M31, brown or blue?

[ashworthacca]

I imaged M31 back in 2011 and when I processed it the galaxy was a brown colour.

I've recently reprocessed it, stacking in Nebulosity 3 and processing both in Pixinsight and PS and its still brown.

I've seen a few recently on the forum where the galaxy arms have been a lovely blue colour but I just can't get that colour in mine.

What am I doing wrong? Is it because I used a modded 1000d and CCD CLS clip filter or what?

I can Dropbox the calibrated image tomorrow if anyone fancies a go...

Thanks

[Josh Wilson]

It could be LP. Was it taken at a dark sky or in an urban area?

Josh

[Steve Ward]

I've been reliably informed that this is THE image of M31 to aim for . . . . . http://www.robgendlerastropics.com/M31NMmosaic.html

[Gina]

Well that's certainly a corker!

[Steve Ward]

Think I prefer Olly and Co's version myself , but that's a slightly biased opinion having had some great advice from Olly over the last year . . .

http://ollypenrice.s...LxLS5&lb=1&s=XL

[JamesF]

I think it may depend on the sensitivity of the camera you're using, how long you expose for and how you process it. Even in the linked image the inner part of the galaxy has a more yellowy-brown hue, whereas the outer part is more bright bluey-white. Perhaps because there's more dust collected around the inner region?

James

[laser_jock99]

I've been reliably informed that this is THE image of M31 to aim for . . . . . http://www.robgendle...31NMmosaic.html

Brown and blue then! Older golden stars in the middle and young, blue stars in the outer arms.

Tricky one really getting the colour of anything right- pretty much personal taste.

My own way of getting natural colour is to assume that the spectral response of my DSLR CCD is relatively linear. Then it's an easy task to overlap the colour histogram and end up with something that looks 'right'.

The other 'test' I apply to my processing is whether the sky background (the part that's meant to be black) looks neutral in colour. Any hint of excess green, magenta etc suggests I need to adjust the overall colour balance of the image.

My own take on M31 might be too 'blue' some tastes.

Here is a 'warmed up' version (4800K) but I like the sky background better in the 6500K version above.

[ashworthacca]

Sorry for the delay in replying to my own thread

It was taken in my back garden which suffers from mild light pollution, the main culprit is a neighbours security light that constantly goes on and off.

When I get the chance I'm going to try it again with my non-modded 550d to see if it's the modification that is giving it the warmer tones.

Thanks guys

[ollypenrice]

Think I prefer Olly and Co's version myself , but that's a slightly biased opinion having had some great advice from Olly over the last year . . .

http://ollypenrice.s...LxLS5&lb=1&s=XL

You're very kind but the link is to Olly's old solo version. 'Olly and Co's' version (Les, Tom, Ian and myself) is this one; http://ollypenrice.smugmug.com/Other/Best-of-Les-Granges/i-FGgG233/0/X3/M31%20LES%20OLLY%20TOM%20TEC%20CORE-X3.jpg

I think Rob Gendler's version is probably more correct though I often feel he drifts somewhat into the magentas. However, he's a rigorous imager and knows a sight more than I do. A lot depends on whether or not you apply an Ha layer to emphasize the star forming regions. In our version Les captured an excellent depth of Ha.

There are various ways to go about colour calibration, the most favoured being the G2V method. What we call white light is the light from the sun, a G2V spectral class star. So you image a G2V star and set your colour weights to give white for that star. I'm lazy and let Pixinsight do it for me but Peter Shah has used the G2V calibration and you can't argue with his stunning colour balance.

More than one way to swing a cat. (Curious expression!!)

Olly

[Steve Ward]

Whoops , sorry Olly . . .

[Gina]

Olly & Co's version is superb - I love it That is what I shall strive for - on a clear night with no moon - ha ha...

[IanL]

I think it is important to bear in mind that colour is not a physical reality, it is created by human perception. Whilst there is a physical reality in that different objects do emit different spectra of light, that absolutely does not mean that you will perceive a given spectrum as a given colour. Your perception will be affected by the overall scene and the juxtaposition of different objects. The colour of an astro-image is a matter of personal choice and taste.

So for example in the M31 image, try masking off the galaxy and all the stars and then changing the colour balance of the background, try a black background, then a blue tint, then yellow, then red and you should see that the 'colour' of the galaxy and stars will appear to change even though you haven't adjusted them at all. Ultimately the 'true' colour of all nebulae is grey, since they are too faint to stimulate the cones in your eyes.

