From my short time on ISS, I have to say it is an very enthusiastic group, polite, but often not very critical. By critical I am not suggesting you go around calling someone a boof-head!
When I look at a deep sky image the things I like to see are, excellent focus (this is the one thing you can nail every time), excellent tracking (i.e. round stars) accurate colour balance.
There are other aspects, but for now the above three really are essential to a good image.
As a case in point you would not be happy with a picture of your family/dog/cat that was out of focus, blurred and had a strong green colour cast from having the wrong white balance setting.
Focus and tracking are self evident. But getting the colour right on an object you can not physically see in colour with a warm human eye at the eyepiece is a little more tricky.
Happily there are many professional images out there, for example the treasure-trove of David Malin images taken with the AAT. A good deal of effort has been made to get the colour right in these images and can be considered as the gold standard of colour accuracy.
A case in point, the Tarantula Nebula is not blue, or cyan as rendered by many DSLR’s.
It’s RED.
Hence to post an image of it looking blue, unless you have used a narrow-band Hubble telescope like filter set, is just plain wrong. So how do you get the colour right?
After all, images you take in daylight look spot on.
The problem stems from the internal I/R cut-off filter used by most DSLR manufacturers, which only passes around 20% of the critical Hydrogen Alpha light in which most nebulae glow. As a result emissions from elsewhere in the spectrum will swamp what little red data is being let through. So you end up with a blue-green spider. There are 3rd party companies that will replace this filter for a more suitable one. Canon also made the 20Da for a time which has a factory fitted version.
Modifying a not inexpensive DSLR would make many hesitant. However armed with the knowledge your image data has very little deep red signal, you can use software like Photoshop to boost the curve of the red data and once again approximate reality. If you have not already done so, I suggest you give it a go. You may be surprised by the results!
Hi Peter, I am one of the guilty who post blue images of Tarantula nebula. I think I understand what you are saying with regard to emmission nebula colour like that displayed in the Tarantula Nebula. I have one question. I notice on your web site an image of Eta Carinae Nebula taken with an SBIG CCD camera. It appears to all wants and purposes to have the same colour cast and tone as my images of Eta Carinae. I realy have trouble understanding why an item like Eta C can be so similar yet Tarantula so diferent. If the blue colour isn't there in reality, where is it coming from? Images of tarantula taken and corrected by Hubble show quite a blue core with red to pink surronds which I would have thought indicated a large reflective base for this nebula.
How on Gods earth do you decide which image requires added red. We are not all fortunate or financial enough to purchase a large format CCD which displays true colours, much and all as I would like to. After all most of us are taking photos to develope some skills and for our own gratification and if the colour is wrong does it realy matter.
Colour is so subjective in these images. I'd be more interested in getting a pleasing, detailed result than attempting to replicate some colour that might be visible in a huge apeture.
Most professional images are false colour to highlight features of interest anyway. Surely getting as much detail as you can in an aesthetically pleasing way is the ultimate goal for the amateur?
Ahh yes, the colour balancing can be a hairy issue. My old 300D did just that, showing even known to be fairly red nebulae such as Eta carina as a bluish pink, and the tarantula a strongly blue. Photoncollector (Paul) showed me that using Curves in photoshop on the red channel could bring out more red without making too much of an overall red cast. Also, I found out that by using a nebula filter like the Baader UHCS, Lumicon Deep sky or equivalent, and using high ISO to get around the filter factor, a good deal more red would come through.
The software I use, Iris, has a commandline function called "white". I draw a small box on a dark area of sky with no stars inside, then enter "white". This tells Iris to balance the image to get a neutral grey on the sky background.
The Human eye, however, sees astronomical nebulae more like an unmodded DSLR. As an unmodded camera looses HA sensitivity dur to the IR cut filter, the human eye looses HA sensitivity because the Rod cells in the retina cannot detect low levels of HA light, and the more sensitive rod cells cannot either, only shorter wavelenghts, only very bright nebulae in big scopes such as M42 in a 20 inch dob will show much red. Some peoples eyes are better than others perhaps, but I struggle to ever see any red with my 10 inch scope and low power eyepiece.
With my modded DSLR, using a nebula filter brings out plenty of red, so much so that I need to be careful not to make it too intense.
here are some raw images processed only with canon's own Digital Camera professional software to show us the images without any external colourbalancing or any other adjustments done, they are "as shot" Similar shots with my old 300D would show the sky background as bluish, even with the filter in place.
Ive read that the one true way to colour balance, is to take a picture of a white card, preferably an official photographers colour chart, using the exact filterset that you will use to take astro images with, take note of the adjustments deeded to render the white card as white, then apply the same settings to the astro images. I might see about trying that out myself, take a pic of something I know to be white, in midday sunlight, run the image through Iris, and take note of the red-blue-green adjustments in order to get it looking white. Will be an interesting exercise. Another way would be to take a short exposure of a deliberately out of focus non saturated star of spectral type G2 (same as the Sun) and balance that as white.
