Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

ROI and 'magnification'

Astrophotography CMOS Planet
  • Please log in to reply
11 replies to this topic

#1 SolarSystem96

SolarSystem96

    Lift Off

  • -----
  • topic starter
  • Posts: 11
  • Joined: 21 Apr 2021
  • Loc: Boyont Beach - FL

Posted 15 May 2021 - 01:28 AM

Hi everyone,

 

Wanting to buy a monochromatic planetary camera, I thought I could use the ROI mode to increase the fps and, in order to better understand the FOV, I went on Stellarium (I chose Jupiter as the planet to shoot, being the bigger) and entered the telescope and sensors measures. And so far so good.

 

From now on I refer to a practical example to make a better idea of my final doubt.

 

  • I'm using a Meade 12" ACF f/10 (D = 304.8mm - FL = 3048mm) from this I know that the theoretical maximum magnification is ~2.5xD = 760x;
  • my theoretical optimal sampling in the visible band (~550nm) is 0.121"/pixel (to be able to apply the Nyquist criterion);
  • the chosen camera is a QHY5L-II because the 3.75um pixels offer me (with a 2x barlow) a resulting sampling of 0.126"/pixel (the closest to the theoretically calculated one);
  • the maximum resolution I can aspire to, determined by the average of local turbulence, is around 2".

 

All that said, to calculate the FOV, given the maximum resolution of (1280x960)pixels of the camera, I calculate the size of the chip or (4.8x3.6)mm, therefore a diagonal of 6mm.

 

Now to calculate the 'corresponding magnification', instead of using the FL as usual for an eyepiece, here I use the diagonal just calculated and it comes out 508x considering there is the 2x barlow (always in order to achieve the optimal sampling), this becomes 1016x which is already much beyond the fateful maximum theoretical magnification allowed by my telescope, previously calculated, but it's still acceptable.

 

Now, however, at (1280x960)pixels, or at full resolution as you want, the camera in question takes a measly 30fps, very few! So I take advantage of the smallest ROI available, looking for the right compromise on the one hand for the highest fps and on the other in order to get the captured planet, e.g. Jupiter, in the FOV.

 

Well, taking into account that Jupiter has an average angular diameter of 39.4" and that the sampling is 0.126"/pixel we immediately obtain ~(313x313)pixel which is therefore the minimum ROI to be adopted to get it in.

 

The problem is that the minimum resolution available for shooting with this camera is only 320x240 at 200 fps (for the height we are in, but not for the width) and so I have to opt for 640x480 even if this means going to only 80.4 fps...

 

But now the fun comes, because so far so good, or almost.

 

Using this lastest resolution I therefore go to calculate, as previously done, the corresponding 'magnification'  finding a diagonal of 3mm which leads to 1016x meaning 2032x with the barlow lens: definitely out of scale compared to the theoretical 760x.

 

Now you will tell me: "Just remove the barlow", but so the sampling is completely wrong, does this mean that I cannot use the ROI?

 

I've thought quite a lot about it, but I can't find a solution.

 

The first thing that came to my mind is that ROI is nothing more than a software tool and that, in essence, the actual sensor size doesn't change, but only the data of fewer pixels is transmitted at the output (even if, unfortunately, the increase of the fps is not proportional) and therefore it isn't necessary to recalculate the size of the sensor which always remains, in this case, of (4.8x3.6)mm. 

 

On the other hand, this means that if I initially take a much larger sensor (e.g. 1600MM Pro), with the same pixel size (3.75um) and the same ROI of (640x480)pixels, I will always frame a region of sky much larger than the QHY5L-II?

 

Maybe I'm getting lost in a glass of water, but this is where the doubts comes out because the reasoning seems counterintuitive and I hope someone could help me, thanks.


Edited by SolarSystem96, 15 May 2021 - 02:36 AM.


#2 Tapio

Tapio

    Skylab

  • -----
  • Posts: 4,302
  • Joined: 24 Sep 2006
  • Loc: Tampere, Finland

Posted 15 May 2021 - 01:44 AM

I think you are overthinking this.
Magnification has no relevance here, as do ROI (other than FPS) .
I believe 2x barlow is good starting point here. And seeing dictates if that is good or you can use more or even that is too much.
I remember that you can use whatever numbers in ROI, not just predetermined.

Edited by Tapio, 15 May 2021 - 01:45 AM.


#3 SolarSystem96

SolarSystem96

    Lift Off

  • -----
  • topic starter
  • Posts: 11
  • Joined: 21 Apr 2021
  • Loc: Boyont Beach - FL

Posted 15 May 2021 - 01:56 AM

Magnification has no relevance here, as do ROI (other than FPS) . 

 

Yes, I know that using the term 'magnification' is misleading so I always put it in quotes.

