ZWO ASI2600MM-Pro vs Cooled and Modded DSLR image quality

[Merlin66]

Whooops, sorry missed that!

[MarsG76]

6 hours ago, sharkmelley said:

The main difference is the noise reduction from the cooling.  Otherwise a modded DSLR/mirrorless can produce very similar results because in many cases dedicated one-shot-colour astrocams use the same sensor as a modded DSLR/mirrorless camera.

Mark

I'm starting to think that I'll have minimal benefit from the camera as my DSLR is active cooled.... only hope I have now is detail captured due to no bayer matirx on the mono camera. Also hoping for faster photon aquisition.

 

 

[sharkmelley]

There will be a big advantage because the modern sensor has at least twice the QE of your 40D sensor.  Also you have the advantages that mono gives you over one-shot-colour i.e. you can shoot luminance and narrowband is not impeded by the bayer matrix.

Mark  

[rsarwar]

On 01/04/2021 at 00:44, MarsG76 said:

I'm starting to think that I'll have minimal benefit from the camera as my DSLR is active cooled.... only hope I have now is detail captured due to no bayer matirx on the mono camera. Also hoping for faster photon aquisition.

 

 

Dont forget you will get less noise - meaning to get to the same quality as a DSLR, you would need to stack far less frames. i.e. your exposure time for a single frame will not shorten by more than 1-1.5 times, but the total integration time will be lot shorter, albeit you need 3-4 channels to work with.

 

I am still awaiting for my QHY268. got some clear skies today with quater moon only rising at 5 am. 

[MarsG76]

19 hours ago, rsarwar said:

Dont forget you will get less noise - meaning to get to the same quality as a DSLR, you would need to stack far less frames. i.e. your exposure time for a single frame will not shorten by more than 1-1.5 times, but the total integration time will be lot shorter, albeit you need 3-4 channels to work with.

 

I am still awaiting for my QHY268. got some clear skies today with quater moon only rising at 5 am. 

Do you know this for a fact or are you speculating/estimating like me?

 

 

[rsarwar]

17 hours ago, MarsG76 said:

Do you know this for a fact or are you speculating/estimating like me?

 

 

Speculation as i still dont have the camera, but i am confident. 

Qhy268 has smaller pixel and lower read noise and thus will have better SNR. So to achieve the same level of SNR seen in a DSLR, you will need stack less number of frames to achieve the same level of detail /SNR as a DSLR. 

 

also your new camera is 16-bit, that will also play a major part in bringing out more detail. 

 

What i am saying is there is no comparision to be had between QHY268/ASI2600 and a DSLR - it will whip the floor with any DSLR from any generation. However, your comparison based on exposure time needed for each frame, is not a the right one to be looking at.

Edited April 5, 2021 by rsarwar

[rsarwar]

So i got my QHY268 earlier today. 

but cloudy outside. 

 

Was wondering if anyone knows how to map ISO values from DSLRs are to be mapped onto the Gain scale. Without the right mapping, this comparision will not be possible.

[SteveNickolls]

I'd be happy seeing a best effort with a DSLR v the astro-camera and how they simply look as a test result. 🤗

Good luck and clear skies.

Cheers,
Steve

[sharkmelley]

14 hours ago, rsarwar said:

Was wondering if anyone knows how to map ISO values from DSLRs are to be mapped onto the Gain scale. Without the right mapping, this comparision will not be possible.

This site gives you the ISO for unity gain (i.e. 1e-/ADU): DxOMark Derived Sensor Characteristics (photonstophotos.net)

The QHY268 gain setting for unity gain (1e-/ADU) will be available probably in the documentation.  You can then adjust upwards and downwards from there.  You will want to choose a gain where both DSLR and QHY268 are operating in the same mode i.e. high gain or low gain.

Mark

[rsarwar]

3 hours ago, sharkmelley said:

This site gives you the ISO for unity gain (i.e. 1e-/ADU): DxOMark Derived Sensor Characteristics (photonstophotos.net)

The QHY268 gain setting for unity gain (1e-/ADU) will be available probably in the documentation.  You can then adjust upwards and downwards from there.  You will want to choose a gain where both DSLR and QHY268 are operating in the same mode i.e. high gain or low gain.

Mark

Thanks Mark. I will have a read through it after work. 

I am using my ASI EWF, so have to wait for some M48 extentions for my flattenern but i would definately try to work this out

[rsarwar]

Okay I did some testings on the heart nebula. very rough ones. but using the same target and same telescope (130 pds). I live is bortle 6.

