Re: Stall and Control box
|
Hi Tim, As I discover more about getting better planetary images, I will give that advice freely.? ?And I do have some new things to share...??here is what I discovered: First of all...the photo attached is surely my best final imsge.? Again, I used the C14EdgeHD ... manually focussed with the telescope focus knob (expecting a motor focusser to come in later this week).? ?Seeing was better than average that night, but not as perfect as Clear Sky Chart sometimes shows rarely.? That picture should have turned out badly on a 14 inch scope.? But it didn't.... ?It was shot on the ZWO ASI224 camera (3.75 um pixel size) at f/11... With a ZWO ADC atmospheric dispersion corrector in place in front of the camera.? It was shot at prime focus with a 1.25 inch visual back...no Crayford for fine focus. A nice Crayford I have put the camera image plane too far back of the 5.75 inch recommended image plane distance.? In that scope you can be inside of 5.75 inches, but image sharpness gets worse past 5.75 inches.? It's a critical element of the EdgeHD scopes that I did not know when I bought it (used earlier this year).?? At the time of that video recording, I did not use a UVIR filter and so about 20% of the Red light was from long wavelength IR.? I only discovered that difference the following night...so use a UVIR filter to improve contrast.? I now use that UVIR in the cameras front nosepiece. There were a few "rules of thumb" that were told me years ago when I first tried planetary imaging.? They were: 1. Rule: Only try imaging when the Clear Sky Chart shows excellent (darkest color bars) for seeing.? I find this to be wrong:? I'll explain.? The Clear Sky Chart is not able to say that there may be short windows of excellent seeing maybe 30 minutes or so, between periods of turbulent air.? It just gives the overall prediction for its time window.? Sometimes you get very lucky if you are set up and shooting... And sometimes it predicts perfect seeing and that wish does not come true either. It really is "lucky imaging." [ See the incredible images from? Anthony Wesley of Australia....here: Anthony Wesley explained that his best image came from a night he was not expecting great seeing, but then, the sky stabilized and...voila.? (He has a fabulous scope, a Robert Royce 16 inch Newtonian mirror...essentially flawless.? That sure helps!) ] 2.? Rule: You must use the f# that is "5x your video camera pixel um size". You are supposed to use a 2x or 2.5x (best is a Tele vue Powermate) to get the f#.? Example: on my ASI224MC camera, with it's 3.75 um pixels... this means I should be trying to get f/18.5.? so by that rule, I should have used a 2X tele to get my f11 scope to be f/22.? Yet I find this wrong too...I got the incredible image attached using that ASI224 at f/11.? ? Why did f/11 and no 2x tele work??? When you use a 2X tele, you get 1/4th the light into each pixel.? So to get the exposure to be say 80% of full, you need 4X the exposure time.? If the seeing is so-so not perfect, the longer exposures blur out the image frame.? Using no 2X at all gives a 4X faster exposure so can "freeze" the seeing and get you clearer frames. And for the same exposure time, say 180 sec, you get 4X the number of frames without the 2X tele.? I got 44,000 frames in each of the 180 sec exposures.? autostakkert3 (AS3) can throw out a lot of frames and have plenty to stack. 3. Don't shoot more frames than like 60 sec else the planet Jupiter will spin and the frames at the front of the video won't stack with the rotated frames at the end of the video.? No longer true: The free program WinJuPos solves this, by creating a replacement video that "de-rotates" the frames.? I do not understand yet how it performs this magic trick...but it does.? (I'll send a separate document on how to use WinJuPos.) Last night I tried a different ASI camera.? The ASI224MC has 3.75 um pixels, so ideally (by the rule of thumb) requires f/18.5 .? The ASI178MC has smaller 2.4um pixels...so by the rule requires f/12.? My scope us f/11...that's close enough to f/12 that I don't need the 2X tele, and my exposures can be faster. Also the ASI178 is made to use Backside Exposure.? This put the light into the pixel silicon region with no metal wiring to block the photons.? The 224 uses frontside exposure and has the metal pixel connections in the way if some of the light. The 224 is more sensitive to red and far less sensitive to blue. The 178 is more sensitive to blue than red due to the backside detection.? I have yet to process my images of last night but I think they were better.?? I used to take maybe a few thousand frames.? Now I see the benefit of taking a huge number of frames.?? Rule: use the deNoise capability of Registax wavelets:? in Registax6 wavelets, I used to only use Linear Gaussian and I was putting the DeNoise number to be 2X the Sharpen number.? What that did was to blur out the final image... it certainly denoised it, but too much.? Now I use the Default Linear wavelet setting and there is no DeNoise being applied.? For my present work that works better. I follow up the Registax6 TIF 16 bit output with Photoshop (CS5).? Then I use the Image Size tool to expand to 200% of the original.? I use the Bicubic_Smoother selection as Photoshop says "best for enlargements." Then I copy layers,? use the Unsharp Mask at different settings like 1.7 pixel or 0.9 pixel, to gradually sharpen up the final image.? I also may have to look for noisy areas of the image, Mask them and use the noise reduction tools.?? In photoshop, after you generate your final image, flatten it. Then use the Channels tab to look at the Red, the Green, and the Blue separately as B&W images.? The blue likely has the most noise: the air scatters blue light much better than it scatters Red light.? It's why the sky is blue!? You can then use the Noise reduction tools in Photoshop to reduce noise in the Blue channel.? Likewayse examine the green and the Red channels.?? I also found that if I sharpen too much in Registax6, any colorful moons floating in the picture can become totally white and blown out.? RS6 is not smart enough to rescale the pixel intensity so it remains within the gamut.? (That seems to me a silly processing problem...should have been easy to program it to stay in gamut bounds...but it doesn't. ) So I process the same AS3 output image twice in RS6...once for the planet sharpening, on slightly less sharp for the moons process.? Then I combine the two RS6 images in Photoshop using small masks to get the colorful moons in with the sharper planet image.? ? Hope this stuff helps you...but I need to write another document for you...showing the steps with examples... Stay well and enjoy life and health, Michael On Sun, Jul 19, 2020, 8:26 AM Astronut <hg2u@...> wrote: Hi Michael, |
Michael Herman