Scott
EDIT: I should also say that turning OFF the Automatic White balance in the SLR camera might be a good idia too, and set it to Daylight instead, although with Iris, as it does its own debayering, the AWB might not matter
The centre of Eta, indeed a number of nebulae have quite a lot of H-Beta...with a strong blue component....which is what you may be seeing.
The core of the Tarantua has this as well, but the outer regions of both are well and truly H-alpha red.
Also I am not saying don't get out there, have a go and have some fun.
But when you post a blue NGC 2070, as I said earlier, it's like people with green skin....fun perhaps...but accurate? No.
The fix is so simple, and does not require a $10K CCD. Just ramp up the red channel in Photoshop.
Why strive for accuracy? Images with excellent colour fidelity tell a story about the real physical processes going on in an object. Plus getting it right, I believe helps other aspects of you image processing skills.
Quote:
Originally Posted by Hagar
I notice on your web site an image of Eta Carinae Nebula taken with an SBIG CCD camera. It appears to all wants and purposes to have the same colour cast and tone as my images of Eta Carinae. I realy have trouble understanding why an item like Eta C can be so similar yet Tarantula so diferent.
Last edited by Peter Ward; 15-01-2008 at 01:33 PM.
Reason: links added
Why strive for accuracy? Images with excellent colour fidelity tell a story about the real physical processes going on in an object. Plus getting it right, I believe helps other aspects of you image processing skills.
At best this must be an approximation, so
I think that should read material distribution.
Spectral analysis gives us the proper story.
Now to the modded DSLR.
How do you subtract the now strong IR componant in the image?
Do you use a IR filter as us CCDers do?
From my short time on ISS, I have to say it is an very enthusiastic group, polite, but often not very critical. By critical I am not suggesting you go around calling someone a boof-head!
Don't make bland assumptions. To start with, I recieve very good (and extremely helpful) criticism of my images from talented and polite people like Jase, Bert, and a few others. Not everyone says just 'Nice image'.
Quote:
Originally Posted by Peter Ward
When I look at a deep sky image the things I like to see are . . . . accurate colour balance.
As a case in point you would not be happy with a picture of your family/dog/cat that was out of focus, blurred and had a strong green colour cast from having the wrong white balance setting.
But getting the colour right on an object you can not physically see in colour with a warm human eye at the eyepiece is a little more tricky.
. . . and those horrible neon green, blue, red etc reverse looking images are ghastly, and yet they are called 'magnificent'. Yes, the detail is nice. I might not like the colours used, but I admire the detail!
Quote:
Originally Posted by Peter Ward
Happily there are many professional images out there, for example the treasure-trove of David Malin images taken with the AAT. A good deal of effort has been made to get the colour right in these images and can be considered as the gold standard of colour accuracy.
But are they having 'fun'??? I know I am with my incorrectly coloured images.
Quite frankly, I am getting fed up reading your posts! All they ever seem to say is 'you ain't doing diddly-squat unless you use all my expensive equipment and become like a professional!
Stuff that!!!
I may only use an ED80 on an EQ6 with an El-Cheapo Toucam, but I am having a blast, and if I'm happy with my images, who gives a poop? You apparently!
Enough said before I go further and it gets deleted!
Last edited by ballaratdragons; 15-01-2008 at 02:31 PM.
Reason: spelling
Jeff, that narrow band image I did of Eta Carina has purely arbitrarily assigned colours, and I'd be a little more than concerned if an image from my DSLR ended up looking like that .... You are never going to get accurate colour by imaging through narrow bandpass filters that only look at very specific parts of the spectrum, hence this is not really relevant to my original post....I simply wanted to make sure people don't get the two confused.
The colour of a celestial object can tell you a lot about it. Malin's "Colours of the Stars" devotes a couple of hundred pages to the topic, and is well worth the read.
With a modified DSLR, as with CCD's stars will saturate to white from the IR component. Using an IR cutoff is pretty much standard issue as the fix
You missed the point entirely...... Some of my most satisfying images were taken with a DSLR, photo-tripod and 80mm APO.....
Yep, and cranking up the red channel in PS costs a motza.... Cheesh.
Quote:
Originally Posted by ballaratdragons
I am getting fed up reading your posts! All they ever seem to say is 'you ain't doing diddly-squat unless you use all my expensive equipment and become like a profsessional!
Stuff that!!!
I may only use an ED80 on an EQ6 with an El-Cheapo Toucam, but I am having a blast, and if I'm happy with my images, who gives a poop? You apparently!
Enough said before I go further and it gets deleted!
Quite frankly, I am getting fed up reading your posts! All they ever seem to say is 'you ain't doing diddly-squat unless you use all my expensive equipment and become like a professional!
Stuff that!!!
Come on Ken! No one's trying to put anyone down! I think the gist of Peter's post is quite clear. A standard unmodified DSLR shot is low in red because of all the reasons mentioned. He clearly states that it would be good to fix that in PS to give a more accurate rendition.
Why does everyone on here get all edgy when someone gives constructive criticism? I didn't see the post say "buy an SBIG and your problems will go away". At best, there are some examples that go to show what can be done. I think if everyone just took a step back, took on board some of the advice of one of the leading astrophotographers in the country, we would all learn a lot.