People have preconceived ideas that a given object should be blue-ish or red-ish based on many preceding images that they have seen. Some of the examples above may look 'wrong', but only because you already have a fairly fixed idea of what is 'right', rather than them being intrinsically wrong. The Hubble palette is completely unnatural, but there are certain popular objects that amateurs now image using narrowband filters, and they wouldn't dream of choosing a different set of colours for the composite because it would deviate from what they and the audience now expects.

And before anyone says 'well what about photographic film images, they must be true representations of colour", the combination of different colour films, processing and the slide/print stock used to print them produce different colour tones depending on the dyes that are used, never mind the printing process (or your computer screen) which further changes the 'true' colour of an object.

If you really think an object has a true 'colour', take a look at these and then get back to me:

http://www.lottolab.org/illusiondemos/Demo%2012.html

http://www.lottolab.org/illusiondemos/Demo%2015.html

http://www.lottolab.org/illusiondemos/Demo%2010.html

http://www.lottolab.org/illusiondemos/Demo%2024.html

Main page is here if you want to find out more:

http://www.lottolab.org/articles/illusionsoflight.asp

[ollypenrice]

I think it is important to bear in mind that colour is not a physical reality, it is created by human perception. Whilst there is a physical reality in that different objects do emit different spectra of light, that absolutely does not mean that you will perceive a given spectrum as a given colour. Your perception will be affected by the overall scene and the juxtaposition of different objects. The colour of an astro-image is a matter of personal choice and taste.

So for example in the M31 image, try masking off the galaxy and all the stars and then changing the colour balance of the background, try a black background, then a blue tint, then yellow, then red and you should see that the 'colour' of the galaxy and stars will appear to change even though you haven't adjusted them at all. Ultimately the 'true' colour of all nebulae is grey, since they are too faint to stimulate the cones in your eyes.

People have preconceived ideas that a given object should be blue-ish or red-ish based on many preceding images that they have seen. Some of the examples above may look 'wrong', but only because you already have a fairly fixed idea of what is 'right', rather than them being intrinsically wrong. The Hubble palette is completely unnatural, but there are certain popular objects that amateurs now image using narrowband filters, and they wouldn't dream of choosing a different set of colours for the composite because it would deviate from what they and the audience now expects.

And before anyone says 'well what about photographic film images, they must be true representations of colour", the combination of different colour films, processing and the slide/print stock used to print them produce different colour tones depending on the dyes that are used, never mind the printing process (or your computer screen) which further changes the 'true' colour of an object.

If you really think an object has a true 'colour', take a look at these and then get back to me:

http://www.lottolab....os/Demo 12.html

http://www.lottolab....os/Demo 15.html

http://www.lottolab....os/Demo 10.html

http://www.lottolab....os/Demo 24.html

Main page is here if you want to find out more:

http://www.lottolab....ionsoflight.asp

Interesting and I certainly vary my background sky colour weighting in order to make it look 'right' in the context of the object. A bog standard background sky isn't an option (to my eye) because, as you say, our perception of it is affected by the other colours seen against it.

However, I thnk that using carefully contrived extreme examples, as in the demos above, can be slightly misleading or can overstate the case somewhat. There remains a fundamental difference between RGB imaging and narrowband colour mapping. One can plot the peak spectral emission of parts of spiral galaxies and find that the arms peak at shorter wavelengths than the core. The core Popuation 2 stars are redder than the young spiral Population 1. Recent UV images of Andromeda confirm the presence of bright emission is parts of the arms, suggesting that the sometimes-controversial blues may well be justified.

Essentailly I'd say that the measured peak emission of parts of an object should be presented in a colour that we see for that wavelength on Earth. Yes, what is next to it will slew our impression of that colour but don't you think that some correlation should still exist? And if we could make an F0.1 ten metre telescope would we still see grey? I think we'd see colour and I think that good RGB images should show those colours.

Some (considerable) guesswork has to go into it, of course!

Olly

[cfpendock]

That is very interesting, Ian. I have often thought about colours of nebulae and stars, especially when adding a touch of Ha or other. Of course it is nice to see the normally invisible radiation, but I find it difficult to remember that these images are not "correct" from a human observers point of view, as well as the fact that slight differences in background or nearby star colour can have a huge effect on the perceived colour of the main object. I like the G2V approach described above by Olly.