Instead for the ROI, I don't understand, what do you mean with "has no relevace here"?

 

I remember that you can use whatever numbers in ROI, not just predetermined.

 

Oh really? That is a great news. I still have a lot to learn.


Edited by SolarSystem96, 15 May 2021 - 01:57 AM.


#4 Kokatha man

Kokatha man

    Hubble

  • *****
  • Posts: 15,380
  • Joined: 13 Sep 2009
  • Loc: "cooker-ta man" downunda...

Posted 15 May 2021 - 02:22 AM

...you're certainly over-thinking this! smile.gif (& excuse any typos, I'm typing without my glasses! lol.gif )

 

With the 3.75um pixel sizing in your camera the "rule of thumb" dictates an image scale as calculated for an f ratio of 5x3.75= f19 approximately.

 

Your Meade has a native fratio of f10 meaning you need a barlow amplifying that by just under 2x...many barlows when using just the lens body element will give varying amplification factors from about 1.2x upwards...this chart demonstrates the TeleVue factor which is roughly similar for most barlow makes btw. You could also choose a 1.5x barlow (or something like the Antares 1.6x) & use a bit of extension to get to around f19, although the length of your imaging train might provide that automatically.

 

Televue2XGraphModified.jpg

 

As to the ROI, this is a means of using portions of the sensor's array of rows & columns - by decreasing the height of the ROI one finds that a higher frame-rate (fps) is attainable...if using FireCapture you have default ROI's up to the full-sensor mode, but you can also "draw" custom sizes from the full sensor. (remembering to "centre the chosen dimensions of aforesaid before capturing btw)

 

No experience with this QHY camera btw...so you'll have to reference what fps you might be able to obtain with reduced ROI's.....if you do get about f19 you can work out the image scale in pixels to determine how big an ROI you need to create/select to be able to comfortably fit any planet in said, allowing for any tracking corrections &/or the vagaries of winds etc shifting the image around in the ROI/fov.

 

8bit mode is all that is necessary for planetary imaging & with a lot of planetary cameras this allows for higher fps by virtue of enabling "High Speed" in the capture software. wink.gif



#5 Tulloch

Tulloch

    Mercury-Atlas

  • *****
  • Posts: 2,812
  • Joined: 02 Mar 2019
  • Loc: Melbourne, Australia

Posted 15 May 2021 - 03:03 AM

Hi there, you have certainly over-thought this. Kudos for doing your homework, but it appears you were doing the wrong homework smile.gif.

 

In addition to reading through the advice above (and having a look at the websites mentioned), I would recommend watching these tutorial videos, I found them very useful when I began, the only issue is that Steve's choice of preferred camera has changed since he recorded these videos. Personally I would recommend the ASI224MC or ASI462MC for a planetary colour camera, the ASI290MM or ASI183MM for mono.

http://planetaryimagingtutorials.com/


Edited by Tulloch, 15 May 2021 - 03:11 AM.

  • George Bailey likes this

#6 SolarSystem96

SolarSystem96

    Lift Off

  • -----
  • topic starter
  • Posts: 11
  • Joined: 21 Apr 2021
  • Loc: Boyont Beach - FL

Posted 15 May 2021 - 04:22 AM

I think you are overthinking this.

 

Hi there, you have certainly over-thought this.

 

...you're certainly over-thinking this! smile.gif

 

Am I over-thinking this? lol.gif

 

In any case, I managed to understand the point where I got confused, considering magnification and sampling to be comparable and therefore the mistaken concern of 'overcoming a critical threshold' (i.e. 760x) that in fact does not exist for the latter.

 

(& excuse any typos, I'm typing without my glasses! lol.gif )

 

Don't worry sir. no typos detected cool.gif

 

8bit mode is all that is necessary for planetary imaging & with a lot of planetary cameras this allows for higher fps by virtue of enabling "High Speed" in the capture software. wink.gif

 

Thanks for the valuable advices especially for this here and for the possible modification of the barlow lenses: I have just a 2x of 2" Televue and the graph is very interesting!

 

(and having a look at the websites mentioned) 

 

I don't see any links except yours... Am I going crazy?! confused1.gif bawling.gif

 

In addition to reading through the advice above (and having a look at the websites mentioned), I would recommend watching these tutorial videos, I found them very useful when I began, the only issue is that Steve's choice of preferred camera has changed since he recorded these videos. Personally I would recommend the ASI224MC or ASI462MC for a planetary colour camera, the ASI290MM or ASI183MM for mono.

http://planetaryimagingtutorials.com/

 

I'll watch it asap, thanks. 

 

Regarding the CMOS you mentioned, what do you think of the 1600MM Pro in ROI? 

Because it has 3.75um pixels, which for me would be great in terms of samplig and in addition I already have an excellent 2x barlow that I could use... otherwise I would have to buy everything. 


