Mini-Histogram Stretch Button

[MarMax]

Robin,

When doing DSO live stacking with an OSC camera and using the Live Stack Histogram it can be difficult to see the fine adjustments to RGB, especially when using the ±0.1dB changes. With Lock Stretch ON the Mini-Histogram RGB curves can provide a better visual for fine adjustments but it's a fairly small graph to use as-is.

You added a Horizontal Zoom to the regular Histogarm a while back and I wonder if a similar feature can be added to the Mini-Histogram. Possibly another button below the Lock Stretch button to expand the horizontal axis for making fine RGB adjustments while live stacking with Lock Stretch ON.

Thank you,
Mike

[clouzot]

+1 on this request, or anything that could make the Mini Display Histogram more readable. It's my favorite tool for fine tuning the color balance: as the livestacked image is already stretched, the "Mini" will readily show any color imbalance there may be on the sky background, which is a good neutral reference in a majority of cases. The Livestack Histogram displays the unstretched data distribution, so it's far more difficult to see what's going on.

Another possibility could be to have an extra "Display after stretch" checkbox on... the Livestack histogram itself. But now that I think of it, that would restrict the use cases to live stacking only.

[mikenoname]

When doing DSO live stacking with an OSC camera and using the Live Stack Histogram it can be difficult to see the fine adjustments to RGB, especially when using the ±0.1dB changes. With Lock Stretch ON the Mini-Histogram RGB curves can provide a better visual for fine adjustments but it's a fairly small graph to use as-is.

Mike,

I guess I'm about to learn something because I don't understand either of the above sentences.

When I make RGB changes in the livestack histo, they are very apparent as they move the associated R, G or B line. And the display histo is so small that absolutely nothing is easier to see on it than the much larger live stack histo. So I'm scratching my head here...

Mike

[mikenoname]

The Livestack Histogram displays the unstretched data distribution, so it's far more difficult to see what's going on.

Ben,

Could you explain the above? I thought when you stretched the livestack histo, it showed the stretched curves?

Mike

[admin]

Hi,

for reasons lost in the mists of time, the live stack histogram does reflect the adjustments made by using the vertical R/G/B sliders to adjust colour balance of the stack, but does not include the effects of the live stack stretching. That means that the live stack histogram itself can be used to align the colour channels if you are trying to do colour balance by aligning peaks, etc. I don't know if that helps with the situations outlined above or not...

The small size of the mini histogram makes it hard to add new functionality to it - there just isn't space for much more to be added. One thing that might be possible is a 'zoom on hover' effect - there is a similar effect on the drift graphs in the live stacking area where they pop up to a larger size if you hold your mouse over them. Would that be useful?

One long standing issue with live stacking is that if you are *not* using darks then there is a constant background from the camera offset level in all data, which is affected by the colour balance adjustments - this can make it hard to adjust the colour balance successfully for both the background and brighter objects in the frame. One day I will think of a good way to deal with this and implement a fix - for now it is something to be aware of...

cheers,

Robin

[MarMax]

When doing DSO live stacking with an OSC camera and using the Live Stack Histogram it can be difficult to see the fine adjustments to RGB, especially when using the ±0.1dB changes. With Lock Stretch ON the Mini-Histogram RGB curves can provide a better visual for fine adjustments but it's a fairly small graph to use as-is.

Mike,

I guess I'm about to learn something because I don't understand either of the above sentences.

When I make RGB changes in the livestack histo, they are very apparent as they move the associated R, G or B line. And the display histo is so small that absolutely nothing is easier to see on it than the much larger live stack histo. So I'm scratching my head here...

Mike

I'm learning as I go as well. From the manual, adjustments to the Black Level, Mid-Level and White Level do not affect the shape or position of the histogram shown in the LIve Stacking panel. When I'm making the final small adjustments to an image prior to "saving exactly as seen" (and suitable for posting in the CN EAA Forum), I'd like to use the Mini-Histogram to make the final RGB and stretch adjustments. I've been using histogram stretch 6 quite a bit lately and can use every visual aid available to dial in RGB balance and final stretch settings.