Chris

[nephilim]

I've been reliably informed that this is THE image of M31 to aim for . . . . . http://www.robgendle...31NMmosaic.html

Hmmm.....asking a bit too much of my XBoxcam

[nephilim]

I think it is important to bear in mind that colour is not a physical reality, it is created by human perception. Whilst there is a physical reality in that different objects do emit different spectra of light, that absolutely does not mean that you will perceive a given spectrum as a given colour. Your perception will be affected by the overall scene and the juxtaposition of different objects. The colour of an astro-image is a matter of personal choice and taste.

So for example in the M31 image, try masking off the galaxy and all the stars and then changing the colour balance of the background, try a black background, then a blue tint, then yellow, then red and you should see that the 'colour' of the galaxy and stars will appear to change even though you haven't adjusted them at all. Ultimately the 'true' colour of all nebulae is grey, since they are too faint to stimulate the cones in your eyes.

People have preconceived ideas that a given object should be blue-ish or red-ish based on many preceding images that they have seen. Some of the examples above may look 'wrong', but only because you already have a fairly fixed idea of what is 'right', rather than them being intrinsically wrong. The Hubble palette is completely unnatural, but there are certain popular objects that amateurs now image using narrowband filters, and they wouldn't dream of choosing a different set of colours for the composite because it would deviate from what they and the audience now expects.

And before anyone says 'well what about photographic film images, they must be true representations of colour", the combination of different colour films, processing and the slide/print stock used to print them produce different colour tones depending on the dyes that are used, never mind the printing process (or your computer screen) which further changes the 'true' colour of an object.

If you really think an object has a true 'colour', take a look at these and then get back to me:

http://www.lottolab....os/Demo 12.html

http://www.lottolab....os/Demo 15.html

http://www.lottolab....os/Demo 10.html

http://www.lottolab....os/Demo 24.html

Main page is here if you want to find out more:

http://www.lottolab....ionsoflight.asp

There was a QED documentary about colour perception a while back, very interesting, theres also a school of thought that says personal upbringing & surroundings define how we individually percieve colour & that one colour to one person can differ quite alot to anothers. Thanks for posting.

Steve

[IanL]

Maybe I am being a little provocative, but I really don't think I am.

The examples above absolutely prove that context is everything when it comes to colour, and not just in extreme or contrived examples. You would swear that the grey object is bright blue or bright yellow, and even when you switch between the two views, as soon as you put the grey in context you perceive it as one or the other colour. There is no way you can force yourself to see they grey that intellectually you know is really there. This process happens continuously in real life, but you just aren't aware of it most of the time. Further, relative brightness of objects also affects perception of colour; see some of the other examples linked from that site's main page.

I think it is fair to say that measuring the spectrum of a given object will tell you what colour most people would say it appears to be in isolation of anything else (i.e. if you had a big patch of it an no other colours in the field of view). I also agree that by measuring wavelengths, you can calibrate the imaging system and processing to ensure that each object is represented in its true/natural colour. But the point is, that won't necessarily result in an image that looks true or natural.

Take an example: You make an Ha emission lamp and image a big block of the light with your camera. Now you take the image and balance it on screen so that it appears the same colour on screen as it does to your eyes in real life. So we have a baseline for Ha in isolation. Now image (say) Andromeda. You can now balance the colour in your image so that the Ha pixels match the hue of your baseline image. Let's assume you do the same sort of thing for the other main objects in the image, baseline and balance the equivalent spectra in isolation, and then balance the galaxy image to match.

So now take a look at the resulting image. Does it look 'right'? I very much doubt it would look right, since you are now looking at your perfectly balanced set of colours in context with each other and, as the examples above show, they will appear to be different colours than you were aiming for. How different depends on the context, I do not dispute that, but there will be a difference.

So what do you do next if you Ha appears 'too magenta' or 'too orange' in the context of the image. Why you rebalance the colours to match your expectations of course! But the point is, that is an artificial choice that you have made, not the true reality of how the object would appear through our hypothetical f/0.1 telescope, where the Ha would appear to your eyes to be magenta or orange or something else.

[Steve Ward]

Not to mention the level of "colour-blindness" that we all have to some degree . . .

Hence my first "wide" shot of M31 that was distinctly purple to most people but looked lovely to me . . .

Have since tweaked it 100 times and every version is different . . . . but still lovely to me . . .

[ollypenrice]

Maybe I am being a little provocative, but I really don't think I am.