DSLR: D7000, 16 MPix APS-C IMX071 (same was ASI071) with full spec modd + custom firmware to disable onboard post processing, i.e. true darks + no star eating stuff. The pictures (always the right hand side of each screenshot) were taken a weeks ago (zero moon with the nebula fairly high in the sky) when i got the M48 Astronomik 12 nm Ha filter and wanted to see how well it worked and how an OSC handles it. ISO was set at 800 and time was 600 sec. according to the link sharkmelley gave, and according to this link, it equates to a gain of approx 3.3 * unity. ambient temp was about zero. only showing the red channel here - as it is  a Ha filter

QHY268M with the same filter, no moon, and readmode=0, offset=25 and gain=56 (6.6 time unity) but took a 300 sec frame (i was having issues with backfocus and wanted to investigate it so wanted a shorter exposure - it did saturated some of the stars). so I did not frame the target - just point and shoot. camera cooled to -15 C. image was taken when the nebula was fairly high in the sky.

 

Firstly, the autostreched images on PI. the left image, taken by qhy is clearly shows more detail and is brighter.

Secondly, same images but zoomed in to the core; ignoring the obvious backfocus/focus issues the Max value of a pixel on the nebulocity was 0.178 and 0.131 for the qhy and d7000 respectively. that makes the qhy 1.36 times faster. [D7000 is 4.7um with an eff of 35-40pc, so i was would have expected 1.20 times better.] However the level of detail is phenomenal in my opinion. qhy has less noise. so less stacking is needed. I also have a 1200 sec frame from the D700, and it is still not able to match the level of details

final picture is that of the same images, but using the same stretch.

@MarsG76 I dont think you will be disappointed

 

Edited April 13, 2021 by rsarwar

[SteveNickolls]

Thank you so much for doing this comparison, it's now a question of saving up he he.

Best regards,

Steve

[MarsG76]

On 13/04/2021 at 20:36, rsarwar said:

Okay I did some testings on the heart nebula. very rough ones. but using the same target and same telescope (130 pds). I live is bortle 6.

DSLR: D7000, 16 MPix APS-C IMX071 (same was ASI071) with full spec modd + custom firmware to disable onboard post processing, i.e. true darks + no star eating stuff. The pictures (always the right hand side of each screenshot) were taken a weeks ago (zero moon with the nebula fairly high in the sky) when i got the M48 Astronomik 12 nm Ha filter and wanted to see how well it worked and how an OSC handles it. ISO was set at 800 and time was 600 sec. according to the link sharkmelley gave, and according to this link, it equates to a gain of approx 3.3 * unity. ambient temp was about zero. only showing the red channel here - as it is  a Ha filter

QHY268M with the same filter, no moon, and readmode=0, offset=25 and gain=56 (6.6 time unity) but took a 300 sec frame (i was having issues with backfocus and wanted to investigate it so wanted a shorter exposure - it did saturated some of the stars). so I did not frame the target - just point and shoot. camera cooled to -15 C. image was taken when the nebula was fairly high in the sky.

 

Firstly, the autostreched images on PI. the left image, taken by qhy is clearly shows more detail and is brighter.

Secondly, same images but zoomed in to the core; ignoring the obvious backfocus/focus issues the Max value of a pixel on the nebulocity was 0.178 and 0.131 for the qhy and d7000 respectively. that makes the qhy 1.36 times faster. [D7000 is 4.7um with an eff of 35-40pc, so i was would have expected 1.20 times better.] However the level of detail is phenomenal in my opinion. qhy has less noise. so less stacking is needed. I also have a 1200 sec frame from the D700, and it is still not able to match the level of details

final picture is that of the same images, but using the same stretch.

@MarsG76 I dont think you will be disappointed

 

Good stuff...

I'm still waiting for my camera.... taking a while...

 

Looking at your test shots, you can definitely see improvement, and what makes me happy is that you're comparing 300s QHY subs to 600s DSLR subs.. and the QHY subs have more signal... so that is at least more than half the exposure time... possibly 4 time quicker, as you said that a 1200s DSLR exposure was not as bright at the 300s test...

I can't wait to do a test of my own.

[rsarwar]

14 hours ago, MarsG76 said:

what makes me happy is that you're comparing 300s QHY subs to 600s DSLR subs.. and the QHY subs have more signal... so that is at least more than half the exposure time... possibly 4 time quicker, as you said that a 1200s DSLR exposure was not as bright at the 300s test...

I dont want to deflate you however, the 1200 sec from D7000 was "brighter" than QHY. What i suggested was that It did not have comparable detail (i.e. qhy had more detail) - possible due to superior bit-depth and dynamic range.

