I've primarily noticed a difference from the RAM disk in Dev and Autodev, and Bin. As you said, Ivo, the larger the image, the bigger the difference...no surprise there given that the amount of info being swapped is a direct function of the number of pixels involved.
Just finished overclocking this critter to 4ghz, and there seems to be a noticeable, but not "holy cow wow" improvement, mostly in rendering of previews and such. Again, not a surprise given that's where a bunch of the math is occurring.
Building a new system
Re: Building a new system
Sounds good Ivo! Happy Holidays!
Re: Building a new system
Took some time today to do my first "real" process with the new machine.
This is M2 from a set of data taken on 2012-08-21.
(Click to view AstroBin tech card and larger resolution)
I'd previously given up hope on this data, there's some HORRIBLE gradients (I'll be uploading an unstretched TIF shortly, link will be in the AstroBin card). Rummaging around this morning for some data I hadn't tried in ST yet (it appears I'll have time to try everything HST has ever shot in St before the weather clears) and decided to give this one a shot.
ST Log :
The new machine just FLEW through this thing. Wipe was notably quicker than before, and I was able to work comfortably with a much larger image.
The new machine is definitely a success.
This is M2 from a set of data taken on 2012-08-21.
(Click to view AstroBin tech card and larger resolution)
I'd previously given up hope on this data, there's some HORRIBLE gradients (I'll be uploading an unstretched TIF shortly, link will be in the AstroBin card). Rummaging around this morning for some data I hadn't tried in ST yet (it appears I'll have time to try everything HST has ever shot in St before the weather clears) and decided to give this one a shot.
ST Log :
- Binned 71%
- Cropped 50 pixels off each edge.
- Develop Gamma .94, DAF 5.3 pixels, all others defaults.
- Wipe started with the Gradient preset, Precision 512x512, Aggressiveness 98%
- Wipe again with the Gradient preset, to attack the 2nd gradient, Aggressiveness 95%
- Basic star mask, unselected most of the core, inverted.
- Filter, Reject mode, started cleaning up most of the gradient left in the lower left corner.
- Contrast
- Sharpen
- Mask, Threshold 85, Highlights>Threshold
- Color, Sample, Fill Mask, Saturation 125%
- Stop tracking, Denoise defaults
- Created a Star mask, tweaked to split multiple star groups and deselect much of the core
- Repaired with Warp
- Magic Shrink Core
- Color saturation 115%
- Synth - Refractor - Adjusted to taste
The new machine just FLEW through this thing. Wipe was notably quicker than before, and I was able to work comfortably with a much larger image.
The new machine is definitely a success.
Re: Building a new system
Hi Gordon,
What a great shot & thank you for sharing the flow. How did you do the annotation?
Nice job! The stars are very punctual and the colors are right on. Nicely done!
M.
What a great shot & thank you for sharing the flow. How did you do the annotation?
Nice job! The stars are very punctual and the colors are right on. Nicely done!
M.
Re: Building a new system
Thanks for the kind words, Miquel
Annotation was done by astrobin. It platesolves any image you upload, if you request it.
Annotation was done by astrobin. It platesolves any image you upload, if you request it.
Re: Building a new system
Thanks for the tip on astrobin. I appreciate it.
I noticed one more thing on your flow. How did you split group stars? I saw that in one of Ivo's videos but I'm not clear on the process. Thanks for your tips & advise. M.
I noticed one more thing on your flow. How did you split group stars? I saw that in one of Ivo's videos but I'm not clear on the process. Thanks for your tips & advise. M.
Re: Building a new system
I just did it manually. Zoom in on a pair of stars covered by one "blob" of mask, set the Brush Mode to Single Pixel Off, and just draw a line of unmasked pixels between the two stars.
Does that make sense?
Does that make sense?
Re: Building a new system
Yes, I think so. Thanks a lot Gordon!