Aperture, F-ratio myth and telescope basics

[codemonkey]

I've never really stopped to think too much about how telescopes work and how the combination of focal length and aperture work together to produce images of a certain brightness and field of view, so now that I am thinking about these things, I find I have questions.

Stan Moore has an interesting article on the F-ratio Myth (http://www.stanmooreastro.com/f_ratio_myth.htm) basically saying that varying the F-ratio does nothing for exposure time, only varying the aperture makes a difference.

On the face of it, this makes sense... focal length controls FOV, aperture controls the amount of light. Bigger aperture means a bigger light bucket and thus more light.

But why does a bigger aperture mean more light from the same source? Shouldn't it mean more light, but from different sources? Take a point light source, it doesn't seem like it should matter whether you have a 80mm aperture or a 500mm, it's a point source, so it seems like you should either get it or not.

Does it work differently for point light sources and diffuse light sources?

[Atmos]

Basically, point light sources radiate in all directions and not just directly at anything in particular. Because they radiate in all directions, a larger aperture is able to capture more of that "all directions". If I wasn't at work I would draw a diagram on my white board, could do that when I get home :-)

[Stefan Buda]

The author of the article you mention deserves a jail sentence for causing a great deal of confusion among astrophotographers.

There is no F-ratio myth. The laws of optics are clear and simple.

[glend]

So why are there all these guys running reducers to get lower ratios, when they are obviously the same aperture on their one scope. :question:

[multiweb]

Damn!... I'd better list my C11-hyperstar in the classified then. 30s subs at F/2 vs 15min subs at F/10 is not good enough ;)

[RickS]

The way I look at it is that only two optical parameters matter: clear aperture (amount of light collected) and image scale (how thinly that light gets spread over the sensor.) Seems pretty simple...

[astroboof]

Hi there,
Figure aperture to light like you might a dam full of water- (the light source), and a hole in the bottom of a weir (the aperture), thus, the bigger the hole, the larger the flow of water out of it. I know that doesn't answer your query sorry.

Or perhaps, electricity supply and an appliances draw on it, if you'd rather.

Also apertures effect can be diluted by a FL, but the effect is far from linear or standard, or practical sometimes for loads of reasons.

Rough analogies, but things are rarely absolute or black and white, especially when dealing with the EM spectrum, where nothing is ever wasted or used, just converted into another form of energy. The laws of optics are only clear and simple if you like to order your mind with absolutes, for your own convenience. Truth is the many variables behind optical performance would be near endless is you stopped to consider all the interactions.

Years ago I had to do some of this with math for imaging optimising in post-grad Uni astro studies for our optical trains, it was a real eye-opener, and a pain.

If optical physics prompts you to ask more questions than it does answer them, than you are on a track to truth, which is a pandora's box to say the least.

As for us hobby imagers, if it does the job and we get satisfaction from our efforts, then it is good enough and worthwhile imVho

Cheers and all the best. :)

[Atmos]

The only determinant factor is aperture for capturing photons. The focal length and therefore focal ratio just change the way that the captured photons interact with your camera.

[Peter.M]

I agree completely. Except then you need to consider obstruction, reflectivity or transmission of every surface, ect ect. I think it's best to just judge equipment on aesthetic value of the images they produce. I can find hundreds on fsq images that I think are high quality. The same is not true for c11 hyperstar systems.

[julianh72]

The last two comparison images at the bottom of the page http://www.stanmooreastro.com/f_ratio_myth.htm are very misleading. If one was indeed taken at f/12.4 and the other at f/3.9, then they have been scaled differently, or were taken with different sized sensors. For example, if the f/12.4 image is the full capture over the sensor, then the f/3.9 image is roughly a 1/3 x 1/3 crop of the whole image. (Either that, or the sensor used for the f/3.9 image is only 1/3 the size of the f/12.4 image sensor.)

Using a focal reducer on a long focal length telescope lets you capture a bigger sky area on the same sensor, but doesn't alter the brightness of point sources.
[Edited to correct an overly general statement]

[Peter.M]

You are wrong. It does alter the brightness of extended sources. Because now at a lower sampling rate more area of sky fits into each pixel

[gregbradley]

A reoccurring question.

I imagine all a focal reducer is doing is focusing more of the collected light onto the sensor that normally misses it and is wasted and so there is a gain and also gives a wider field of view where more light sources are emitting.

Greg.

[multiweb]

2 degrees field worth of light on a chip is worth more than 30min field in the bush ;)

[rustigsmed]

I guess it gets confusing when the FOV changes all the time, generally when you change aperture the FL generally changes too.

However in this is perhaps a good example?

an 8" f5 newt (FL 1000mm) vs a 10" f4 newt (FL 1000mm) same field of view an extra 2" light gathering power hitting the sensor in the same spot, makes sense the 10" (or f4) is "quicker".