I've found that the image on my laptop screen is not the best visual aid to do this so I'm looking to use the histograms and the "hover zoom" feature Robin is suggesting may help to see how final stretch adjustments impact the Mini-Histogram. I suggested a zoom horizontal scale button because you will need to make final stretch level adjustments while observing the zoomed Mini-Histogram.

Robin> If you feel there is no value added in using the Mini-Histogram in this way you can disregard this feature suggestion.

[mikenoname]

From the manual, adjustments to the Black Level, Mid-Level and White Level do not affect the shape or position of the histogram shown in the LIve Stacking panel.

for reasons lost in the mists of time, the live stack histogram does reflect the adjustments made by using the vertical R/G/B sliders to adjust colour balance of the stack, but does not include the effects of the live stack stretching.

Oh that's interesting. I never considered the effects of stretching making a difference to the shape or position of the histo curve. I've always just played with the color balance. So I guess what you're saying is that if you do a stretch on the display histo, then the live stack histo should march to the right indicating it has brightened? Is my thinking right there? If so, that would help with faint objects whereby the histogram peak is off the left side of the chart and bring the peak of the curve to the right and allow unclipping of the black levels...

- M

[clouzot]

Oh that's interesting. I never considered the effects of stretching making a difference to the shape or position of the histo curve. I've always just played with the color balance. So I guess what you're saying is that if you do a stretch on the display histo, then the live stack histo should march to the right indicating it has brightened? Is my thinking right there?

Mike,

I don't think so. The Display Histogram does what it says on the tin: it stretches the display, and the display only; it doesn't modify anything in the underlying live stack data, so the Live Stack Histogram should (and does) stay the same.

for reasons lost in the mists of time, the live stack histogram does reflect the adjustments made by using the vertical R/G/B sliders to adjust colour balance of the stack, but does not include the effects of the live stack stretching. That means that the live stack histogram itself can be used to align the colour channels if you are trying to do colour balance by aligning peaks, etc. I don't know if that helps with the situations outlined above or not...

I think I remember the reasons, but it's been a long time. It all started while Sharpcap was still in v2, with a bunch of piteous French EAA-ers asking for some kind of display to help them tune the colo(u)r balance. We were already using the Mini Histogram as a monitoring tool, but wondered if there could be some better form of visual feedback. In addition, the raw livestack data being by definition unbalanced, the Sharpcap v2 LiveStack histogram would show three successive peaks (one per channel), and it was confusing to understand how to place the black/mid/white points.
If I remember well, you explained that introducing a "post-stretch" Live Stack histogram display would further confuse users, so the fallback solution was to only apply the color transformations to histogram.

--Ben

[admin]

Hi,

yes, Ben is right - the mini histogram shows the histogram of the image the comes out of live stacking (with live stack stretch applied), and the stretct in the mini histogram affects the display only - there is no upstream effect from the mini histogram/display stretch on live stacking.

Thinking again about the reason the colour balance affects the live stack histogram - the live stack histogram helps you set up the right stretch level between input brightness and output brightness. Colour adjustments happen *before* the stretch while the image is still being processed linearly, so they count as part of the *input* to the live stack stretch, so it makes sense for them to show effects in the histogram graphs.

cheers,

Robin

[MarMax]

With all this discussion so far I'm still not sure if the Mini-Histogram can be a final adjustment measurement tool.

I've reached my "level of detail" limit visually using the laptop image and Live Stack Histogram. That is, I'm not able to discern which final 0.1dB R-G-B adjustments best align the curves. Each final 0.1dB adjustment can make a noticeable visual change to the displayed image yet I'm still not sure if I've best aligned the curves.

I have a decent ear for music but for art and colors I much prefer to use a tool like a caliper or voltmeter to know where I stand. If the Mini-Histogram can provide measured detail to best align R-G-B it may useful.