A side by side (but not very scientific) comparison of QHY and Risingcam IMX571 OSC cameras

[tomato]

Having recently acquired the Risingcam (Touptek) version of the IMX571c sensor I said I would run a side by side comparison with the QHY 268c using the same scopes on my dual rig. I managed around 90 mins last night in reasonable conditions and although I am sure the holes in this brief exercise will soon be exposed I'll post what I have and let folks draw their own conclusions. Yes, I could run them both through the Sharpcap sensor analysis but what I'm really interested in is what do the images look like and are they comparable given the price differential?

The cameras were attached to Esprit 150s, both equipped with IR/UV cut filters, both running at -20C. The Risingcam was run at Gain 100, Offset 10 (apparent optimum DSO settings from reading various forum posts) and the QHY was running in the camera's DSO photographic mode, Gain 30, offset 30. All exposures were 180 seconds. Both scopes were imaging simultaneously, so both would have encountered very similar changes in sky conditions.  I will put my hands up now and say there is one rather large bluebottle in the ointment on this comparison and say the Risingcam had an 2" Baader (not CMOS optimised) IR/UV cut filter, while the QHY had an  Astro Essentials 2" IR/UV cut filter.

The first thing to note was the HFR stats from NINA (after running the auto focus routine at the same times on both scopes) were always better on the Risingcam subs than those on the QHY scope,  ~3.6 cf ~4.0. In the absence of anything else I would put this down to the different filters. The Risingcam also consistently had a bigger star count on each sub. I briefly ran some subs on the QHY at Gain 100, offset 10, the star count went up but so did the HFR. After calibration with their respective darks, flats and flat darks,  I have posted below  the best frame from each camera selected on quality by APP, autostretched in APP and converted to JPEG. I have also posted the stacks from each frame, each 32 x 3 mins auto stretched in APP. One obvious difference is the Risingcam stack has a big vertical background gradient which is absent in the QHY stack.  I presume  this is attributable to the different  gain settings on the cameras, given they were imaging the same bit of sky at the same time. I was trying to operate both cameras at their respective optimum DSO imaging settings.

I suppose I should swap the filters over and run the comparison again, or better still, put my hand in my pocket and buy 2 shiny new Baader CMOS optimised filters.🙄

All I'll  say at this point is I really like what both of these OSC cameras can deliver in a relatively short integration time, I will post the result of the combined data in the Deep Sky Section.

Best Risingcam sub

Best QHY sub

Risingcam stack

QHY stack

 

[nfotis]

Quite interesting comparison, thanks for the demonstration. I suppose that there's a need for some settings optimization still?

N.F.

[tomato]

I agree, but I’ll be interested to see what folks think I should do first, presumably sort out the IR/UV cut filters?

[nfotis]

The RisingCam stack has definitely a lighter background at the top half of the image, I suppose that's due to a different offset setting for the black point?

I don't know if the IR/UV cut filter can influence the end result so much?

N.F.

 

Edited December 11, 2021 by nfotis

[Laurin Dave]

Interesting comparison Steve..  are you using an OAG to guide? If so and that’s the side with better star size maybe there’s a bit of flex causing the difference..  also did you use flats to calibrate,  if so it’d be worth looking at the uncalibrated subs to see if the gradient is the same..  

Dave

[tomato]

Yes, the OAG was on the Risingcam scope so that fits with your suggestion, since scope flex is always a possibility on this rig.

I’ll take a look at the uncalibrated subs.

[newbie alert]

Qhy looks far superior on the background and to my eye a little sharper on the target 

[Starflyer]

There's definitely a gradient on the Risingcam stack, were both stacks fully callibrated?

[tomato]

1 hour ago, Starflyer said:

There's definitely a gradient on the Risingcam stack, were both stacks fully callibrated?

Yes, both calibrated with fresh flats, flatdarks and darks, taken the day after.

[Starflyer]

3 hours ago, tomato said:

Yes, both calibrated with fresh flats, flatdarks and darks, taken the day after.

Could it be the OAG casting a shadow, or perhaps a light leak?  That's a fair old gradient with what are very similar cameras.

[ONIKKINEN]

10 offset is way too low for the risingcam, you will have quantization errors with that. The offset should be at least 500 ADUs to counter this. Also, which mode with gain 100? The results are very different with the LCG and  HCG mores on the risingcam. There is a 3x increase in read noise with LCG mode.

[tomato]

2 hours ago, Starflyer said:

Could it be the OAG casting a shadow, or perhaps a light leak?  That's a fair old gradient with what are very similar cameras.

