Jump to content

Canada's top-tier Telescopes & Accessories
Be as specific as possible when reporting issues and *ALWAYS* include the full version number of the application you are using and your exact *CAMERA MODEL*
NEVER POST YOUR KEY IN ANY PUBLIC FORUM, INCLUDING THE O'TELESCOPE SUPPORT FORUM ::: IF YOU DO YOUR KEY WILL BE DEACTIVATED WITHOUT NOTICE!
  • 0

Dither Algorithm


dts350z

Question

I seem to remember reading somewhere here that BYEOS to PHD2 dithering is "random"?

 

If so, is that a function of BYEOS or PHD2?

 

If BYEOS, is there an opportunity for improvement based on this:

 

"For best results, dithering should be non-random to avoid statistical clumping and should not repeat the same location for the entire stack."?

 

from here: http://hiddenloft.com/notes/NeedDarkFrames.htm

 

 

 

Link to comment
Share on other sites

  • Answers 12
  • Created
  • Last Reply

12 answers to this question

Recommended Posts

OK so I'm imaging at either 0.54 arcsec/pixel or 0.86 arcsec/pixel and my guide rig is 1.54 arcsec/pixel, so a guide pixel is either ~2.9 image pixels or ~1.8 image pixels.

 

How many image pixels should I be dithering? From there I can calculate the guide dither pixels and values for the formula you posted. 

 

Thanks!

Z

 

 

Link to comment
Share on other sites

"How many Image Pixels should I be dithering?" - That's hard to say, as the goal will be to facilitate Stacking that overcomes much of the Thermal Noise and Bad Pixels specific to your Camera.  There is a school of thought which proposes "Large Dithers" for DSLR Images - 10's of Pixels per Dither - given their assertion that DSLR Sensors record Thermal Noise in "large-ish patches".  So, if you are going to try for a Image with Dither of 20 Pixels average, you'd need a 40 Image Pixel Range, which would equate to about 12 Guider Pixels.  So, setting BYE/BYN for an Aggressiveness value of 3 (2x multiplier within PHD) would require a PHD Dither Scale of 6.

 

Does that make sense ??

 

Beyond that rough suggestion, you'll need to decide for yourself the relative positive impact on your Images (based on Stacking results) knowing that these results will be affected by your unique combination of Camera and Environment and Process.

(Sorry...  there is no hard/fast Rule to follow...)

Link to comment
Share on other sites

I'd be rather careful citing much from the sources that you referenced.  They are obviously rather dated sources, given that the "New Dither Algorithm" discussed "as implemented in CCDAutoPilot v2.06" is discussing an "innovation" coded in either 2008 or 2009.

 

Current Dithering applications such as PHD and MetaGuide and SGP and MaximDL have come to the conclusion that "Random Dithering" is the most appropriate form to support.  

It is the ONLY form supported by PHD (which performs the actual Dither operation - BYE simply signals a "request" that a Dither be performed).

 

In fact, for DSLR Imaging, the current consensus is that Large-scale/Aggressive Random Dithers are the best for Noise Reduction.

 

 

Link to comment
Share on other sites

OK thanks.

 

Does that come down to recommended settings (other than the defaults)?

 

...

 

The other stuff on that site is about sub exposure length. Is there a more up to date/appropriate source for DSLR sub exposure, beyond getting some black space on the left of the histogram?

 

 

Link to comment
Share on other sites

Does that come down to recommended settings (other than the defaults)? 

The default settings for PHD2 (currently v2.3.0) are usually quite good for a starting point, and PHD2 does do some automatic optimization of those parameters when you identify your Guide Camera and Mount in the Setup profile.  Beyond that, the best first suggestions are to raise the Max Calibration Step Time to 2000-3500ms and to raise the Max Duration setting to 1000ms.

 

The other stuff on that site is about sub exposure length. Is there a more up to date/appropriate source for DSLR sub exposure, beyond getting some black space on the left of the histogram? 

 

That is usually the best Rule-of-Thumb.  

For much more, look to resources such as:

 

Books and e-Books:

The Deep-sky Imaging Primer by Charles Bracken
Astrophotography for the Amateur by Michael A. Covington
A Guide to Astrophotography with Digital SLR by Jerry Lodriguss
A Beginner's Guide to DSLR Astrophotography by Jerry Lodriguss
 
Videos:
Forest Tanaka series on YouTube
 
Link to comment
Share on other sites

All good info thanks. I'll dive in. 

 

Basically I'm looking for the trade off between shorter exposures, with higher stack counts against longer exposures (lower iso and/or farther to the right on the histogram) with smaller stack counts.

 

It seems to me (and what I've read) that iso is actually pretty irrelevant, compared to sensor temperature noise, and sky fog with space at the left of your histogram, and with my current DSLR cooling rig I can hold the sensor temp to 46 degrees F for long (10 min or less, haven't tried longer) exposures. 

 

For example I could image at 6o seconds @ 12800 iso and see the same amount of camera noise as a 480 second exposure @ 1600 iso. I need a utility to actually measure the noise though, vs. eyeballing a DARK or looking at the histogram.

 

I guess one complication of more shorter exposures is the mirror lock, dither, and dither settle times start to become a larger percentage of your time on target.

 

Speaking of Dithering again:

 

What about the dither aggressiveness in BYEOS? 

 

I've been running with:

 

Aggressiveness: 3

Settle at: 0.25

Calm down: 15 (which also serves as time for the imaging camera sensor to cool down)

Dither every image

 

 

 

Link to comment
Share on other sites

The problem with resorting to Higher ISO levels is that in DSLRs these are implemented by multiplying the photon counts captured, unlike actually increasing the sensitivity of High ISO Films.  The only result is that you generally reduce the available Dynamic Range of your DSLR Sensor.  It is highly unlikely that any DSLR (maybe excepting the new 6D) will produce Less Noise at 12800ISO than at 1600ISO (there are numerous numerical analyses that support this finding).

Link to comment
Share on other sites

Speaking of Dithering again:

 

What about the dither aggressiveness in BYEOS? 

 

I've been running with:

 

Aggressiveness: 3

Settle at: 0.25

Calm down: 15 (which also serves as time for the imaging camera sensor to cool down)

Dither every image

 

Those Dither values are quite reasonable.

If you are using PHD2 (or the older original), you'll also want to check the "Dither Scale" value as that is set in the PHD "Brain" Global Tab.  It is as much that value as the "BYE Aggressiveness" setting which dictates the actual size of the Dither.  "Aggressiveness" values range 1-5 which translate to PHD as values (0.5, 1, 2, 3, 5).

"Dither Scale" is in Guider Pixels (0.1-100.0), while "Aggressiveness" is a multiplier, with the final calculation:

Dither Pixels = Aggressiveness x Scale  x randomizer%.

 

 

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.


×
×
  • Create New...

Important Information

This site uses cookies to offer your a better browsing experience. You can adjust your cookie settings. By closing this banner, scrolling this page, clicking a link or continuing to browse otherwise, you agree to the use of cookies, our Privacy Policy, and our Terms of Use