Canon T3I Dark Current Analysis

Because I found learning calibration challenging during a demo for PixInsight (where horrendous results were obtained initially through simple mistakes), I also started to pay more attention to bias, dark and dark flat calibration frames.  Whereas in Astro Pixel Processor, calibration is so easy that no attention need be paid, in PixInsight the user is down in the weeds looking at data... trying to figure out what's gone pear shaped.  And so, I began to do some searches on the behavior expected for the Canon T3I, whether ISO1600 was (in fact) suitable for lights and darks, whether ISO100 is suitable for flats, etc.

This led to an old (2015) page from the Blackwater Skies Blog on their PixInsight workflow for the Canon EOS 500D that (similar to other posts) noted that the dark current visible in Canon EOS DSLR cameras is modified before written to the RAW CR2 files to manage the dark current levels.  At one extreme, this may suggest fixed pattern noise is no longer an issue in modern Canon EOS sensors... and darks are unnecessary.  Other posts suggest this in-camera solution is incomplete.  Bizarre effects can be seen in some analysis (including lower median dark amplitude as a function of camera temperature).  I utilized the same Batch Statistics Script insight PixInsight as well as exiftool (and Libre Office Calc) to dump median brightness and median absolute difference (MAD) statistics for 281 dark frames collected at ISO1600 across a range of exposures and temperatures.

These results do not show an inverse relationship between median brightness and temp as seen in other Canon dark frame analysis.  This suggests perhaps a more predictable relationship between temperature and dark current for this DSLR.




Comments

Post a Comment

Popular posts from this blog

Markarian's Chain

Image
Started a mult-session collection on 03/26 (52) and continuing on 04/15 (47), 04/16 (109) and 04/19 (~20) against Markarian's chain  with the venerable Orion ST80 using the generic dual narrow-band filter , clear Astronomik EOS clip-in filter (for consistent focusing with the EOS CLS clip-in filter) and Canon T3I (astro modded for Ha sensitivity).  180s collections used for each (except for 3/26 at 300s) and flats collected against morning blue sky when needed based on re-focusing (03/26, 04/16).  Canon configured with UTC time.  Based this collection in Kstars/Ekos/Indi by centering on  NGC 4438 . For this processing, utilized both 60x180s and 30x300s darks as well as 130 bias (all at ISO1600) frames for calibration.    Collection notes from 04/19:   Ran entirely on the Odroid XU4 via remote desktop protocol (rdesktop client with compression) to simulate a remote location and provide some robustness for network / wifi drop-outs. ...
Read more

Great North American Eclipse

Image
4/8/2024 On Akron OH captured a time-lapse of short exposures through Baeder solar film through the venerable Orion Short Tube 80 with the Svbony 405cc using AstroDMx Capture until totality. Results were great considering the high level cirrus clouds.  After light editing in digikam, found some evidence of corona!
Read more

M94 and M106

Image
 3/7/24 -- first moonless, cloudless night with workable temps.  Attempting to capture M94 high in the north sky from the Meade LX10 EMC with UV/IR cut filter, Meade 0.63 focal reducer and the Svbony 405cc camera.  Set camera temp to 0C and using 0 gain and 0 offset. Telescopius simulator result for M94 using the SCT with the 0.63 focal reducer and the SV405CC camera. Telescopius M95 altitude plot for 03/07/24. Cleardarksky forecast for 3/7/24 for Montgomery County showing fairly clear skies and reasonable seeing expected after midnight. Meteoblue forecast for 03/07/2024 for Silver Spring illustrating clear skies expected for the evening. Before mounting the OTA and counterweights, checked the mount as level and utilized the polar scope (due to a clear view of Polaris through the backyard tree -- a rarity).  Adjusted the altitude up slightly and did a sanity check to ensure ~39 degrees. Initial alignment measured with PHD2 from the Odroid was ~ -11 after several minu...
Read more