QHY600PRO binnig 16 bit 18 bit 19 bit 20 bit error

Post Reply
deneb
Posts: 15
Joined: Tue Jul 14, 2020 5:58 pm

QHY600PRO binnig 16 bit 18 bit 19 bit 20 bit error

Post by deneb »

Hi Robin.
I have a qhy600 pro and I like photometry.
The subject for which I write to you is the following.
When trying to perform photometry with this camera, the manufacturer states that the equipment through software can have 18 bits of resolution at 2x2 binning, 19 at 3x3 binning and 20 4x4 binning.
In testing with sharcap, I can only get to 65535 which corresponds to 16 bit in any read mode.
I think it may be qhy's sdk, but I don't know.
If the camera manufacturer's characteristics are correct, it would have to be
Native depth 65535
Binning depth 2x2 262 140
Binning depth 3x3 589,815
Binning depth 3x3 1,048,560
Is there a problem that can come from the conversion by limitation in sharcap ?, or as I suspect, is it a problem with the QHY SDK that does not change the format to 32Bit when binning?
User avatar
admin
Site Admin
Posts: 6954
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: QHY600PRO binnig 16 bit 18 bit 19 bit 20 bit error

Post by admin »

Hi,

I suspect that the SDK may actually support a 32 bit readout, but SharpCap certainly has no code to enable or use that feature. I can see something in the SDK definition that *might* be related to this, but there is no mention of it in the documentation :(

Of course, since the camera uses a CMOS sensor, there is no actual image quality benefit to binning as there is in CCD. The higher bit depth values are just obtained by digitally adding up the readout values from 4 or more adjoining pixels. You could obtain identical data by capturing at full resolution and then binning to higher bit depth in post-processing (I expect PixInsight would cope with higher bit depths, possibly other software too).

There will be a benefit in reduced file size and possibly readout time that you miss out on by capturing at full resolution of course.

cheers,

Robin
deneb
Posts: 15
Joined: Tue Jul 14, 2020 5:58 pm

Re: QHY600PRO binnig 16 bit 18 bit 19 bit 20 bit error

Post by deneb »

Hi Robin.

I would like to define hoaxes or little information about binning and I would love for you to remove or agree with logic.

In CCD analog cameras. Binning was misused or rather, taken advantage of.

For years, binning has been done digitally, since before I think it was done analogically using storage blocks.

The way would be.

Analog data arrives from the CCD.

they are amplified by the programmable gain amplifier.

It is converted to digital by the "8bit, 10bit, 12bit, 16bit" analog to digital converter and stored in a buffer.

After this step, the microprocessor structures the blocks that will create the pixel binning and will save it in another memory block that will be the final image.

This system has the advantage that it increases the sensitivity because it adds the values ​​which is the same as increasing the surface of the pixel. It also improves the signal-to-noise ratio but not to a great extent.

But this system does not get all the power that it would have if it added the values ​​but in 32 bits.

Remember that each pixel value after the 16bit converter would have 65535 levels.

When the binning system in CCD reaches its maximum bit capacity, it no longer goes up as it is structured at 16 bits.

It is true that linearity is better in binning since the bad linearity is due to the semiconductor or pixel when reaching the limitation of the anti blooming or maximum of the quantum well.

You have to think that in 2x2 binning, a virtual pixel with 262,140 values ​​would come out. Which due to quantification limitations remains at 65535 values. The remaining signal is neglected as it disappears in the sum.

Therefore we have used a ¼ signal from each pixel.

If the internal logic of addition were quantized to 32 bits, those 18 bits could be reached with a resolution of 2x2 binning and the remaining .262,140 values ​​or 589,815 values ​​in the case of 3x3 binning or 1,048,560 values ​​in the case of 4x4 binning. .

It is true that linearity for photometry would leave, in the case of useful 2x2 binning, values ​​up to 199,226 levels. 76% more or less of the signal, but you would have 3 times more dynamics than in a classic 16-bit system.

Imagine the depth or rather the contrast that can be achieved.

The famous STOPs skyrocket 17 .. 18 .. 19 ..

That is to say, at the photometric level, to be able to analyze stars of very low magnitudes and very high magnitudes without losing linearity and increasing sensitivity in all their ranges.

In a cmos it is exactly the same, except in qh and that theoretically add the values ​​but in 32 bits.

This is a very important improvement that nobody has been able to take advantage of correctly.


In CCD it is ridiculous to have a quantum well of 200,000 values ​​per pixel since you will never use it. Unless the analog converter has more bits. Something that the manufacturers of astronomical cameras have failed to take advantage of.


So if it is not much of a problem. You could implement this 32-bit binning idea.

As in the sensor analysis, it can be evaluated.

I don't know if I'm wrong in what I say.
But how beautiful it is to dream.
I control all this after spending the whole night doing astronomy and not sleeping.
forgive if I say something outrageous
User avatar
admin
Site Admin
Posts: 6954
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: QHY600PRO binnig 16 bit 18 bit 19 bit 20 bit error

Post by admin »

Hi,

my understanding is that CCD sensors bin by collecting the electric charge from multiple pixels (all the pixels making up a single binned pixel) into the ADC at once. That is, if your individual pixels are collecting ~100e and you are using 2x2 binning then ~400e gets delivered to the ADC from four pixels for conversion to digital. The advantage of this approach is that you only get one set of read noise. As you mentioned, this is not helpful for bright images as you saturate at 1/4 of the normal full well depth, but it is helpful for faint things.

CMOS binning is always digital, after the ADC. In the same scenario, you do 4 ADC conversions of ~100e each, each readout value has it's own dose of read noise added. This would be a big problem, except for the fact that CMOS tends to have much lower read noise than CCD. Once you have the 4 ADC values, the 'binning' can either be adding them together or averaging them (different manufacturers do it different ways).

Anyway, binning is not usually a technique that is used to try to enhance the FWD, more to bring out faint stuff more quickly.

cheers,

Robin
deneb
Posts: 15
Joined: Tue Jul 14, 2020 5:58 pm

Re: QHY600PRO binnig 16 bit 18 bit 19 bit 20 bit error

Post by deneb »

Hi Robin.
It seems that they already have the function you asked for integrated.
Greetings

https://www.qhyccd.com/user-manual-of-3 ... 11-qhy461/
User avatar
admin
Site Admin
Posts: 6954
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: QHY600PRO binnig 16 bit 18 bit 19 bit 20 bit error

Post by admin »

Hi Deneb,

thanks for spotting that - I will put high bit depth onto the to-do list. Unfortunately adding support at the 'talking to the QHY SDK' is only part of it - lots of other parts of SharpCap expect images to either be 8 bit or 16 bit, so there is quite a bit of work to do to cope with 32 bit images throughout.

cheers,

Robin
Post Reply