I don't think it is the OAG prism, it does not protrude beyond the step down on the reducer after the OAG and the subs show no sign of vignetting. 

2 hours ago, ONIKKINEN said:

10 offset is way too low for the risingcam, you will have quantization errors with that. The offset should be at least 500 ADUs to counter this. Also, which mode with gain 100? The results are very different with the LCG and  HCG mores on the risingcam. There is a 3x increase in read noise with LCG mode.

I was hoping I might get a response on the offset value from someone with more experience than myself with the camera. I picked up on CN that a small positive value is all that was needed. The default value is 768, is that too high?

I didn't change the mode from default, I confess I don't know what it is set on but I will find out.

[ONIKKINEN]

6 minutes ago, tomato said:

I was hoping I might get a response on the offset value from someone with more experience than myself with the camera. I picked up on CN that a small positive value is all that was needed. The default value is 768, is that too high?

I didn't change the mode from default, I confess I don't know what it is set on but I will find out.

768 ADUs is small for the camera. Since its a 16 bit camera that is only 1.17% of the total available dynamic range so not even remotely something you could notice even with your very long subs. On the subject of long subs, 3 minutes is a bit on the long side. Nothing wrong with that if the mount plays along nicely but you also have no real benefits from shooting such long exposures with these sensitive cameras, unless you are doing narrowband or are shooting from a very dark location (maybe B5 or better). But actually even then as long as you swamp read noise and sky noise you probably are fine with 60s subs. If you change the camera to the HCG mode (you should) you may notice that 3 minute subs blow out starcores, so a more reasonable 1-2 minutes would probably be appropriate. Older CMOS cameras like the 1600MM with a 12bit ADC are very different. With a 12bit camera you only have 4096 values to work with, so it really is crucial what you set there as there is no room to spare. With a 16bit camera you never have to worry about this, so just set it to the default value and forget about it. Its convenient also because if your drivers or settings reset for whatever reason you will automatically be in the correct offset.

Also this could be related to the gradient in your RisingCam stack as the single sub does not have the gradient that the stack has. You probably have non reproducible banding in your darks and bias frames due to the low offset. Offset too low clips blacks as the read noise can result in a value that is less than 0 = which obviously is clipped to 0 and so all information is lost. Read noise is not an added value, but an average deviation from the "right" value. Offset of 10 means that probably somewhere around half of your pixels have lost all useful data with bias and darkframes. Also if you used the LCG mode you had an unnecessarily high read noise to begin with. I found that i could not reliably use my camera below offset 500 because of this. This will probably also make your flats difficult to calibrate since the offset is not stable.

I studied this a bit because im very interested in seeing this comparison roll out but found that this will be difficult/impossible to do "evenly". This is because the QHY version has some really strange design decisions that are very different from the ZWO/other brand offerings. It looks like QHY has gone for optimizing full well capacity instead of read noise like the other brands. Nothing wrong with that, it just means that ideally the cameras would be used differently with different exposure times.

Take a look at this site (one of the only sensor analysis graphs i found): https://eu.lunaticoastro.com/product/qhy-268-m-c-camera/

Then compare the graphs/measurements to the RisingCam measurements in the aliexpress site: https://www.aliexpress.com/item/4001359313736.html?spm=a2g0o.productlist.0.0.6f047164JGhOx6&algo_pvid=88c7fc7f-59b2-4b58-9bdc-a75b08237944&algo_exp_id=88c7fc7f-59b2-4b58-9bdc-a75b08237944-0

You'll find that at pretty much no point are any of the performances in any of the modes for both cameras "equivalent" to each other. Interestingly the QHY camera has consistently more read noise in all modes to the RC.

But my judging of the graphs is that the Rising cam HCG mode at gain 100 and the QHY with the high gain mode at gain 60 are the closest to eachother. The RisingCam in this way has less than 1e of read noise while the QHY will have more than 1.5, but the QHY will have far more fullwell. Fullwell depth is not an issue for most targets (certainly not for M33) and is just dealt with an appropriate exposure time.

The other "fair" comparison would be to set the RisingCam to LCG mode gain 100 and the QHY to the stock(?) photographic mode (mode 0 in the graph) to gain 30. This looks like the same thing for both cameras, although the QHY has a bit more read noise it seems. This might be the closest comparison you could make but in my opinion not really the fairest since the HCG mode is vastly superior to the LCG mode in almost all situations. So you would be doing the QHY a favor and the RisingCam a disfavor with this method.

[tomato]

Thanks for the informative reply. I'll follow your recommendations next time out, but have a full cloud forecast for the next 5 nights, and the moon is getting bigger...