The examples above absolutely prove that context is everything when it comes to colour, and not just in extreme or contrived examples. You would swear that the grey object is bright blue or bright yellow, and even when you switch between the two views, as soon as you put the grey in context you perceive it as one or the other colour. There is no way you can force yourself to see they grey that intellectually you know is really there. This process happens continuously in real life, but you just aren't aware of it most of the time. Further, relative brightness of objects also affects perception of colour; see some of the other examples linked from that site's main page.

That there is some truth in this is beyond doubt. But if we take a view of the matter as extreme as yours how do we account for tha fact that we can say to each other, for instance, 'You have a dark blue Escort and I have a red one,' and make any sense to each other? Clearly we can say this and make sense so we must perceive there to be some innate or consistent quality in the colour of our respective cars which we can meaningfully name. Yes, we might say as we head out of the observatory at three in the morning, 'Funny, our cars look the same colour at the moment' but we are still happy to say one's red and one's blue, and an overwhelming majority of people will agree. If context really were everything, as you suggest, how would we ever agree on what was a blue car and what a red one, as I suspect you'll agree that we can? How would the paint industry function? Perhaps it would make only one colour and advise its cutomers to pay attention to the context? I'm convinced that there is something consistent in our colour perception, though not immutable in different contexts.

Olly

[opticalpath]

I do initial RGB balance using G2V calibration which seems to work pretty well (subjectively) for stars. But as soon as there is an extended object in the field like a galaxy or nebula, I'm nearly always dissatisfied with the initial G2V-balanced result and end up tweaking it. When I start adding Ha, all bets are off and it's just down to what looks right; there's a lot of just eyeballing it. I agree with the view that we're much influenced by preceding 'reference' images, and probably by changing fashions too.

I suppose we could be truer to the data by just displaying the separate monochrome images taken with each filter - like science data - but wouldn't it be dull to stop there!

Adrian

[Steve Ward]

Another example of the difference in peoples ideas of "correct" colour will have been seen by most of us when watching television in someone else's house .

I'll bet there are not many who've not had the urge to adjust the picture somewhat . . .

And as for walking round in Currys surrounded by a multitude of massive , garish sets all set to max Colour and max Contrast . . .

[IanL]

Another example of the difference in peoples ideas of "correct" colour will have been seen by most of us when watching television in someone else's house .

I'll bet there are not many who've not had the urge to adjust the picture somewhat . . .

And as for walking round in Currys surrounded by a multitude of massive , garish sets all set to max Colour and max Contrast . . .

Yep, most TVs have a 'shop display' mode hidden away somewhere that turns everything up to 11. I remember when we switched from our giant Sony CRT to a flat-screen Samsung, took me about two years to get used to the thing due to the much more saturated colours. Now I don't even notice.

[Steve Ward]

Best you hide the remote if I come round then . . . .

[Qualia]

Just adding a shot to Ian's very informative posts...

We can note that the philosophy of colour is a very rich and fascinating area of study dating right back to Galileo. Indeed, as science informs us, the seas and the skies are not really blue, nor are apples really red. As the great empricist Hume went on to put it in 1738, "Sounds, colours, heat and cold...are not qualities in objects, but perceptions in the mind."

Keeping this brief, we can note, then, that observation cannot be simply a matter of mere perception; instead, it is a cognitive achievement that involves perceiving that something is or is not the case. In this sense, we can make the distinction between perception and inference. If our brains held no hypotheses, no inferences about the world when encountering it (or, alternatively, if these hypotheses were fixed-in-stone) then we would not be able to learn and adapt. This is to say that observations and experiences have to be interpreted and it is this unavoidable involvement – even at the level of the brain functioning – that constitutes a singular 'problem' for science, namely, that of theory-ladenness.

The realization - if correct - is that because perceptions, observations, facts are not neutral and that they are theory-dependent means we can no longer appeal to a body of facts as an impartial tribunal that judges the validity of a theory. All observations rest on some form of conceptual framework that is brought to the table of experience. Ergo, facts presuppose some kind of conceptual framework a given body of people, society, whatever, are working within and that by taking on board the enormity of the theory-ladenness of observation and in plain terms, of classical empiricism as a source of knowledge of truth, we come to the realisation of its 'falsity'

However, I personally don't think this is something we should lament, but merely the realisation that judging what there is or not in the world or universe by merely passively observing it was too simplistic a notion. The refrain “What we see depends mostly on what we look for” is not a cynical statement but an encouragement to look again and again in different ways as part of a truly reflexive practice and if adopted will surely aid our nightly observsations and wonder.

[Steve Ward]