I am comparing 300s QHY with 600s DSLR, however, the operative consideration there is that qhy was operating in 6.6 * unity gain whilst DSLR was on 3.3 * unity gain. It's like comparing two DSLR at ISO 400 vs ISO 800. the one with the higher ISO will have more signal and hence half the exposure time, however will sacrifice dynamic range and welldepth. with 6.6 * unity gain, qhy manage to saturate 30% of the stars, so the gain i ran the test on is not a usuable gain in most case.

i calculated a reduction of exposure time by a factor of 1.36. yours will depend on the QE and pixel size of your dslr.

 

I have to state - QE does not indicate performance of a camera. it is just another arbitary param (look at the QE curve from ASI2600M and QHY268M, their curves vary vastly for the same chip). asi294 and asi2600 has almost the similar QE (infact asi294M has better QE compared to asi2600M), however asi2600 will outperman a asi294 given it is 16-bit. For example, nikon and canon camera have comparable QE. however I would never pick a canon camera as my primary interest is night time photography - because of the poor dynamic range of canon cameras. you just can pull out details. things that matter are read-noise, dark current, dynamic range, bit depth and well dept

Edited April 15, 2021 by rsarwar

[rsarwar]

Running a proper mock session on the heat nebula with readmode=1, gain = 56 and offset = 25. this would be my standard setup. this is approx 2. compared to D700 at iso800 (3.3 * unity gain) the brightness is exactly the same (runing 600 sec subs). however, the level of detail is still as impressive as before thanks to DR and bit depth in my opinion.

[MarsG76]

19 hours ago, rsarwar said:

I dont want to deflate you however... poor dynamic range of canon cameras. 

You wont deflate anything... don't worry... what Canons were you using? The Canons I used delivered better images than any Nikon I've seen, especially in AP... there's a reason why Canon DSLRs are the DSLR of choice in the vast majority of astroimagers. 

[MarsG76]

15 hours ago, rsarwar said:

Running a proper mock session on the heat nebula with readmode=1, gain = 56 and offset = 25. this would be my standard setup. this is approx 2. compared to D700 at iso800 (3.3 * unity gain) the brightness is exactly the same (runing 600 sec subs). however, the level of detail is still as impressive as before thanks to DR and bit depth in my opinion.

I'd like to see some of your astroimages... post some here.. or do you post them to another site??

[rsarwar]

I only started astro last year. I used Canon before in early 2010s - 550d and 70d. The nikons cameras post 2013 all use sony sensors and have better SNR and DR. Prior to 2013, i nikon's offerings were below below par. Notice the last three entries:

I take a lot of images of fireworks

This image was taken with D7300 and it was completely blown out. i had similar blown out shots from canon and they were not recoverable. that is why i sold all my canon gear to move to nikon. initially to d7100 and now D750

 

I took these last night. Stacked. total integration 2 hour for both

QHY 300 sec (gain = 2.2*unity)😞

D7000 600sec (gain = 3.3*unity):

same stretch, qhy on the left.

 

Edited April 16, 2021 by rsarwar

[rsarwar]

having said that - i dont think D7000 is something that should be put aside. it is a very capable camera.

this was with IDAS V4 filter.

 

Unfortunately i have to wait till next year to try horse head/orion/rossetta nebula as they are blocked by some trees

Edited April 16, 2021 by rsarwar

[rsarwar]

45 minutes ago, MarsG76 said:

You wont deflate anything... don't worry... what Canons were you using? The Canons I used delivered better images than any Nikon I've seen, especially in AP... there's a reason why Canon DSLRs are the DSLR of choice in the vast majority of astroimagers. 

Here is the heart nebula from last night. Before i kept my guide camera, sv305, at 2x2 which gives a pixel size of 5.8. this worked well with D7000@4.7um. but was inadequte and i only realized after running the session for 1.5 hours  - so ended up losing half the frame. the binning on the M81 was 1x1 and i was able to keep all 41 frames before the sun came up.

10 min subs at 2.2 * unity gain. Halpha filter

peak on nebolosity region is 0.0123 and the peak on image from D7000 i posted previously was 0.0136. that is factor of 1.11 in favour of D7000, but qhy was operating at a gain 1.5 times lower. so it all boils down to qhy being 1.35 times faster.

Edited April 16, 2021 by rsarwar

[MarsG76]

8 hours ago, rsarwar said:

i dont think D7000 is something that should be put aside.

My DSLR will still be in operation once my cam arrives.. but I imagine it'll be near exclusively used for natural color data to colorize the luma mono exposure stacks.

That said, who knows how long it'll be before I get the cam.. been over 2 months now and I'm not getting replies from the shop.... it's a big shop so I know that they are doing a runner.. but the service from china (ZWO) is abominable... pffft china.