Russ

[Stefan Buda]

Looks like you guys missed the recent and excellent post on this topic by Ray:
http://www.iceinspace.com.au/forum/showthread.php?t=136008

I said post rather than thread because there is no need to read past the first post.

It should have been nailed to the top but it hasn't even been pinned..

[Shiraz]

Hi Lee. Stan Moore's article is very misleading, because in debunking one myth he inadvertently created a new and even more insidious one. The problem seems to be that he neglected to mention the key part of his argument which should be "F-ratio makes no difference to sensitivity for a given aperture IF THE PIXEL SIZE IS ADJUSTED TO COMPENSATE FOR THE CHANGE IN FOCAL LENGTH". The bit in caps is the bit he left out and now people continue to quote the first part of the argument and assert that F-ratio makes no difference - even experienced photographers who surely must know that it clearly isn't true. Suggest that you ignore that paper.

the things that matter are the aperture, which determines how many photons get in and the angular size of the pixels, which determines how many of the available photons end up in each pixel. if you fix the focal length and pixel size, the sensitivity increases with reducing FNo as Russ' example shows. This applies to extended objects and stars for most systems (except highly undersampled ones). The main point is that you cannot consider aperture, focal length, pixel size (or FNo) on their own - you must consider them all to get the system sensitivity.

re the way a lens works, please forgive me if I have misunderstood your question, but here goes. If you hold up your hand and look at it, you get almost no information on where the light that you see came from - it could have been from many different sources in the room, but you can't tell anything much about those sources by looking at the light reflected from your hand. A similar scrambled pattern of light falls all over the aperture of a lens or mirror, but the lens or mirror has the ability to unscramble that light pattern and develop a map of the angles that the light came from and how bright it is - that is the image that you detect. Light from a point source falls all over the lens aperture and it is then transformed (unscrambled) by the lens back into a point in the focal plane. Light from an extended object also falls all over the aperture, but it is transformed into a 2D representation of the source object. If the aperture is bigger, more light from both point and extended sources gets into the system.

[codemonkey]

That makes sense, thanks mate.



Nope, I got that, but that doesn't answer my original question about how aperture relates to more light instead of different light. Cheers for the link though, it was a good thread.



Spot on Ray. He's technically correct, but his presentation is a bit misleading. I actually got that part and my main question didn't really relate to his article all that much, it's just what I was thinking about at the time. I think I should have omitted that from my post as it's largely unrelated to my main question and has served only to confuse people.



The main thing I was confused about is why does a bigger aperture not mean light from different sources rather than more light from the same sources. To make that extreme, imagine I had a mirror the size of my thumbnail. Obviously it could only reflect light from a small area... but then if I sold my thumbnail-sized mirror and bought one the size of Queensland, it would reflect light from a much greater area simply due to its larger size, but (assuming all else is equal) the intensity of the reflected light would remain constant.

Now if the bigger aperture means that more light is focused into an area of the same size, I can see why the intensity, and resolution would increase... but how does this not also increase the "area" of light gathered, i.e. an increased FOV. Unless it does but because the recording medium was already fully illuminated (presumably), the extra FOV just bypasses it? In which case if you had a system where the sensor was not fully illuminated, keeping everything constant except the aperture would result in a sensor that is more fully illuminated with data that was previously not recorded...?

[gregbradley]

I think you got a good grasp of the subject and have cleared that area up nicely. I did not get it at first but then that F-ratio myth was blocking my understanding of it. I agree its missing that exact component as you have just stated.

You've also highlighted the importance of correct camera pixel size for a certain focal length.

Well done Ray. You've made a very good contribution here. Your article should be exported around other sites.

Greg.

[Bassnut]

Yes,excellent Ray. I always found Stans paper unintuative and confusing, something wasn't quite right. You have put it simply.

[barx1963]

The way I think about this is a light source such as a star emits light in all directions. If it simply emitted a single pulse of light there would be a sphere of photons heading out into the universe expanding all the time. If you were looking at the star at the time that sphere hit, you would collect s many photons as would fit into your pupil and that would determine how bright that the star appears at that moment. All a telescope does is collect photons from a larger part of the sphere and concentrate them into your pupil, hence making the star appear brighter. I think your reference to a star being a point source is confusing you, the light which is emitted is definitely not a point.

Malcolm

[astroboof]

If you had an aperture the size of Qld, but also had a FL that was equal to the average distance from our moon, what kind of light collection and inherent FOV quality would you think we might gain?

So, how long is a piece of string.

[RickS]

Lee,

I think the thing you are missing is that we are so far from these objects that the light rays from them are effectively parallel. If you're imaging M8 then increasing your aperture gives you an increase in the number of photons you collect from M8 which is proportional to the increased area.

Cheers,
Rick.

[Shiraz]

my guess is that the long fl image was taken at 3x3 binning Julian - that would explain the results.

