ZWO ASI120mm - Interlace, Interlace everywhere.

[CKemu]

I've had this camera for some time now, about 3-4 years, and initially I used it for some solar imaging, but these days I use it for a guidecam, and it does a perfect job of that, however, tried some solar imaging and it's showing a really grim interlace pattern, I've tried different shutter speeds, gamma settings etc. I've tried different USB ports and even different cables. No matter what, I can't get rid of it. am I missing a trick, is this a software issue (drivers)?

Using FireCapture and ASI2.

 

[Stub Mandrel]

Looks like your driver thinks it's the 120MC.

Or worse, you've muddled up your MC and MM!

[CKemu]

8 hours ago, Stub Mandrel said:

Looks like your driver thinks it's the 120MC.

Or worse, you've muddled up your MC and MM!

Well if certainly says MM on the camera and FireCapture/windows is reporting an MM being plugged in.

[symmetal]

It seems you posted a thread on mentioning this on your ASI120 over 4 years ago when imaging the sun. You didn't find out what caused it since then. 

Is it only there when in imaging in Ha? Is the image section shown (and on your previous thread) at full size (100%) or is it enlarged from that. Also is it a single image taken or one frame from a video? These cameras don't operate in interlaced mode. Basic de-interlacing filters just average successive pairs of lines, so halving the vertical resolution which could hide the effect you have. 

Alan

[CKemu]

13 hours ago, symmetal said:

It seems you posted a thread on mentioning this on your ASI120 over 4 years ago when imaging the sun. You didn't find out what caused it since then. 

Is it only there when in imaging in Ha? Is the image section shown (and on your previous thread) at full size (100%) or is it enlarged from that. Also is it a single image taken or one frame from a video? These cameras don't operate in interlaced mode. Basic de-interlacing filters just average successive pairs of lines, so halving the vertical resolution which could hide the effect you have. 

Alan

No one ever got back to me from ZWO, and no one else had any clues, past few years I have been focusing on deep sky stuff, had a few sunny days, so thought "new rig, new drivers", might be worth a go, nope, still giving me the same garbage performance, incredibly frustrating.

I don't recall seeing the interlace patterns on lunar images, my image is 100% and sharpened, here's another one showing the limb, again sharpened, you can see how awful it is in terms of ruining the view of proms etc.

[symmetal]

Yes, most odd. On this close up you can see that the dark lines have a 1 to 2% drop in value compared to their brighter neighbours. As you say the moon shots don't show it is it triggered by a substantial portion of the image being clipped (sensor overload). CMOS cameras have their A-D (analogue to digital) converters embedded within the sensor chip with one A-D per column, so thousands of converters working in parallel, unlike CCD cameras which have one A-D converter external to the sensor . Whether a column is a row depends on how you view the chip so maybe each line of the CMOS chip in this case has a separate A-D and the A-D reference voltage is different between alternate lines giving these results.

This is just speculation and would point to a faulty sensor or control circuitry. This being an early release of the camera could make this more likely.  It's unlikely to be a cause but I assume you have tried it on a different computer just to eliminate possible USB power issues.

Alan

[CKemu]

22 minutes ago, symmetal said:

Yes, most odd. On this close up you can see that the dark lines have a 1 to 2% drop in value compared to their brighter neighbours. As you say the moon shots don't show it is it triggered by a substantial portion of the image being clipped (sensor overload). CMOS cameras have their A-D (analogue to digital) converters embedded within the sensor chip with one A-D per column, so thousands of converters working in parallel, unlike CCD cameras which have one A-D converter external to the sensor . Whether a column is a row depends on how you view the chip so maybe each line of the CMOS chip in this case has a separate A-D and the A-D reference voltage is different between alternate lines giving these results.

This is just speculation and would point to a faulty sensor or control circuitry. This being an early release of the camera could make this more likely.  It's unlikely to be a cause but I assume you have tried it on a different computer just to eliminate possible USB power issues.

Alan

Yeah, tried running it through two laptops and my desktop, no difference. Sadly this is starting to seem like something out of my control, and will require me to save for some time to get a new mono camera. Annoying, but manageable, it's thankfully solar min currently.

[symmetal]

Yes, I think it needs a new camera CKemu.  They should at least be cheaper than when you bought it before.

Just out of interest, you have an 120MC in your equipment list. You could try taking an image with this under the same solar conditions and compare the pixels on adjacent lines before de-bayering. As a green pixel is every other pixel on each line this should show if it has the same problem as your mono camera. If it doesn't then it does prove a fault with your mono camera.

Alan 

[John78]

This thread with a different camera seems to show an equally odd thing, not sure if @vlaiv ever reached any conclusions, but a mono sensor with colour firmware might produce the effect shown.

 

There are certainly 2 firmware versions for your camera on ASI's site one to try and improve ekos/kstars/indi compatibility, maybe try loading either or both of those.

[vlaiv]

On ASI1600 it is related to pixel sensitivity rather than column A/D or anything like that. I first noticed it on flats - H alpha flats to be more precise, but then I examined my other flats and same thing appeared (I just didn't pay attention to it before since it calibrates out nicely when using flats).

Did you calibrate your movies? This could possibly solve the issue. Pipp can help you with .ser calibration. You need two different calibration files to fully eliminate this if it is either due to bias (I think this might be issue with your ASI120) or due to different QE (like in ASI1600). Just shoot "dark" movie at exact same settings as your regular movie and do that calibration first (no need for bias - darks cover bias and darks effectively on non cooled cameras for such short exposure time). Next thing to do would be to do flats. It still might be a flats issue - different QE of different columns, could be manufacturing thing rather than firmware thing. You probably did not notice it on the moon because signal level was so low compared to this or something similar.

[Oddsocks]

5 hours ago, CKemu said:

Sadly this is starting to seem like something out of my control, and will require me to save for some time to get a new mono camera. Annoying, but manageable, it's thankfully solar min currently.

I have the same two early series ASI120 cameras as you, 120MM and 120MC

Comparing your Ha image in the opening post against two images I took this evening in white light, using MaxImDL for capture with mono camera selected for both MC and MM, it looks as though it might be possible you have a colour detector that has had the mono firmware loaded, though as I don't know how much you stretched the image I can't be really be sure. My MC image in white light won't be exactly the same as your Ha of course since you will only be recording primarily in the red while I was capturing in red green and blue.

A really quick test you can do in daylight is to remove the camera from the telescope, do not use the supplied wide angle lens just use the camera with the sensor exposed and point the camera at a brightly lit primary coloured object, red, blue or green, held a few centimetres in front of the sensor and in live view mode look at the histogram display.

A monochrome sensor lit with a primary colour will show a histogram that contains just a single peak, as in the image attached below, a colour sensor will show two or three separated histogram peaks when lit with a primary colour, also shown below.

It is not beyond the realms of possibilities that during a production run a colour sensor board ended up being flashed with the mono firmware and labeled as such.

Another quick test you can do is to look at your MM and MC side-by-side, I really wanted to post a picture for you but my camera just won't focus closely enough so a description will have to do, tilt the camera so that a bright light is reflecting at a sharp angle off the sensor so that you see the refection of the light in the sensor, the colour sensor is considerably darker than the mono sensor when the angle of reflection is the same. If both the sensors look around the same brightness that is another indicator that they are both colour sensors.

I final test might be to download the ASI120MM and ASI120MC firmware from ZWO's website, they still have the firmware load tool and both the MC and MM updated firmware for the early versions of the ASI120MM/MC. It would be interesting to load the MC firmware, connect the camera and wide angle lens and see if it captures a normal colour image, if so that proves you have a colour sensor in a camera pretending to be a mono. if not, then reload the mono firmware. AFAIK the only difference between them is the device ID string that tells the capture software whether a colour or mono camera is connected so that it knows whether to debayer the image or not. I must warn though that reloading flash firmware is not without its risks, as flash memory frequently fails during erase and load operations, as long as you are aware of the risk it might be something to consider. I have re-flashed both my MC and MM cameras, the later firmware is supposed to help with the capture lock-up bug and blue-screen-of death-problems on Windows 7 64bit that the early cameras had.

Comparison image below:

HTH

William

 

 

[symmetal]

Interesting comparison William. On the colour version you can see the green pixel has the similar brightness on each line. The green pixel next to the red one has a slightly higher output compared to the one next to the blue pixel probably due to light leakage, hence the two green peaks on the histogram. On the 'funny' mono camera all the pixels on the dark lines are darker than the ones above or below implying there is no similar 'green' filter on each line. If there is a bayer matrix present on CKemu's camera it's not the standard one. 

Alan

[Oddsocks]

26 minutes ago, symmetal said:

If there is a bayer matrix present on CKemu's camera it's not the standard one. 

Hi Alan.

My quick test is not really a direct comparison because I think CKemu’s posted image is from a Hydrogen Alpha solar telescope, so primarily only the red pixels should be active, though I have no real idea under what conditions the image was produced, how much light leakage into the optical path, exposure time, gain setting, thermal dark current etc, whereas my sample images were taken in white light illuminating a dark red object at just 2ms exposure time and minimum gain.

A real test would be to compare a white light image of a deep red object at similar camera settings with the samples from my MM and MC cameras. It just seemed that CKemu’s image looked so much like a bayered image when captured in mono mode to be just random chance.

An interesting problem nonetheless. 

William.

[symmetal]

36 minutes ago, Oddsocks said:

Hi Alan.

My quick test is not really a direct comparison because I think CKemu’s posted image is from a Hydrogen Alpha solar telescope, so primarily only the red pixels should be active, though I have no real idea under what conditions the image was produced, how much light leakage into the optical path, exposure time, gain setting, thermal dark current etc, whereas my sample images were taken in white light illuminating a dark red object at just 2ms exposure time and minimum gain.

A real test would be to compare a white light image of a deep red object at similar camera settings with the samples from my MM and MC cameras. It just seemed that CKemu’s image looked so much like a bayered image when captured in mono mode to be just random chance.

An interesting problem nonetheless. 

William.

You're right William. I was forgetting it was through an Ha filter.  In that case it would show how good the bayer filters really are as the blue output is still fairly high.

Alan