Edited by SolarSystem96, 15 May 2021 - 04:35 AM.


#7 Tulloch

Tulloch

    Mercury-Atlas

  • *****
  • Posts: 2,812
  • Joined: 02 Mar 2019
  • Loc: Melbourne, Australia

Posted 15 May 2021 - 05:58 AM

 

I don't see any links except yours... Am I going crazy?! confused1.gif bawling.gif

 

 

Regarding the CMOS you mentioned, what do you think of the 1600MM Pro in ROI? 

Because it has 3.75um pixels, which for me would be great in terms of samplig and in addition I already have an excellent 2x barlow that I could use... otherwise I would have to buy everything. 

Sorry, the link was in Darryl's (aka Kokatha man) sig file https://momilika.net/Index.html

 

Do you already have a 1600MM? If so, that would be a great camera to use for planetary if you already own one, probably overkill to buy it new if you were only going to use it for the planets. Don't get too worried about ROI, even on Jupiter you will only get a few hundred pixels on the planet's disc with the Barlow, which is why a small sensor camera is perfect for the planets. If you want to capture the full disc of the moon in one go, then a large sensor is required (183MM or 1600MM if you have one).

 

Watch the tutorial videos - hopefully they will explain things clearly :)

 

Andrew



#8 TOMDEY

TOMDEY

    Cosmos

  • *****
  • Posts: 9,710
  • Joined: 10 Feb 2014
  • Loc: Springwater, NY

Posted 15 May 2021 - 06:21 AM

Hi, 96!

 

Your path is good and normal practice to this point. ~~~ Seeking optimum so run the numbers and ask for opinions. In this case which camera to buy for imaging Jupiter with your scope. Desire to get the best pictures in the least total integrated time e.g. maybe something like an hour a night whenever the conditions are favorable, your availability and mood.

 

And your computations (based on rules of thumb) yield a rather fuzzy answer. --- which is entirely normal!

 

I notice that some of your input parameters pull in different directions. Especially your diffraction-limited theoretical best optical resolution of 0.2 arc-sec vs anticipated best atmospheric-supported resolution of ten times worse. In a sense that helps the choice, because somewhere between will likely be where you will find the most gratifying images without burning yourself out collecting too much data.

 

I'd recommend getting a camera and configure the optics to support a Q in the neighborhood of unity. The chip should be generously oversized, allowing you to then play around with Q, ROI, etc. 2x2 or 3x3 binning might even be on your menu of things to try; just explore trade-space in the field and off the calculator. Sometimes reality and theory can disagree by as much as a factor of two. [Q is a parameter expressing the ratio of pixel pitch to impulse response size.]

 

Implicit Assumptions are most always worth visiting overtly: >>>

 

> The one camera assumption. You might want to consider ordering two or more different cameras. Try them all and settle on the one that actually works best from your yard on the better nights. The others remain available for other conditions or other applications. That way your choice is field-optimized, the cost of the extras justifying finding that optimum empirically. The advantage is that you don't get forever wrapped around the axle in Systems Engineering over-analysis space. This prevents what psychologists call the "Stickleback Dilemma" where that fish hovers above and between, guarding two attractive food dishes... and starves to death guarding both, never actually visiting one or the other. What we engineers and scientists recognize as "Analysis Paralysis".

 

Other implicits:

> different telescope - smaller, larger, better...

> different target(s) - Saturn, planetary nebulae, remote globulars...

> related hobbies - Birding, microscopy     Tom



#9 TOMDEY

TOMDEY

    Cosmos

  • *****
  • Posts: 9,710
  • Joined: 10 Feb 2014
  • Loc: Springwater, NY

Posted 15 May 2021 - 06:40 AM

Oh! I just realized --- It is indeed normal and good to intentionally overthink and overanalyze at the front end. I always do that but also pre-set a "decision date" which could be an hour, day, week, month or year later (depending on the magnitude of the decision and expense in $$$, time, effort and risk). And, as you guys probably have experienced --- It's also normal to analyze the heck outa things --- and then buy what you actually wanted in the first place!    Tom



#10 SolarSystem96

SolarSystem96

    Lift Off

  • -----
  • topic starter
  • Posts: 11
  • Joined: 21 Apr 2021
  • Loc: Boyont Beach - FL

Posted 16 May 2021 - 07:41 PM

Sorry, the link was in Darryl's (aka Kokatha man) sig file https://momilika.net/Index.html

 

Thanks, precious images and advices on here.

 

Do you already have a 1600MM? If so, that would be a great camera to use for planetary if you already own one...

 

Yessir: you're officially the first person who tells me this, until now I have met all people who denigrated it at the thought of a planetary imaging use! 

 

Hi, 96!