Come on Ken! No one's trying to put anyone down! I think the gist of Peter's post is quite clear. A standard unmodified DSLR shot is low in red because of all the reasons mentioned. He clearly states that it would be good to fix that in PS to give a more accurate rendition.
Why does everyone on here get all edgy when someone gives constructive criticism? I didn't see the post say "buy an SBIG and your problems will go away". At best, there are some examples that go to show what can be done. I think if everyone just took a step back, took on board some of the advice of one of the leading astrophotographers in the country, we would all learn a lot.
Turbo
It wasn't about DSLR's and IR filters that was getting to me. That info is obvious. It's the constancy in other threads (and a little in this one) about how we 'need' the top gear. Read back thru them all if you've forgotten the theme in them (theme being the need for 'expensive gear'). Usually with many replies saying much of what I implied above. We often can't afford top gear and we are having fun with what we have for now.
It gets to me coz I tend to think of newbies coming in and reading about astrophotography. It would turn them off attempting it when they read that they probably need Paramounts, SBIG's etc to get an image worth posting. And now the colour balance has to be perfect (whatever perfect colour is).
Some are even worded and slanted to make me feel like I own crap equipment!
The information is probably correct, maybe just worded badly.
Hi folks.
Here are 2 Tarantula neb images. Both have been dark subtracted, flatfielded, aligned and stacked, but no stretching, curves or the like.
The green one is not colour balanced in any way, I then drew a small box in a starless area of sky, entered "white" in the Iris commandline, and presto, a more balanced image
Scott
If someone posts an image and I like it I will say "nice image". My criteria for what makes a nice image has as much to do with the experience, equipment and observing location of the imager as it has to do with the result itself.
If someone with little experience posts a green looking Tarantula and I can see detail in it to me that is great. It reminds me of the view at the eyepiece. I have never imaged the Tarantula so I'm not going to criticise someone who is achieving something I've never done myself.
I'd be thrilled to see anything at all in my first Tarantula image
Scott's excellent example shows in a nutshell what I'm suggesting, which is not perfect colour, just accurate colour.
It's just a mouse click away
P.S.
Ken. Unless English isn't your first language, for the life of me I can't see how you construed "buy brand X or live in misery" from this post. Have a lie-down and a Bex
Quote:
Originally Posted by tornado33
The green one is not colour balanced in any way, I then drew a small box in a starless area of sky, entered "white" in the Iris commandline, and presto, a more balanced image
Scott
Well i've said no. What determines accurate colour? Is it what you see through the eyepiece? or what you would see if you could get close to it?
Or is it what a very long exposure would show? Or what a spectrograph would show?
Open to whatever interpretation the individual desires i guess.
INDIVIDUALITY RULES!!!!!!
P.S.
Ken. Unless English isn't your first language, for the life of me I can't see how you construed "buy brand X or live in misery" from this post. Have a lie-down and a Bex
"Ken. Unless English isn't your first language". Don't be a smart *rse!
I never said you said "buy brand X or live in misery". You do however imply that we 'need' the top shelf gear or our images aren't very good (in other threads, not this one).
"Have a lie-down and a Bex" Read my reply in the first line of this post again.
well now that you have mentioned the war - what is a dlsr killer Peter? somehow i cannot see how a $6200 hunk of plastic and metal can be that much better than a modified cooled dslr or for that matter a non modified camera for instant pleasing results. geez for that money i would want hot and cold running maids to type in the cooords of the next object. My filter is shown in the diagram below on a Canon 350D. All up mine would be worth 2600 australian for the body and the modification. I know this camera impressed a certain award winning gentleman who shall remain nameless for the instant result it gave. In the hands of a genius it will give more than a pleasing result (in my hands maybe not but i certainly will enjoy the butchers picnic). as for low noise for that particular model - show that will you. some little testing comparisons that were done are here just with my camera.
as for the colour - does anyone take in an ionised atmosphere, street lights, other IR sources and just processing because of a crappy monitor or eyesight. It all comes down to the person and artistic license. What looks good to you?
I am getting fed up reading your posts! All they ever seem to say is 'you ain't doing diddly-squat unless you use all my expensive equipment and become like a professional!
Stuff that!!!
or what mental leap did you make to work out my inexpensive suggestion on getting accurate colour, was a thinly veiled attempt to get everyone to run out and buy a metre class RC?
Unless it is purerly for science it is a MUTE! point.
As i said individuality rules. Lets not start WW3 over such whimsical nuances.
Just my opinion. If i could afford $100K worth of gear it might mean something different to me,LOL.
Cheers
.... You are never going to get accurate colour by imaging through narrow bandpass filters that only look at very specific parts of the spectrum, hence this is not really relevant to my original post....I simply wanted to make sure people don't get the two confused.
. My filter is shown in the diagram below on a Canon 350D. All up mine would be worth 2600 australian for the body and the modification. I know this camera impressed a certain award winning gentleman who shall remain nameless
)
