pass-band happens to coincide with a missing bit of the LED's output spectrum.

Although covering a different, but related topic, the linked web document does contain a useful paragraph on spectral response for flat illumination sources and is worth a read: http://diffractionlimited.com/flat-fields-stray-light-amateur-telescopes/

Unfortunately the specs produced for general lighting LED panels such as in the first link you provided for the 300mmx3000mm £15 panel are woefully inadequate to make a judgement, however it should be noted that this type of general lighting panel is way too bright, with an output equivalent to a 100W tungsten lamp. The panel is non-dimmable, which is a good thing since many of the general purpose lighting panels use P.W.M. controllers to reduce the light output which can induce banding in the flats, so it would require several sheets of a neutral density absorber to reduce the light output to a level suitable to take good consistent flats. A suitable neutral density gel filter can be found here: https://www.sblite.co.uk/67-neutral-density a single sheet is big enough to cut three 300mmx300mm individual layers from and you would need a sheet of the densest filter they stock, 1.2, grade. Since the spec quoted for that GP lighting panel gives an output range between 800 and 1600 lumens I have no idea how many layers of ND filter you would need in practice. Hopefully a single sheet of ND gel filter would be sufficient when cut to produce three layers but this may still be inadequate and require as many four layers, or more, so two sheets of gel filter at £7 each, making the inital purchase price for that panel not so attractive.

With GP lighting panels it is a bit of an unknown how even the light distribution across the panel will be, a purpose built art tracing panel 'should' be made to a higher standard with even light distribution right up to the corners of the panel.

The second panel you linked to, the A3 tracing panel, that you already have, has equally inadequate specs to make a judgement. The specs say it is dimmable and flicker free, however it is not stated how they achieve this. If light output is controlled by a linear regulator then the flats won't suffer from banding, if they use a DC to high frequency pulsed DC regulator for brightness control then possible it will, you would need to test it to see. Nor do they state how accurate and stable the output light is when using the dimmer function, leading to possible difficulties achieving consistent flats. It is generally better to use the tracing panel on full brightness to avoid possible banding artefacts or output variability caused by deficiencies of the panel's brightness regulation control and use layers of ND filter between panel and OTA to reduce the panel's transmitted light and produce consistent flats. Finally they don't quote a colour temperature for the tracing panel however if you try it and find the three RGB histogram peaks from your OSC flats are widely separated then you can add a complementary colour correction gel filter in front of the panel: https://www.sblite.co.uk/63-ct to pull back the strongest channel and bring the three RGB histogram peaks back into balance.

The gel filters I linked to are the cheapest form of transmitted light reduction and colour correction possible, made from dye coated polyester they are tough but susceptible to scratching so if you are in the habit of balancing the panel on top of the OTA to take the flats the filters should be protected by a sheet of clear acrylic to prolong their life. In an observatory the panel can be hung on a wall and the OTA pointed to the panel, in this case the filters will not need additional protection. A few surface scratches won't effect the quality of the flats but over time a build up of scratches in one area can lead to uneven light transmission through the filter.

I know of several users of USB tracing panels who produce good flats with monochrome cameras so without further manufacturers specification being available I would guess the tracing panel you already have would be the best option but if testing shows a wide separation of the three RGB histogram peaks for your OSC camera then add a colour correction gel filter between panel and OTA to balance the spectral output, use the panel on full brightness to minimise banding and variability in output and use a few layers of ND filter cut from a single filter sheet to reduce the output light to a sensible level and this should be ideal.

HTH

William.

[pubquiz]

Hi William

Thanks very much for that detailed expalnanation....greatly appreciated.

Looks like lots of experimenting will will be required.

Prior to getting the tracing pad I took apart an old unused 12" PC monitor I had hoping to use the LED screen from it but found it used tiny fluorescent tubes and part of its circuits board to produve the current to drive it and didnt fancy messing with mains voltage circuits etc .....It did however have 3 or 4 sheets of a Mylar type plastic to even out the light so I kept these and may be able to utilise them as light reducing filters.

Lots of usefil info there...thanks again

 

Tom

 

[michael8554]

There are so many variables I believe there are many right answers.

For instance:

The night sky even at a dark site isn't  black or dark blue, it's a muddy brown, not from sky pollution.

So sky and tee shirt flats might be best, but twilight sky is probably still  blue.

So which camera white balance should you use to match the real night sky?

In any case camera white balance isn't applied to Raws, unless you've ticked the box in DSS etc to apply it.

Does shooting Raw flats in Monochrome Picture Style cover all screen colour balance,  colour temp, and spectral response difficiencies?

Michael 

[pubquiz]

yes it gets ever more complicated i am thinking of taking up knitting as a hobby