 

Your path is good and normal practice to this point. ~~~ Seeking optimum so run the numbers and ask for opinions. In this case which camera to buy for imaging Jupiter with your scope. Desire to get the best pictures in the least total integrated time e.g. maybe something like an hour a night whenever the conditions are favorable, your availability and mood.

 

And your computations (based on rules of thumb) yield a rather fuzzy answer. --- which is entirely normal!

 

You couldn't have described my mind process in a better way. At the same time I think it can only be done by those who have already been there.

 

just explore trade-space in the field and off the calculator. Sometimes reality and theory can disagree by as much as a factor of two. 

 

An important and always true remind, thanks.

 

> The one camera assumption. You might want to consider ordering two or more different cameras. Try them all and settle on the one that actually works best from your yard on the better nights. The others remain available for other conditions or other applications. That way your choice is field-optimized, the cost of the extras justifying finding that optimum empirically. The advantage is that you don't get forever wrapped around the axle in Systems Engineering over-analysis space. This prevents what psychologists call the "Stickleback Dilemma" where that fish hovers above and between, guarding two attractive food dishes... and starves to death guarding both, never actually visiting one or the other. What we engineers and scientists recognize as "Analysis Paralysis".

 

I think I got it very well even if still am only an Astrophysics student: I knew the 'Buridan's donkey paradox' version.

 

Oh! I just realized --- It is indeed normal and good to intentionally overthink and overanalyze at the front end. I always do that but also pre-set a "decision date" which could be an hour, day, week, month or year later (depending on the magnitude of the decision and expense in $$$, time, effort and risk). And, as you guys probably have experienced --- It's also normal to analyze the heck outa things --- and then buy what you actually wanted in the first place!    Tom

 

I have only read now, so I confirm what I wrote above grin.gif

 

Just one last question:

what formula should I follow to obtain an optimal focal length based on the sampling?

 

Because the book that I follow contains formulas, which give very different results from those I find on the net and I'm going crazy to understand which is the right one!

 

Manuel


Edited by SolarSystem96, 16 May 2021 - 08:39 PM.


#11 Kokatha man

Kokatha man

    Hubble

  • *****
  • Posts: 15,380
  • Joined: 13 Sep 2009
  • Loc: "cooker-ta man" downunda...

Posted 16 May 2021 - 08:14 PM

Just one last question:

what formula should I follow to obtain an optimal focal length based on the sampling?

 

Because the book that I follow contains formulas, which give very different results from those I find on the net and I'm going crazy to understand which is the right one!

 

Manuel

Hi again Manuel - it sounds like you need to try & correlate various theoretical appraisals for determining imaging parameters, nothing wrong with that whatsoever waytogo.gif ...although your <"At the same time I think it can only be done by those who have already been there."> acknowledges that wheels don't need to be constantly re-invented! wink.gif

 

The 5x the pixel sizing in the sensor is a very tried & true maxim...seeing always has the final word & often the seeing will not support anywhere near optimum resolution, although it is rare indeed ime to find good seeing supporting more than what this rule of thumb suggests.

 

My former comments about working with an image scale using slightly less than that which a 2x barlow achieves still holds here: all the images on our website (in siggy below) were using this formula...but going back to slightly earlier times I did conduct many experiments to validate this figure. wink.gif


  • SolarSystem96 likes this

#12 TOMDEY

TOMDEY

    Cosmos

  • *****
  • Posts: 9,710
  • Joined: 10 Feb 2014
  • Loc: Springwater, NY

Posted 16 May 2021 - 10:22 PM

This chart sorta relates to the greater context of camera configuration for imaging. I created if for "Popular Audience" talks that I gave here and there. I worked ~Remote Sensing~ R&D most of my career: Radio, Radar, Imagery, Lasercom, Holographic Interferometry, Imaging Sats, Drones, etc. It's that Q in the upper left corner that is implicitly/properly dominating our conversation here. All those equations and references are just different aspects of Q-Selection/Optimization. And that is most affected by thoughtfully answering the question. "What are you trying to accomplish?"

 

In my fantastic cartoon, we Earthlings are attempting to accomplish ~defense from attacking Martians~. So the Q of the imaging satellite is chosen to optimize the Actionable Information content of the downlinked images.

 

Here on Cloudy Nights, we mostly want to accomplish ~impressive, pretty images of the heavens~. Boundary conditions would probably be like e.g. "Best that I can achieve with my telescope dedicating a week to the selected target." And implicit is that is to have self, friends, family, and enemies all exclaim "Wow - that's fantastic!" Nothin wrong with those Mission and Vision Statements!    Tom

 

~click on~ >>>

Attached Thumbnails

  • 28 remote sensing Toms chart.jpg

  • SolarSystem96 likes this


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics





Also tagged with one or more of these keywords: Astrophotography, CMOS, Planet



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics