After the last 2 weeks of rain I finally have a Tarantula. I captured the shots back in March during 3 sessions, 8-03, 11-03 and 14-03.
The frame is cropped as I used 0.63 reducer/flattener and didn’t have any flats. Processed in PixInsight as 3 separate batches (MBias and MDark applied) then re-integrated again, with a final touch in LT4 and StarSpikes added in PS.
3x40x2min worth of data @ ISO1600
Taken with 60Da, CPC925 x 0.63, HD-wedge, guided.
Why only 2 minute subs though? Also ideal ISO for a 60Da is probably more like ISO800. ISO1600 is probably only adding noise compared to ISO800 and losing dynamic range and only makes the same signal easier to see on the LCD screen. There is an ISO sweet spot for each camera. Going above that is simply amplifying noise as well as signal. The engineers tend to make optimizations for read noise at different ISO levels. Higher ISO levels damages Dynamic Range. Above about ISO800 Dynamic Range would be dropping significantly, not that this is a too high a dynamic range image (although the centre is quite bright). But you probably have lost 1 stop of dynamic range if not more by ISO1600 on a 60Da which has small pixels.
ISO is a bit of a funny thing because of course the signal is the signal received by the sensor and higher ISO is simply turning up the volume so to speak, noise along with it. It doesn't add signal, simply amplifies what was received. The only way for more signal is a faster lens or longer exposure time.
Would 5-10 minutes at ISO800 and boost in post processing be a better strategy? Do you expose for 1/3rd of a histogram wide for each shot?
3 x 40 x 2mins = 4hours which is a decent amount. Does it take a long time for Deep Sky Stacker to process that many subs?
Sorry for the long post, - its a nice image - really!
Why only 2 minute subs though? Also ideal ISO for a 60Da is probably more like ISO800. ISO1600 is probably only adding noise compared to ISO800 and losing dynamic range and only makes the same signal easier to see on the LCD screen. There is an ISO sweet spot for each camera. Going above that is simply amplifying noise as well as signal. The engineers tend to make optimizations for read noise at different ISO levels. Higher ISO levels damages Dynamic Range. Above about ISO800 Dynamic Range would be dropping significantly, not that this is a too high a dynamic range image (although the centre is quite bright). But you probably have lost 1 stop of dynamic range if not more by ISO1600 on a 60Da which has small pixels.
ISO is a bit of a funny thing because of course the signal is the signal received by the sensor and higher ISO is simply turning up the volume so to speak, noise along with it. It doesn't add signal, simply amplifies what was received. The only way for more signal is a faster lens or longer exposure time.
Would 5-10 minutes at ISO800 and boost in post processing be a better strategy? Do you expose for 1/3rd of a histogram wide for each shot?
3 x 40 x 2mins = 4hours which is a decent amount. Does it take a long time for Deep Sky Stacker to process that many subs?
Sorry for the long post, - its a nice image - really!
Greg.
Thanks Greg for very educational lecture on ISOs.
You see, my major issue is that I’m taking all images from my court-yard close to the Perth’s CBD and because of the location the mathematics on ISO/exposure time is twisted. Probably if I went to a dark location somewhere far away all what you’ve said would be perfectly applicable, but here I have to do my best to satisfy my appetite for reasonable images.
Quote:
Would 5-10 minutes at ISO800 and boost in post processing be a better strategy?
After 10mins the image is as opaque (or white) as a sheet of paper .. LOL
So, my experience is to take max 2min shots at higher ISO (sometimes up to 2500) and degrade the images by, for example, black clipping in the processing stage.
Quote:
3 x 40 x 2mins = 4hours which is a decent amount. Does it take a long time for Deep Sky Stacker to process that many subs?
I use PixInsight and integration of 40subs takes about 5mins ..
I’ll give your suggestions more thoughts and when the weather allows – do some more tests with lower ISO. Thanks again,
Ian
Why the three separate batches? Did you integrate all the frames in one go, or as two separate steps?
Thanks Dave,
Quote:
Why the three separate batches? Did you integrate all the frames in one go, or as two separate steps?
In my experience once you reach the number of subs around 35-40 there is not much difference when you add another 10, or more, and process them in one batch as 50, however, if you do the same as two separate batches and then integrate the two images – the result is different.
Thanks again,
Ian
In my experience once you reach the number of subs around 35-40 there is not much difference when you add another 10, or more, and process them in one batch as 50, however, if you do the same as two separate batches and then integrate the two images – the result is different.
Without any rejection you should get the same result either way. With a rejection algorithm you will get different results but I would expect most rejection algorithms to perform better with more subs - they would have a more robust estimate of each pixel value to work with. Might be an interesting experiment to see if the actual results match theory...
hello ian firstly let me say i envy your astro pics
the first and only dslr i have ever had is my 60da and for the year i have had it i have tried just about every combination of exposure and iso, and i concur with mr bradley that my sweet spot i have found to be iso 800 and between 180s and 240s exposures in mount hawthorn
i too use the f6.3 reducer plus an LP filter which i dont think you use(?)
pat
You’re probably right Rick, if you integrate all subs or integrate integrated images – mathematically speaking it shouldn’t matter, but what if you first process each of integrated images and then re-integrate them and “polish” the resultant photo?
Pat, I don’t use the LP filter because I am pretty convinced that with light pollution it also removes some valuable data leaving the image dominated by blue and purple.
Perhaps on some other targets such as M45, where the nebulosity is almost entirely blue, the LP filter may help.
I’m going to try this on M45 when the weather improves ...
... I don’t use the LP filter because I am pretty convinced that with light pollution it also removes some valuable data leaving the image dominated by blue and purple....
You're doing pretty well without a filter. Nice image!
However if you have heavy LP, I wonder if a gentle filter like a Baader Semi-Apo or moon and skyglow filter would give an extra little boost? A more gentle filter may offer a boost in contrast without messing up the colours.