 

[MarsG76]

15 hours ago, rsarwar said:

I only started astro last year. I used Canon before in early 2010s - 550d and 70d. The nikons cameras post 2013 all use sony sensors and have better SNR and DR. Prior to 2013, i nikon's offerings were below below par. Notice the last three entries:

 

I took these last night. Stacked. total integration 2 hour for both

QHY 300 sec (gain = 2.2*unity)😞

D7000 600sec (gain = 3.3*unity):

same stretch, qhy on the left.

 

That comparison data sheet is exactly what I'm trying to get away from... this is obviously biased and canon biased comparisons are there too.. one can find what ever data or statistics they want to see, hence why I started this thread asking about real world comparisons.... and what you're showing is that there is not a phenomenal difference between the qhy and the nikon... than again you're posting severely scaled down low res examples, how about full res or full res center cropped comparisons... there has to be more improvement!!!

 

[rsarwar]

14 hours ago, MarsG76 said:

That comparison data sheet is exactly what I'm trying to get away from... this is obviously biased and canon biased comparisons are there too.. one can find what ever data or statistics they want to see, hence why I started this thread asking about real world comparisons.... and what you're showing is that there is not a phenomenal difference between the qhy and the nikon... than again you're posting severely scaled down low res examples, how about full res or full res center cropped comparisons... there has to be more improvement!!!

 

how can data be biased? Interpretation, sure that can be baised, but not data. You may choose to dismiss data if they donot fit your world domination plans, but one can not just find whatever data or or statistics they want to see, otherwise we would surely be able to find a dataset which showed a APS-C sized camera is able to take a picture that equats to a full frame camera in terms of field-of-view when using the same lens (focal length). if your comment relates to the canon vs nikon debate, i donot want to go into a fanboy based argument, but i know what the data states and i know what my experience has been thus far based on which i made choices that had financial consequences for myself

ill uploaded the fits  file to this post (not sure if you use PI or someother tool). Also just to clarify, i did not show there is no phenomenal difference between the qhy and the nikon - but on the contrary. The contrast between the two cameras is very straigtforard - qhy is wins hands down. and i have already said why and not really willing to rehash them again.

What i showed is that your expectation for exposure time is unfounded. Exposure time, i must add, has not bearining on the quality of of the pictures produced - just the amount of time it is needed to complete a project. infact the only reason i ran the test is because i was secretly hoping that going mono did not mean i have to spend 3/4 times more time taking picture - coming from OSC i am not thrilled (i am guessing that is why you are also interested in this question). and i have the answer i was looking for - the SNR, DR,  bitdepth and the welldepth allows superior details, meaning if i was to to be happy producing a picture equivalent to DSLR quality, yes i can reduce the total integration time, but the exposure time will have to be similar, or 75% of what i used on my dslr. 

 

QHY_M_82_Light_H_Alpha_300_secs_readmode_1_gain_56_2021-04-16T00-00-58_002.fits QHY_IC_1805_Light_H_Alpha_600_secs_readmode_1_gain_56_2021-04-15T22-09-31_002.fits QHY_IC_1805_Light_Ha_300_secs_readmode_0_gain_59_2021-04-12T22-40-38_001.fits D7000_M_82_Light_600_secs_ISO800_2021_03_06T00_37_33_001_c_cc_d_R.fits D7000_IC_1805_Light_600_secs_ISO800_2021_04_03T21_18_45_001_c_cc_d_R.fits

Edited April 17, 2021 by rsarwar

[MarsG76]

On 16/04/2021 at 22:57, rsarwar said:

QHY 300 sec (gain = 2.2*unity)😞

D7000 600sec (gain = 3.3*unity):

same stretch, qhy on the left.

 

...this does not say to me that there is a phenomenal (or much) improvement at all... when I get a day off and some free time, I'll download your fits and have a look, but my expectations are quite dashed... oh well... at least the physical size of the ZWO2600 will allow me to move into the realm of Hyperstar imaging.

 

 

[rsarwar]

5 hours ago, MarsG76 said:

 

...this does not say to me that there is a phenomenal (or much) improvement at all... when I get a day off and some free time, I'll download your fits and have a look, but my expectations are quite dashed... oh well... at least the physical size of the ZWO2600 will allow me to move into the realm of Hyperstar imaging.

 

 

that is your interpretation. however even in the low res screenshots, i see details in the qhy image that are not present in dslr.  and the obvious lack of noise. All this while the obvious understampling of the subject due lack of focal length - imagine if i had the right focal length on the scope

 

if you are concerned, change your order to asi2600mc. it will work better with hyperstar anyway.

However, as i said, you are looking at the wrong parameter in order to define performance of a camera. you should be looking at SNR - not "brightness"

Edited April 18, 2021 by rsarwar