[Peter.M]

For a given camera, focal length dictates field of view, nothing else. An 800mm FL f2 scope with an aperture of 400mm will have the same field of view as an 800mm FL f10 scope with an aperture of 80mm. In these systems, the only difference in the final image will be the light collecting area defined by its clear aperture. I think you would do well to google a ray trace of a telescope, what is happening is light from the star or nebula or whatever you image hits EVERY part of your mirror and it is focused onto the sensor. That is why bigger mirrors (for a set FL) will give brighter images.

Another way to show this physically, would be to put a longer focal length eyepiece into your scope. going from a 30mm to a 3mm (increasing magnification) makes the background darker.

[Steffen]

I'm guessing Stan uses his DSLR in full-auto :P

The refusal by most people to internalise (or at least accept) simple geometrical relations is why mobile phone cameras are eating DSLRs' lunch. Also, because they're smaller and always with you, but that reason doesn't further my argument :D

[Geoff45]

There is a nice article here http://www.stark-labs.com/help/blog/files/FratioAperture.php by Craig Stark. He discusses Stan Moore's conclusions and puts a somewhat different, but not necessarily contradictory, slant on things. His contention is that focal ratio does make a difference, but it is not as simple as "half the focal ratio makes your scope 4x faster"
Geoff

[Shiraz]

But in this series of slides, he supports the "1/2FNo gives 4x faster" result and points out that Stan's paper was not based on pixel SNR - lots of other interesting stuff as well.http://www.stark-labs.com/craig/resources/Articles-&-Reviews/ImageSampling_Fratios_SNR_RTMC.pdf

[gregbradley]

Good link Ray.

This is why I got an AP RHA 305 F3.8 ad has been my exact experience with it so far. Its also why I am keen to use the reducer on the CDK17 when I can.

A 5 minute Ha image using a 5nm Ha at 1x1 binning is probably as smooth and noise free as a 20minute F7 image.

It speeds up imaging by a factor of 3 or 4 at an educated guess over the more typical imaging at F7/8. Its probably about 2-3 times as fast as an FSQ106 even with a narrower FOV. Couple that with nearly 80% QE of Sony ICX sensors and you've got a very fast imaging system which reduces the problem of the variable of astrophotography weather (cloud, moon, wind).

Greg.

[Geoff45]

He doesn't actually say the focal ratio doesn't matter: He says

Thus there is an actual relationship between S/N and f-ratio, but it is not the simple characterization of the “f-ratio myth”.

Geoff

[codemonkey]

Thanks everyone!



Eh, he says a number of things, including:

"Focal length (and thus f-ratio) has absolutely no effect on the number of photons collected and delivered."

Either way, I derailed my own thread by mentioning that article. I didn't realise it would be so inflammatory.

[Geoff45]

Well the statement in quotes is obviously true. There is no way that f ratio can affect the number of photons hitting and hence being transmitted by the objective. What happens after the photons are transmitted by the objective is where the disagreement arises.

Nothing wrong with robust or inflammatory discussion. Just treat it as all good fun!

[gregbradley]

And how it reaches pixels and then how those pixels translate into a signal and how many photons miss the detector.

Greg.

[codemonkey]

Thanks for the clarification, really appreciate it.

[RickS]

You're a computer geek like me IIRC, Lee? It's not very complicated if you have some basic maths and science skills and give it a little thought. Take a look at the Ray/Shiraz model here: http://www.iceinspace.com.au/forum/showthread.php?t=136008

I came to effectively the same conclusion myself. Clear aperture determines how many photons you capture from a DSO and then image scale will tell you how they get stuffed into pixels. Optical efficiency plays a smaller role determining if some of the photons don't make it to the sensor and then QE tells you how many photons get recorded. Add a basic understanding of SNR and the sources of noise (shot noise, read noise, dark current noise) and you have all the tools you need.

Cheers,
Rick.

[codemonkey]

Thanks Rick :-)

I understand the role of aperture, focal length, central obstruction size, QE and pixel size and their combinatory effect on SNR/acquisition time; the part that I've been scratching my head about is the mechanics of telescopes and whether a change in aperture impacts the way light is captured from point sources vs diffuse sources. This was answered in the first response and it's simple and makes sense.

The other thing was whether (all else equal) a bigger mirror should mean you're capturing more light but from areas previously outside your FOV. Still a bit hazy on this one. Seems pretty obvious that it should, but all that seems to be discussed is how a bigger aperture = more light (implication being from the same source), and only the focal length affects the FOV.

[RobC]

Changing your objective/mirror size is not changing your field of view but changing the amount of focused light ( or photons ) hitting your sensor from the same field of view.

I think you are getting confused over the amount of stray/scattered light outside of your FOV hitting your sensor. That's is why baffling/flocking is used inside of your telescope tube.

You double your objective/mirror size for the same focal length and your quadruple the amount of light hitting your sensor.

[codemonkey]

I think that's it... I think I actually get it now. Thanks Rob!