Canon 5D MkII vs. Touptek IMX294 - back to back

[Rallemikken]

Full moon, not much you can do. Did a small comparison; 5 frames with my new IMX294, and 5 more with my trusted Canon 5D. Not super-impressed by the new kid in town.

Top: Canon 5D MkII, 120 seconds, ISO 800. Stacked with 5 calibration frames of each, color calibrated and a pretty hard stretch.

Bottom: Touptek IMX294, cooled to -10C, 120 seconds, gain 398, offset 100. Stacked with 5 calibration frames of each, color calibrated and a pretty hard stretch.

Scope: 8" RC, target M82. Blindingly bright and full moon, but no moonlight into the tube. No skies. 10 degrees celsius, that means the Canon had a big disandvantage, as I turned on the cooling on the Touptek.

Not sure what to expect. Stacked and photometric color calibrated in Siril. Both images are stretched with levels in Gimp until the histogram looked similar. Pixelscale 1:1, only crop, no resampling. The IMX294 has smaller pixels and brings the galaxy closer. That means that the targets brings more light into the frame, and the background gets darker as long as I insist on identical histogram peaks. M82 will be my next target with this scope when conditions permit, not sure what camera to use after this. 

[GTom]

You also need to scale them to the same size, your Touptek operates at approx 2x higher magnification. Check some very faint stars on the bottom one: they might be invisible on the canon's result.

Also, the canon background seems certainly higher, means lesser S/N that will have a visible impact on your stack result. Small differences that eventually add up and adding up we do...

[Clarkey]

You will get a better SNR with the astro camera. Also, if/when you start using any sort of NB filter, the astro camera will give you much more signal.

[Rallemikken]

12 minutes ago, GTom said:

You also need to scale them to the same size, your Touptek operates at approx 2x higher magnification.

Hesitated to do that, would not resample. Now I see that SGL HAS RESAMPLED AND SCALED, so the whole post has been in vain. But I see your point. 

Some makes an big issue of pixelscale. The IMX294 has pretty big pixels, 4,63um, but the 5D  has even bigger pixels at 6,41um. A scaling would show if there are some truth in this. I got the 8" RC (1600mm focal length) and the IMX294 for use on galaxies and small, but not to dim targets. Season has just started for me, looking forward to test it some more.

[Rallemikken]

10 minutes ago, Clarkey said:

You will get a better SNR with the astro camera.

Hope so. Must learn to tune it. And the smaller sensor size (for me that are used to old Canon's) means that I don't have to bother to much about what happens near the edges. The fullframe 5D can be demanding.  As for filters, don't use them. Bortle 3/4. I'm also hesitant to use any glass between the mirrors and the sensor. I also have two newt's, one f/4 and one f/5. The f/4 must have a flattener, with the f/5 it depends on the target. How much I can crop out.

[GTom]

12 minutes ago, Rallemikken said:

Hesitated to do that, would not resample. Now I see that SGL HAS RESAMPLED AND SCALED, so the whole post has been in vain. But I see your point. 

Some makes an big issue of pixelscale. The IMX294 has pretty big pixels, 4,63um, but the 5D  has even bigger pixels at 6,41um. A scaling would show if there are some truth in this. I got the 8" RC (1600mm focal length) and the IMX294 for use on galaxies and small, but not to dim targets. Season has just started for me, looking forward to test it some more.

Eventually the point is to see if there are more details and if you can spot fainter targets on one of the pictures.

Also, if you use a MONO camera, your color S/N ratio will be in a different universe, especially in HA and SII where only 1/4 of the Bayer pixels see anything even if astro-converted.

Edited September 19 by GTom

[Clarkey]

1 hour ago, Rallemikken said:

As for filters, don't use them. Bortle 3/4. I'm also hesitant to use any glass between the mirrors and the sensor.

That makes sense for RGB imaging such as star clusters, reflection nebulae and galaxies (although some benefit from Ha imaging), but what about all the emission targets? If you have a full moon, NB allows you to image more targets.

[Rallemikken]

51 minutes ago, Clarkey said:

what about all the emission targets? If you have a full moon, NB allows you to image more targets.

Boiles down to philosophy, I want real colors, and I won't let anything tamper with it on it's way to the sensor. And the days around full moon means rest and restitution. Long nights this far north, I know from experience there will be plenty of integration time this winter too!

[GTom]

44 minutes ago, Rallemikken said:

Boiles down to philosophy, I want real colors, and I won't let anything tamper with it on it's way to the sensor. And the days around full moon means rest and restitution. Long nights this far north, I know from experience there will be plenty of integration time this winter too!

You'll always have a UVIR blocking filter of some sort, even when using Newtonians.

BTW many NB targets are simply dark outside their respective emission bands, using filters on them doesn't change their color. True, it's not always the case and also true star colors, if they are significant, are messed up.

Edited September 19 by GTom