STC Multi Spectral Astro clip in filter and using Olympus m43 bodies for astrophotography

STC Multi Spectral Astro clip in filter and using Olympus m43 bodies for astrophotography

Part 2.

STC also provided me with their Multi Spectral Astro clip in filter and when I ran it through my spectrophotometer, the results were much more in keeping with STC’s website specifications and even closer to the QC profile run on every filter they ship.

STCMSAstroFilter2

The transmission profile is very similar to light pollution filters that have been on the market for the past decade, and one of the best known is the IDAS P2.

Screen Shot 2019-11-02 at 6.43.47 PM

This type of filter is particularly useful for astronomers living in cities that are equipped with the older style high pressure sodium vapour  or mercury street lamps but not so much for the newer fluorescent and LED lamps.   I haven’t had much luck in the past imaging with filters of this type from my home in Toronto but I was very curious to try the STC model.  Unlike the narrowband false color method used in Part 1, this filter promises natural color balance so that stars will show true colors and targets like galaxies will be visible.

 

 

insert image here

 

 

 

 

 

When I began my journey in astrophotography some 15 years ago I was a Canon user because Canon DSLRs had the lowest thermal and read noise in their CMOS sensors.  Nikon and Olympus were still using CCDs and it was much more difficult to achieve clean high contrast long exposure astronomical images from them.  And of course like all dedicated amateur astroimagers (I coined this term because a photographer takes short exposure terrestrial photos) I migrated to astronomical monochrome cameras featuring thermal electric cooling (TEC) for best performance but I still like the idea of using a DSLR or mirrorless camera for casual astroimaging.  It simplifies the process and minimizes the amount of equipment one needs to carry to a dark site.

It was the introduction of the E-M1.2 and the E-M1X which has made astroimaging possible with Olympus cameras.  All camera sensors exhibit thermal or dark noise where the heat produced from the densely packaged electronics inside the tight confines of a camera body causes the spontaneous development of electrons even in the absence of incoming photons.   Dark noise increases linearly with temperature and length of exposure and is recorded as a false signal throughout the image.  When it’s bad, it’s like the snow and static burying the image on an analog TV trying to tune in on a weak aerial signal.  You can remove this noise by subtracting a dark frame from the astro image and a dark frame is constructed by subjecting the sensor to the exact same exposure – but completely in the dark.  This is in fact how the camera performs in body long exposure noise reduction.  The camera exposes the sensor twice and performs the dark subtraction.  Astroimagers tend to avoid this method because it reduces their data collection time to half.  If you spend four hours in the field, only two of them will yield exposure data.  So astroimagers create a master dark frame at the end of the imaging session and subtracts all the exposures with this dark frame.   But the subtraction of an improper dark frame can cause more problems by injecting noise into the image.

OlyDarkNoise
The dark noise of a sequential series of 5 minute exposures was measured in Nebulosity software as mean ADU (analog digital unit).  One ADU is the event where one photon hits a sensor well and produces an electron which is converted to a digital pixel value.

The E-M1.2 and E-M1X produces far less dark noise than even the E-M1, and less noise is easier to remove.  The early E-P1 behaves so erratically that dark noise becomes very difficult to remove through dark frame subtraction.  All cameras require considerable time to attain thermal equilibrium indicated by the dark noise reaching a stable plateau.  Again, the E-M1.2 and E-M1X appear to reach this plateau sooner so a dark frame constructed at the end of the imaging session will be fairly accurate for most of the exposures but less accurate for the early initial exposures (approximately in the first half hour).   Elsewhere in this Blog is an article about adding TEC to a Canon T2i DSLR resulting in massive dark noise reduction as well as completely accurate dark frame subtraction since the sensor temperature is maintained at a stable set point (https://jimchungblog.com/2017/01/01/how-to-make-your-own-cooled-dslr-for-way-less-than-late-season-jays-ticket/).

E-M1X5minuteDark
Five minute E-M1X dark frame, heavily stretched to exaggerate dark noise.
E-P15minuteDark
Five minute E-P1 dark frame, heavily stretched to exaggerate dark noise.

Another important issue is to determine what ISO value to use.   The ISO setting controls how much amplification is applied to the ADC (analog to digital convertor).   This is also known as the gain.  Like most cameras, the Olympus bodies are 12 bit so each pixel has a maximum of 4096 levels.   The gain controls how to map these levels to the  well depth or saturation level of the sensor.  The E-M1.2 and E-M1X use the Sony Exmor IMX270 but I can’t find its datasheet online.  The Panasonic Live-MOS sensor used on the E-M1 had a saturation level of 16234 e¯ so one would naturally expect a gain of 16234/4096 = 4 e¯ /ADU but this means single photon hits are not recorded.  A gain of  1 e¯ /ADU would increase the apparent sensitivity by allowing single photon hits to be recorded but it would also lower the level at which bright stars would oversaturate pixels.  A gain of less than 1 e¯ /ADU would improve the contrast between faint signals but further increase the chance that individual pixels would oversaturate, even in the absence of bright stars.

To determine the gain – ISO relationship on the Olympus E-M1.2, three flat pairs were shot at ISO 200, 400, 800 & 1600 at different shutter speeds in RAW and deBayered data from one of the green channels  extracted with LibRaw software.  The flat target was an out of focus sky shot through a white diffuser made from several layers of plastic grocery shopping bags.  Using ImageJ,  the mean ADU of each pair was determined and graphed against the ADU variance (standard deviation squared) of one pair of the flat subtracted from the other pair.  The slope of each ISO straight line is the gain corresponding to that ISO setting.

SystemGainEM12

Since read noise is on the order of 1-2 e¯ , a gain of 2-3 e¯ /ADU would ensure that the signal rises above the read noise floor and would be the optimal gain to be used which corresponds to between ISO 400 and ISO 800.

 

 

Top image is M31 Andromeda Galaxy taken at dark site, one hour total exposure, E-M1 body, 200mm focal length, in camera noise reduction.

Please press [Home] key in the upper right corner to see all the other topics in this Blog.

 

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s