StarTools 1.8.518 Beta 1 now available

[jackbak]

Good Job, Ivo!

SVDecon with my problem child M8-Lagoon is much better - can actually use SVPSF.

HDR is very good as well.

Small niggle - HDR module "Gamma Smoothen" could be more fluidly named "Gamma Smooth" (maybe that's just my bastard American)

Thanks,
Jack

[Mike in Rancho]

Awesome! Thanks Ivo.

Take a well-deserved break while we all try to break 515.

Perhaps Gamma Smoothie? j/k, probably Smoothing sounds best to my ear. More importantly, much new stuff to learn and compare with old HDR.

Seemed good on my first tests - I threw some HDR-needy data at it like M57 and the West Veil/Triangle. I did end up with an odd little gray pixel right in the center of 52 Cyg...

And while my hardware monitors might not be the best, is the initial calculation / context change using enough GPU? It does chew on things for a while, but seemed to be mostly heavy CPU.

[Mike in Rancho]

Small note: I think the log for 515 needs a little clean-up for the new HDR. I don't have it in front of me since it was on the other computer, but I think two parameters are marked as [UNKNOWN] (or something like that), and another one or two are mislabeled or transposed. Easy enough to figure out, but will require some minor editing for any log posts to forums and so forth.

[admin]

Small note: I think the log for 515 needs a little clean-up for the new HDR. I don't have it in front of me since it was on the other computer, but I think two parameters are marked as [UNKNOWN] (or something like that), and another one or two are mislabeled or transposed. Easy enough to figure out, but will require some minor editing for any log posts to forums and so forth.

Indeed. @Guy also noticed this. Nothing escapes you guys!

[admin]

The HDR log entries being scrambled is now fixed in 1.8.516 (now up for download). A "Quality" parameter was also added to the HDR module to speed up local gamma correciton solving, at the expense of some precision.

[LuckyEddie]

One thing I have noticed with the new 1.8.516 alpha is I now sometimes have an immediate crash with no message on the command line or entry in the log.
Once triggered this is a hard failure on every program launch. A complete reboot always seems to clear the problem, so I assume it's some sort of unreleased resource issue, however I haven't seen this particular type of crash with any previous version of ST. Are there more/different start-up checks happening now?
Running Linux64-GPU.

[admin]

One thing I have noticed with the new 1.8.516 alpha is I now sometimes have an immediate crash with no message on the command line or entry in the log.
Once triggered this is a hard failure on every program launch. A complete reboot always seems to clear the problem, so I assume it's some sort of unreleased resource issue, however I haven't seen this particular type of crash with any previous version of ST. Are there more/different start-up checks happening now?
Running Linux64-GPU.

I have seen this as well on Linux distros that run older Intel iGPU drivers, when they resume from sleep. The OpenCL portion of the driver does not seem to initialize.
Is that what is happening?

You may be able to find some entries in the kernel ring buffer (use 'dmesg' command).

[Stefan B]

Finally I had the opportunity to give the 1.8.516 a go with the new HDR. It's great in my eyes! Adjusting things with the sliders worked really well. Choosing the medium option was sufficiently fast and lead to a good result.

But I also had some problems. ST crashed twice during usage of SVDecon during sampling. Thought that was a problem of the past but it showed up again. In order to avoid it I had to sample the stars really slowly.

Second issue had to do with color. I tried a image with an l-eNhance stack as luminance and broadband data as RGB (everything captured by a DSLR). I wasn't able to get rid of some green in the image. Neither with the slider to reduce green nor with the Cap green option. Looking at the MaxRGB image confirmed that green dominant pixels remained. Will try to have a look at a former 1.8 alpha with which I didn't encounter the issue. Maybe it's also my new data... Anyone who also saw this?

Regards
Stefan

[PaulE54]

Finally I had the opportunity to give the 1.8.516 a go with the new HDR. It's great in my eyes! Adjusting things with the sliders worked really well. Choosing the medium option was sufficiently fast and lead to a good result.

But I also had some problems. ST crashed twice during usage of SVDecon during sampling. Thought that was a problem of the past but it showed up again. In order to avoid it I had to sample the stars really slowly.

Second issue had to do with color. I tried a image with an l-eNhance stack as luminance and broadband data as RGB (everything captured by a DSLR). I wasn't able to get rid of some green in the image. Neither with the slider to reduce green nor with the Cap green option. Looking at the MaxRGB image confirmed that green dominant pixels remained. Will try to have a look at a former 1.8 alpha with which I didn't encounter the issue. Maybe it's also my new data... Anyone who also saw this?

Regards
Stefan

Hey Stefan.
As regards your first problem, I've also been having this (discussed on another thread). My experience agrees with yours - crashes during SVDecon unless you sample really slowly / take breaks after selecting a few points. In my case I was running task manager while testing this issue and could see memory utilisation hitting very high values on my 16GB Win 10/64 desktop. Leaving the machine for a minute or two after selecting 5-6 points, and then resuming use of Decon, allowed ram utilisation to drop to normal values. Does this happen to you also?
And yes - new HDR is great!
Paul

[Stefan B]

Hey Stefan.
As regards your first problem, I've also been having this (discussed on another thread). My experience agrees with yours - crashes during SVDecon unless you sample really slowly / take breaks after selecting a few points. In my case I was running task manager while testing this issue and could see memory utilisation hitting very high values on my 16GB Win 10/64 desktop. Leaving the machine for a minute or two after selecting 5-6 points, and then resuming use of Decon, allowed ram utilisation to drop to normal values. Does this happen to you also?

Hi Paul,

I didn't check the utilisation, maybe I'll have the opportunity this evening (running my 8GB Win10/64 notebook). As already noted this happened with some of the first 1.8 alpha versions but wasn't a problem with intermediate versions. Unofrtunately it's showing up again. Don't know if it's the same cause or if it's an independent problem with the same outcome...

Regards
Stefan