Proposal: Replace Gamma with MTF Stretch

[alexhelms]

Would there be any interest in replacing the gamma slider with an MTF stretch, aka "auto stretch", that is used in Pix Insight? This stretch tends to be more reliable and "set it and forget it". No need to mess around with a slider for different exposure settings, changing conditions etc.

If there is interest in adding this feature, I'm willing to implement and submit a PR for review.

Proposal

• Checkbox in Global tab for Auto Stretch, off by default.
• When Auto Stretch is enabled, gamma slider on main UI is hidden.

Thanks for considering this feature.

[EthanChappel]

Bumping this because I also think it would be a great addition to PHD2. In fact, I think it would be better to have it enabled by default.

[bwdev01]

We don't really have any interest in this because is provides zero benefit to guiding performance while adding more processing overhead in critical parts of the code. The only result is "eye candy" at the expense of overhead which isn't a trade-off we want to make - and it encourages the notion that users need to pay attention to the appearance of stars in the real-time display. The trend we've seen over the recent past is the need for faster cadence guiding with strain-wave mounts and the use of smaller, less powerful computing platforms - all of which encourages lower processing overhead, not more. Finally, there is really no such thing as a one-size-fits-all algorithm when you take into the account the huge range of guide cameras, exposure times, optics, image scales, and sky conditions. And the more you have to do to tune non-linear stretches for these situations, the higher the overhead.

[NoahTingey]

One thing worth considering: the image can be heavily downsampled before computing the MTF stretch parameters with essentially no quality penalty for the purposes of PHD2's image display. The median/MAD computation becomes trivial at that point, and it's a much smaller cost than a lot of things already running in the PHD2 pipeline. It's basically just a LUT.

On "one-size-fits-all" -- PixInsight's autostretch is a decent reference for robustness. It handles the range of cameras, exposures, and conditions more reliably than the gamma slider, which in my experience needs manual adjustment whenever anything changes (and sometimes seemingly even when things haven't changed for some reason). If you're worried about edge cases, I think the existing gamma slider could be reused to adjust an MTF stretch to cover edge cases -- though I suspect most users wouldn't need to tweak it from a default value.

Could optionally have a switch to toggle between MTF stretching vs the current gamma slider if users experience slowdowns on very low-power machines.

Appreciate you explaining the team's thinking.

[bvalente22]

Noah

What's the thinking on why a stretch is important enough to add mtf stretch?

referring back to Bruce's comment re: 'The only result is "eye candy"' and 'it encourages the notion that users need to pay attention to the appearance of stars in the real-time display.':

What use case do you see that the user benefits from mtf stretch, knowing that the result is strictly visual, with no impact to actual guiding?

[NoahTingey]

If anything, I'd say the existence of a gamma slider that needs frequent tweaking if the user wants to take a look at their guide stars is itself encouraging the notion that users need to pay attention to star appearance in the real-time display. An MTF stretch that provides good results without any input from the user would avoid the potential for misconception entirely -- if it just works, there's nothing for users to feel like they need to manage.