Focal Ratio vs aperture size.

[roundycat]

I'm glad Martin mentioned that we have both recently got pictures of M63 because I took another 1h 50m of it last night in colour, of which more under Imaging - Deep Sky. I am always urging caution about mixing theory with practice because practice is, if you like, the thing that pays the rent and the final picture must be right regardless of what theory might have to say about the way you took it.

Like others on the forum I too have been playing with longer exposures of which more later. You may recall that my M63 was criticised (correctly) for having imperfect calibration. I went back to the original raw files, calibrated them with the correct flats and re processed the whole thing. I did faff around with some layer masking of a more contrasty but darker centre to the galaxy but the overall result was more noisy than the original.

As I remarked in a previous post somewhere a slightly noisy result can be 'fixed' in Photoshop but how do you explain a second attempt at processing the same data that turns out noisier than the first one? I should add this is a subjective impression, not a serious measurement.

I think this aligns to some extent with the point Martin made about f7 and f10 scopes. You can arrange to take two pictures with the same SNR but if you enlarge the smaller picture to match the bigger one you are going to lose something that overrides all other considerations.

Dennis

[ngc2403]

reading the PDF again there is a term in there pi D squared. Is this a new way of calculating area or have I missed something?

Dennis

The area of a circle ca be found by using the diameter as long as you divide the result by 4.

see in the PDF the is a 4/pi D^2.

i have not seen the image guy but i am going to have wee look at them now.

Needless to say that different people see different thing when they look at images so we everyone is likely to have a different opinion.

ally

[ngc2403]

Having had a look at both of martins images i would say that your first image is comparable to Dennis's however the Jpg has lost all the contrast so i can't really compare the two...

Your reprocessed image looks very different and although i don't think i can say there is a marked improvement in the image i will say that his (Dennis) looks better, the focus looks sharper and the galaxy's dust lanes really stand out.

we can argue this out but i don't think i will go any where unlike the theory with has only one correct answer. The correct answer is one that take all aspects into account and then marks a decision.

[narrowbandpaul]

yes, you have missed something.

A=pi*r^2

but since r=D/2,

A= pi*D^2/4

in any case its only a dimensionless number not far from one. It would not change the dependance on F ratio.

My derivation therefore still holds.

[MartinB]

I didn't post my luminence image but would be happy to put it alongside Dennis's here if people are interested. As I said, it is F7 and 2 hours in total vs F10 3 hours. When viewed at the same size mine is noisier but of similar resolution. The sub exposure times and LP may vary but the optical equipment is very similar.

[dph1nm]

There is no reall mystery to any of this. The bottom line for *imaging* is that what really matters is

1) Telescope aperture (squared)

2) Exposure time

3) Pixel size in arcsecs of the CCD

Term 3) can be altered by changing the F-ratio, but it can equally well be altered by physically changing the CCD or (to a limited extent) by binning up pixels on an existing CCD.

NigelM

[ngc2403]

Dear Martin, you can post your luminance layer if you want i don't know if it will show things clear or not. The thing that i would be interested in is the subs length then we can see how the different focal length have improved the image.

As a wee note this is not good science to compare images processed differently and in truth the SNR talked about in the above posts is just the SNR of the raw image, which is changed as soon as you apply any processing to it.

Nigel, the pixel size in arc seconds is dependent on the focal length so we could say that it is

1) aperture (squared)

2) exposure time

3) pixel size actual (squared)

4) focal length (inverse square)

i.e. 1 and 4 combine to give focal ratio squared....

so just that and time and pixel binning in other words

[narrowbandpaul]

yes

but pixel size in arcsec is governed only by focal length and pixel size.....

and if we divide aperture squared by focal length squared we get focal ratio.

the flux is detemined by aperture, but how that is then spread over the sensor is governed solely by focal ratio. ie W/m2/arcsec^2 only depends on f ratio

not aperture.

not ever.

the derivation is correct, why are we arguing, or are we agreeing?

paul

[acey]

If "pixel size in arcsec is governed only by focal length and pixel size" can we conclude that focal length = arcsec?

Are pixels measured in metres or in arcseconds?

[narrowbandpaul]

pixels size has units of length. They can be expressed in megaparsecs if you like. The usual unit is microns.

in radians...arcsec/pix =f/p, both in metres

[narrowbandpaul]

sorry p/f

[dph1nm]

yes

but pixel size in arcsec is governed only by focal length and pixel size.....

and if we divide aperture squared by focal length squared we get focal ratio.

the flux is detemined by aperture, but how that is then spread over the sensor is governed solely by focal ratio. ie W/m2/arcsec^2 only depends on f ratio

not aperture.

not ever.

the derivation is correct, why are we arguing, or are we agreeing?

paul

Both! Everyone is correct. The flux per unit linear area on the detector is indeed determined by focal ratio. But, the f-ratio myth argument is that what matters in astronomy is the flux per unit angular scale on the sky. This determines what your image will look like. If you think that way then focal-ratio is not important.

NigelM

[narrowbandpaul]

linear area? arent areas 2D?

and if you think flux per square arcsec then F ratio is the only thing that is important.

look at the derivation