asi224 vs asi678mc for planetary / edge 11?

[theskyisthelimit99]

I believe its been said the new 678mc is good with the edge 11 (maybe without a barlow too).. can anyone explain in general terms how this is?  

Im guessing less noise with this iteration, but the pixel size is 2.0 vs 3.75 which goes against the grain i thought for imaging planets.

I wouldnt mind losing the barlow however and i guess doing drizling 
Any thoughts

[vlaiv]

I can tell you in very simple terms what theory says (and no - I don' buy into "theory is one thing while in practice ..." notion that many entertain - if theory is properly applied - it works, otherwise it is flawed and we would know about it - it cease to be current theory as new explanation would be offered).

Thing is - many respected planetary imagers don't follow it (each for their own reasons) - and that creates a lot of confusion, so I won't go into explaining any of that - just state facts.

Optimum F/ratio for planetary imaging is related to pixel size by following formula

f_ratio = 2 * pixel_size / wavelength

where wavelength is wavelength of light of interest. It can be exact wavelength in case of use of narrow band filters - otherwise we must observe that visible part of spectrum is 400-700nm and one should use 400nm as limit of resolution (as more detail is resolved at shorter wavelengths).

However, I often advise use of 500nm instead for regular RGB imaging (with OSC or mono + filters). This is for several reasons, one of which is that luminance carries the most detail information in the image and peak of our luminance perception is north of 500nm. Another is that shorter the wavelength - more distortion there is from seeing (refraction is strongest for short wavelengths - that is why sky is blue - strongest scatter).

This leads to two very simple formulae - you can pick which one prefer:

F_ratio = 2 * pixel_size / 0.5um (500nm) = 4 * pixel_size

F_ratio = 2 * pixel_size / 0.4um (400nm) = 5 * pixel_size

With C11 and 678 - you don't really have a choice, as C11 is F/10 scope - and ASI687 has 2um pixels - so you can pretend that you used 5 * pixel_size formula

Btw, don't do drizzle (except bayer drizzle, which is done by default with AS!3 when processing OSC data).

Back to original question ASI224 vs ASI687

There are only 3 numbers that you should compare:

1. QE

2. Read noise

3. Frame rate

Most of new USB3.0 cameras are capable of frame rates that depend on your exposure length (which should be around 5ms - so around 200fps), so third point is not as important as it used to be (USB 2.0 vs USB 3.0 for example).

In the end - there is little between these two models. Both have QE around 80% (maybe ASI678 has slight edge - but do note that QE is concentrated in blue part of spectrum - and blue carries the least information as far as luminance is concerned) - so I'd give them a tie there. Read noise is also a tiny bit better on ASI678 - 0.6e vs 0.8e, but I've found that quoted read noise is often too optimistic anyway, so there is question if those numbers are accurate. Regardless of that, if ZWO uses same methodology - then relative difference should still hold. ASI678 has very small edge there.

Only advantage (in my view) is that ASI224 can use F/15 (which is 4 * 3.75) while you don't have that choice with ASI678 and need to use F/10 - which is 5 * 2um - so you are stuck with that formula instead of having a choice.

 

[theskyisthelimit99]

5 hours ago, vlaiv said:

I can tell you in very simple terms what theory says (and no - I don' buy into "theory is one thing while in practice ..." notion that many entertain - if theory is properly applied - it works, otherwise it is flawed and we would know about it - it cease to be current theory as new explanation would be offered).

Thing is - many respected planetary imagers don't follow it (each for their own reasons) - and that creates a lot of confusion, so I won't go into explaining any of that - just state facts.

Optimum F/ratio for planetary imaging is related to pixel size by following formula

f_ratio = 2 * pixel_size / wavelength

where wavelength is wavelength of light of interest. It can be exact wavelength in case of use of narrow band filters - otherwise we must observe that visible part of spectrum is 400-700nm and one should use 400nm as limit of resolution (as more detail is resolved at shorter wavelengths).

However, I often advise use of 500nm instead for regular RGB imaging (with OSC or mono + filters). This is for several reasons, one of which is that luminance carries the most detail information in the image and peak of our luminance perception is north of 500nm. Another is that shorter the wavelength - more distortion there is from seeing (refraction is strongest for short wavelengths - that is why sky is blue - strongest scatter).

This leads to two very simple formulae - you can pick which one prefer:

F_ratio = 2 * pixel_size / 0.5um (500nm) = 4 * pixel_size

F_ratio = 2 * pixel_size / 0.4um (400nm) = 5 * pixel_size

With C11 and 678 - you don't really have a choice, as C11 is F/10 scope - and ASI687 has 2um pixels - so you can pretend that you used 5 * pixel_size formula

Btw, don't do drizzle (except bayer drizzle, which is done by default with AS!3 when processing OSC data).

Back to original question ASI224 vs ASI687

There are only 3 numbers that you should compare:

1. QE

2. Read noise

3. Frame rate

Most of new USB3.0 cameras are capable of frame rates that depend on your exposure length (which should be around 5ms - so around 200fps), so third point is not as important as it used to be (USB 2.0 vs USB 3.0 for example).

In the end - there is little between these two models. Both have QE around 80% (maybe ASI678 has slight edge - but do note that QE is concentrated in blue part of spectrum - and blue carries the least information as far as luminance is concerned) - so I'd give them a tie there. Read noise is also a tiny bit better on ASI678 - 0.6e vs 0.8e, but I've found that quoted read noise is often too optimistic anyway, so there is question if those numbers are accurate. Regardless of that, if ZWO uses same methodology - then relative difference should still hold. ASI678 has very small edge there.

Only advantage (in my view) is that ASI224 can use F/15 (which is 4 * 3.75) while you don't have that choice with ASI678 and need to use F/10 - which is 5 * 2um - so you are stuck with that formula instead of having a choice.

 

Thanks that makes total sense and a fantastic summary.. i mean i guess if the 678 can avoid the barlow, then perhaps it takes some glass out of the equation.  I gues

You mentioned to not drizzle.. i was talking about the advanced settings of as3, has off, 1.5, 3 etc, i thought i read someone with the 678 was drizzling 1.5x, though i didnt realize that as3 was doing it by default during stacking?

[geoflewis]

17 hours ago, vlaiv said:

This leads to two very simple formulae - you can pick which one prefer:

F_ratio = 2 * pixel_size / 0.5um (500nm) = 4 * pixel_size

F_ratio = 2 * pixel_size / 0.4um (400nm) = 5 * pixel_size

Excellent explanation vlaiv and the first time I've seen it set down why the 4* to 5* pixel_size is recommended optimum for planetary imaging.

I'm now using ASI462MC (which has 2.9um pixels) with my C14 and ADC. Until using this camera I typically included x2 TV PM in my imaging train, targeting say F20-F22, but the initial results with the ASI462MC camera were terrible, so I took the x2 PM out. With the ADC in train the native F11 of my scope is pushed nearer F12/F13, so bang in the middle of the target of those 2 recommendations (2.9*4=F11.6; 2.9*5=F14.5). Now I won't waste my time experimenting with the x2 TV PM again.