Dithering PHD2 every x pictures


Would it be possible to set a “Dithering every x pictures” being x the Numbers of images taken before PHD2 dithers?

It would be helpful if someone dithers, but not for every image taken.

Best regards


I’m new to SGP but dithering every “X” frames is the option I chose when setting up the program.

milliesand, I think you might be thinking of setting up the auto focus parameters, which has an option to auto focus every “x” images. I don’t believe there is currently an option to dither every “x” frames.

The option to dither every X frames does not exist.

Just wondering what the reasoning is for this request? Generally you want to dither such that noise doesn’t stack on top of each other and overwhelm the signal. If you’re not dithering very often then you’re stacking noise and amplifying the noise signal as well.

Is settling taking too long? I guess I’m after the reason rather than the solution.


Hi Jared

Yes, reason is mainly time. On a night with 100 exposures and 3 min each (hyperstar use and Canon DSLR) I found that almost an hour is taken when dithering after all frame was selected using Backyard EOS linked to PHD. So, a session of 5 hr exposures takes about 6 hr if I choose to dither after every frame.

That time dramatically lowered when I chose dither every 3 or even 5 images. To me, there wasn’t too much difference when I chose this option over dither after every frame. Given that there were about 100 exposures, and dithering after each 5 frames, there was plenty of signal to take advantage of the dithering process. And almost more than an additional hour exposure.

So, It is a valuable resource when the imaging session is composed by a lot of relatively short exposures (May be not so much if you have not so many long frames).

It would be very good to have this option.

Best regards


IMHO taking a few more frames by not dithering is less effective in reducing the noise than the same amount of time spent dithering.

I agree if you have few individual frames. However dithering every x frames still dithers. So a sigma clipping stack will still remove the hot pixels.



Not so sure that is correct but I will defer to the math gurus on this one. I know that I would never not dither every frame as I want all that I can get from my data and it takes however long it takes.

I’d suggest the only time dithering after a few frames will be superior is if you’re cycling through L, R, G, B then dithering. But then unless you’re super-parfocal you may miss out on the focus front.

Yes, that would make sense as the individual frames from each frequency band would still be dithered. OTOH, it would also mean changing filters more often which takes some (although probably a bit less) time as well. The focus could be dealt by using offsets unless I am mistaken.

In truth, watching my system which is set to very high dither, the time to complete a dither vs. not dithering is rarely more than 30 seconds/dither. I still feel the downside to not dithering exceeds the upside to skipping it.

True, it could not “hurt” to include that as a feature, but I suspect it would be little used and as such may not be worth the programming time it takes. The authors clearly need to prioritize their time and features and I suspect this may be a pretty low priority since it is apparent from the posts here that few folks would use it.

As such, I am not really opposed to it. I just think it would be a low demand item.

Hey guys,

Sorry for warming this up but I don’t believe this should be so simply dismissed.
I understand the benefits of dithering and yes, more frequently is better.
I image with an Atik314L+ (which is a very sensitive little camera) from light polluted skies. Means my subs are at times extremely short, obviously in broadband RGB. For narrowband I don’t mind dithering every frame but in the case of RGB subs being 15 seconds duration, dithering takes 10 seconds (!) every frame so it almost doubles my imaging time.
I’m not a mathematician but can hardly believe that this outweighs the benefit of shooting more subs and dithering every 3rd or 4th frame???

You can auto focus every xth frame in SGP so why can’t you add that simple feature to the dithering routine?
Appreciate development time is valuable but, you had a few requests for this in the past so it’s not completely irrelevant; cheap capture software like Backyard EOS does it, so why not the way more pricey and complete package SGP? I can’t picture this would take a lot of development time either. Pretty basic command to be sent to PHD.

Thanks for re-considering.

Just bumping this up. Having the ability to dither every x frames would be very useful. I have just migrated from Backyard EOS and I’m missing this feature. With a DSLR, I was able to get away with avoiding darks by dithering every 4th image. In light polluted skies where exposures may not be much more than 60 seconds, the overhead from dithering every frame is substantial. Please reconsider adding this feature.


I would love to have the option to control how often to dither, also.


I’m in on this one. Dithering every frame isn’t necessary and a user-settable dithering frequency could dramatically increase overall imaging time.


I agree it would be nice to have the choice. The issue isn’t about “should” one skip dithering, it’s about options, particularly for those that are comfortable with skipping or simply prefer to reduce the time overhead in an imaging session.

I’ll take a shot at this. Dithering moves hot pixels around so that they can be removed during sigma clipping stacking. Without dithering, hot pixels line up on top of each other and when a mean is taken during sigma clipping rejection, that mean is essentially completely calculated from the value of the hot pixel. When you dither, the hot pixel may only show up in one frame at any pixel location. So the mean will be much closer to the average background, and the hot pixel can be rejected as being outside of the kappa value used.

If you dither every 3 frames, for example, the hot pixel will show up at the same location in two thirds of the frames. But one third of the frames will only have background pixel values. So the mean will be a lot higher than a mean taken on a stack dithered every frame, but it will still be a lot lower than the mean for a stack with no dithering. Hot pixels can still be rejected by selecting a smaller kappa value for the sigma clipping. So, in effect, dithering every few frames can be almost as effective as dithering every frame. The advantage, or course, is that over a night of imaging, the time saved not dithering will increase imaging time. For some mounts, settling time can be substantial, and this could make a meaningful difference.

Another advantage is for those of us who use SGP “off-label” by imaging with two cameras. Since there is no way to synchronize the cameras such that dithering occurs between frames for both cameras, dithering every few subs allows more subs to be collected on the second camera without dithering ruining them.


I’m hoping they will reconsider because I would rather not use the kludgy solution of using SGP for plate solving and focusing but using Backyard EOS for image capture and dithering.

When taking a substantial number of stacks, you still get perfect hot pixel rejection via sigma clipping with dithering every 4 frames.

There was a recent discussion on Cloudy Nights about this:


Another option that allows you to not dither at all and still have all the hot pixels removed is to use your master dark to remove all the hot pixels. This works fabulously well in PixInsight using the Cosmetic Correction feature of the PI Batch Preprocessing Script. Since I image 2 or 3 scopes/cameras on the same mount, I can’t dither at all without it screwing up the 2 image sets on the scopes that are not doing the guiding. My guiding/tracking is good enough that the image shift over 25 images of 8 minutes each is less than 1 pixel. The master dark CC procedure removes all the hot pixels.

This approach is probably not valid if you are going to drizzle your data to improve the resolution.

I think [quote=“spokeshave, post:16, topic:625”]
If you dither every 3 frames, for example, the hot pixel will show up at the same location in two thirds of the frames.
I think this is only true for each group of three images, where the first of the three is dithered from the prior group of three. Example: a total of 60 images is taken with dithering every 3 images. There will be 20 groups of three images that, as a group of 3, will all be dithered prefectly. That means that there are 20 pairs of images that share the same hot pixel with one other image. With a maximum of 2 images sharing a hot pixel with 1 other image out of a total of 60 images, the net result is 1 out of every 30. The end result will be exactly the same as dithering every image. Very comparable to the rejection of airplane tracks or satellites.

I would also add my vote to adding this very simple feature. It would be very beneficial for users that dither.

I’ll look into adding this in. Initially it will be in a very minimal way. Meaning that it will be more “every 3 dithers” rather than “every 3 frames”. Which essentially means that when you change filters the “frame count” won’t be reset. I think this should be acceptable as a first pass as it will still greatly reduce the dithering but will be less horrible to implement :slight_smile: