Jupiter 130920 - First one I dare share

[wornish]

Had another go last night and think I am making progress but I can't seem to get any of the detail others on here do.

 

This is taken using Firecapture with my ASI290MC on my HD925 best 10% of over 20000 frames stacked in AS3 and Wavelet sharpened in Registax6.

What am I doing wrong

 

 

Edited September 14, 2020 by wornish

[vlaiv]

What sort of exposure length are you using?

Above image is way to large for HD925 and current Jupiter apparent diameter, what sort of barlow are you using?

[AbsolutelyN]

It's better than my last Jupiter the other day. Suspect it is simply because Jupiter is currently so low in the sky and you're looking through tick turbulent atmosphere?

Edited September 14, 2020 by AbsolutelyN

[wornish]

I did increase the image size in PS. The original was 640 x 480

I have attached it here.

I wasn't using a Barlow.  

The details of the file created by FireCapture are:

2020-09-13-1932_1-DT-IR-Jup_ZWO ASI290MC_Gain=401(off)_Exposure=1.4ms.ser

The output info from AS3 was:

ASI290MC_Gain=401(off)_Exposure=1.4ms.ser
  Stack quality 0.00000000 
  Median quality 51.22 ( quantiles 47.68, 51.22, 55.87 )
  Frame count 33703.

The image was a stack of the best 10%.

I did have the ZWO 1.25 IR Cut filter in front of the sensor but that would make things better not worse I hope.

Seeing last night was the best for months, an 8 on the scale from 0 - 10 as published by met check

http://www.metcheck.com/HOBBIES/astronomy_forecast.asp?zipcode=Congleton&locationID=57638&lat=53.2&lon=-2.2

[vlaiv]

1.4ms seems like rather short exposure but that should be no problem (except for increased noise). I was worried you used too long exposure.

How many alignment points did you use? Try with different number of alignment points in AS!3

Maybe also attach raw stack (if you wish) just to check if processing is the issue.

[wornish]

Thanks for the suggestions.

I only used about 8 align points.

I have attached the raw stack as tif file

2020-09-13-1932_1-DT-IR-Jup_ZWO ASI290MC_Gain=401(off)_Exposure=1.4ms_lapl6_ap30.tif

[vlaiv]

Try more alignment points.

Things improve a bit if you do RGB align - it fixes some of atmospheric dispersion. As is, there is too much blurring - if I try to sharpen it further - noise just comes to surface.

In above image, you just pushed wavelets too much. Btw - there is a moon just of the limb of Jupiter? (top left) - if so - place an alignment point there as well.

[wornish]

Thanks for your help.  I will re process in AS3 again. 

Think I might've another go tonight as its forecast to be good seeing again and that happens once a blue moon here.

[vlaiv]

Do one recording with gain 317 and 6ms exposure length just to see how that stacks up.

Leave all other settings on default (no brightness / gamma / white balance adjustments).

[wornish]

Managed to have another go these were taken at Gain 317 exposure 4.16ms. ( pushing it to 6  maxes out the histogram and everything seemed to be just white so didn't give it a go)

22000 frames.

Best 5 %,  25%. and 80%

 

 

[Tommohawk]

21 hours ago, wornish said:

What am I doing wrong

First thing to say is that Jupiter is obviously a bit of a challenge at the moment, so cut yourself some slack - those last images in particular are pretty good!

A few thoughts. I'm guessing you dont have an ADC? I dont either but your images will suffer as a result - AS3 or Registax will align the RGB components but this doesn't entirely fix the dispersion issue. When imaging, the many various wavelengths get smeared across the sensor, but when processing you can only change the basic RGB components, so its not as good a fix as an ADC.

Are you capturing at maximum elevation? Also re exposure, I've found with that camera and the mono counterpart you can use very high gain for minimum exposure. 1.4ms sounds about right.

Cant see what length AVI you did. 

19 hours ago, wornish said:

ASI290MC_Gain=401(off)_Exposure=1.4ms.ser
  Stack quality 0.00000000 
  Median quality 51.22 ( quantiles 47.68, 51.22, 55.87 )
  Frame count 33703.

Seems to suggest about 45 seconds? I'd suggest longer. Even without derotation you should be able to do 120secs. Maybe youre doing longer but FC is dropping some frames?

 

[vlaiv]

3 minutes ago, Tommohawk said:

I've found with that camera and the mono counterpart you can use very high gain for minimum exposure. 1.4ms sounds about right.

Exposure length should be as long as possible without loosing ability to do lucky imaging. Maximum exposure time depends on something called coherence length / coherence time for particular location and particular sky conditions.

It depends on seeing cell size vs aperture size used and speed at which these cells move. This is why jet stream is not good - it moves cells at very fast rate thus shortening coherence time extremely.

If one goes with shorter exposure time they are just lowering their SNR without any good reason. Exposure length needs to be long enough to "freeze" the seeing without effects of motion blur which happen when successive cells move over aperture in single exposure. In any case - in most conditions and on most locations 5-6ms is coherence time for 8" aperture. 10ms exposure is reserved for those nights of excellent seeing.

With 1.4ms exposure - you are in effect not gaining anything with respect to lucky imaging, and on the other hand you are increasing your effective read noise at least two fold (1.4ms * 4 = 5.6ms - stacking 4 exposures of 1.4ms is equivalent of capturing single 5.6ms exposure but with x2 read noise)

9 minutes ago, Tommohawk said:

Seems to suggest about 45 seconds? I'd suggest longer. Even without derotation you should be able to do 120secs. Maybe youre doing longer but FC is dropping some frames?

Only if camera can operate with 700+ fps - which it can't (at least not on 640x480 ROI). USB speed will be limiting factor here.

Zwo website states:

640×480: 184/381.2

as max achievable fps for 640x480 in USB 3.0 version. I would say that real FPS was about 200-250 if 8bit mode was used (due to computer and storage speed). Which means that effective time for each sub was 4ms (again - this shows that you are wasting imaging time if going with short subs). This would put video at two minute duration which is ok, and I would suggest going up to 4 minutes with AS!3 as it can compensate some of rotation with alignment points.

[Tommohawk]

Interesting stuff Vlad - your knowledge of this is beyond mine for sure so I will defer! I've always tried to freeze the seeing with the shortest possible exposure, even in some cases if the camera cant handle it. But sounds like there is a limit.

Hopefully Mars will remain good and I can experiment to check the practical effects of this. Storage is certainly an issue - I did Jupiter followed by Saturn and then Mars and got through about 200Gb. 

[wornish]

34 minutes ago, vlaiv said:

Exposure length should be as long as possible without loosing ability to do lucky imaging. Maximum exposure time depends on something called coherence length / coherence time for particular location and particular sky conditions.

It depends on seeing cell size vs aperture size used and speed at which these cells move. This is why jet stream is not good - it moves cells at very fast rate thus shortening coherence time extremely.

If one goes with shorter exposure time they are just lowering their SNR without any good reason. Exposure length needs to be long enough to "freeze" the seeing without effects of motion blur which happen when successive cells move over aperture in single exposure. In any case - in most conditions and on most locations 5-6ms is coherence time for 8" aperture. 10ms exposure is reserved for those nights of excellent seeing.

With 1.4ms exposure - you are in effect not gaining anything with respect to lucky imaging, and on the other hand you are increasing your effective read noise at least two fold (1.4ms * 4 = 5.6ms - stacking 4 exposures of 1.4ms is equivalent of capturing single 5.6ms exposure but with x2 read noise)

Only if camera can operate with 700+ fps - which it can't (at least not on 640x480 ROI). USB speed will be limiting factor here.

Zwo website states:

640×480: 184/381.2

as max achievable fps for 640x480 in USB 3.0 version. I would say that real FPS was about 200-250 if 8bit mode was used (due to computer and storage speed). Which means that effective time for each sub was 4ms (again - this shows that you are wasting imaging time if going with short subs). This would put video at two minute duration which is ok, and I would suggest going up to 4 minutes with AS!3 as it can compensate some of rotation with alignment points.

Thanks for explanation.  

I will do at lest a two minute run first and push the exposure time up to 6ms and perhaps drop the gain a touch.  I have tried to keep the histogram at about the 70% mark up till now but I think I will push it towards 90% to see what happens.  If conditions allow I will have a go at a 3 minute run as well.  

One thought could the IR cut filter be affecting the quality of the image? I have seen others say the camera benefits by using one hence I fitted the ZWO one right next to the sensor.

 

Think I will splash out on the ADC as well.

[vlaiv]

5 minutes ago, wornish said:

Thanks for explanation.  

I will do at lest a two minute run first and push the exposure time up to 6ms and perhaps drop the gain a touch.  I have tried to keep the histogram at about the 70% mark up till now but I think I will push it towards 90% to see what happens.  If conditions allow I will have a go at a 3 minute run as well.  

One thought could the IR cut filter be affecting the quality of the image? I have seen others say the camera benefits by using one hence I fitted the ZWO one right next to the sensor.

 

Think I will splash out on the ADC as well.

Histogram is really not important when stacking. Main thing to use histogram for is to see if there is clipping. If there is no clipping - it does not matter if histogram is at 10% or 90%.

It is gain with associated read noise that you should be focusing on, read this for better explanation of what gain does to read noise:

and exposure time.

Planetary imaging is balance of the two - shorter exposures mean better freezing the seeing but increase noise which prevents you from sharpening later on, and longer exposures get you opportunity to sharpen more because of lower noise but it also contributes to actually blurring image more due to seeing.

Best case scenario is stay below seeing induced blur (freeze the seeing) and yet have max exposure length possible to lower the noise and help with sharpening.

It looks like ASI290 does not have UV/IR cut filter, but simple AR coated window. You should be using proper UV/IR cut filter. Not sure if the one you are using is only IR cut filter but if it is - get proper UV/IR cut filter to filter both UV and IR part of spectrum - both can lower contrast and blur out features.

[Tommohawk]

4 minutes ago, vlaiv said:

Histogram is really not important when stacking.

I think OP meant when capturing?

FWIW 70% is what I aim for - most folk seem to feel there is no need to go higher than this.  

[vlaiv]

Just now, Tommohawk said:

I think OP meant when capturing?

FWIW 70% is what I aim for - most folk seem to feel there is no need to go higher than this.  

Yes, both OP and I said what we meant - but I was probably not clear enough.

What I meant was - histogram at the time of the capture is not important if you plan to stack your images later. It is important (to some degree) if you take a single exposure (and that depends on dynamic range of the sensor).

Here is a brief explanation - histogram just shows how ADU values behave with respect to some range. If you use full range - then with respect to full range of values, either 8bit or 16bit.

It does not show anything about electron count value - higher gain will produce higher ADU values for same electron count, and consequently it's not telling you anything about SNR of current sub (which is very much related to electron count).

To show how meaningless histogram is, let's examine this example - 8bit mode vs 16 bit mode capture. Gain is set the same, exposure is set the same, everything is set the same except that 8bits are output in first case and 16bits are output in second case.

Let's assume that majority of pixels have ADU value of 220. In 8bit mode, which has range of 0-255, this 220 is equal to 86.3% histogram peak. Nice, well "exposed", everything seems fine. How about if we use 16bit values? In that case 0-65535 is range of possible values, and 220 out of 65535 is in fact 0.3357% - less than 1%, image would be completely black if stretched to full range! No one in their "right mind" would expose for 0.33% of histogram right?

But if you examine data - it is exactly the same data, same numbers, same SNR, completely same image. It is interpretation of histogram that changed. In fact - histogram is interpretation and it has its meaning with DSLR cameras when you shoot single image and want it to be properly exposed and you don't consider dynamic range but want your image straight out of camera and so on ... then it shows something meaningful as it is tied to particular interpretation.

When stacking and otherwise processing your image - then histogram at recording time is not that important, or rather - it is not important at all and you should be using it only to make sure you are not saturating sensor as saturated pixels are lost pixels - no way of knowing what value was in those pixels and thus this condition should be avoided.

[runoffshed]

Hi Dave, I am far from being a particularly experienced imager but I wonder how good the focus was to start with?  I was out the other night imaging Mars and had a hell of a job getting decent focus as the seeing was poor and I don't have a remote/electronic focuser so the image was leaping all over the place.  I must have spent about 10 minutes before each run going back and forth trying to nail it.  I probably never did and am going to try a 'Y' mask the next clear night.  Do you use any mask/software to help with focus?

Anyway, just a thought and good luck.

 

John

[wornish]

1 hour ago, runoffshed said:

Hi Dave, I am far from being a particularly experienced imager but I wonder how good the focus was to start with?  I was out the other night imaging Mars and had a hell of a job getting decent focus as the seeing was poor and I don't have a remote/electronic focuser so the image was leaping all over the place.  I must have spent about 10 minutes before each run going back and forth trying to nail it.  I probably never did and am going to try a 'Y' mask the next clear night.  Do you use any mask/software to help with focus?

Anyway, just a thought and good luck.

 

John

On my refractor I have an electronic focuser which works great.  But on the HD925 I am using my eyes and manual focussing.  I got the scope specifically to have a go at planetary and lunar imaging and was aware of the focus challenges.  This why I also got the Baader Steeltrack focuser to try and make manual focussing a bit easier.  I use the built in focuser on the back of the scope to get the best focus I can and as you say the image does jump around all over the place when adjusting it that way. But then I use the fine adjustment on the Steeltrack to try and get even sharper focus and using that method the image stays put.  As the scope is new I haven't tried to do any collimation and when I look at a star it seems to be OK.  But I think you might be right getting it spot on is a challenge and not helping.  The focus assistant tool in Firecapture and the one on Sharpcap don't impress me but I haven't really had a serious go using them.

I see the images posted on here and get so frustrated that I can't seem to get anywhere near the level of clear detail they show..  I know some are taken in other countries and that the jetstream is a nightmare in the UK.  But there are people who live in nearby counties to me that seem to be able to just nail the imaging.  It's all part of the challenge and fun and it would be boring if everything just worked:)   I might take the advice posted earlier and get the Atmospheric Disturbance Corrector.  Would like to hear from anyone that's using one if they really do make a difference, ideally with pictures to show an image taken with and without one.

Edited September 15, 2020 by wornish

[runoffshed]

Sounds like you're a lot better set up for good focusing than I am Dave.  

I know what you mean about the frustration.  My recent images of Mars were the best I've taken but they aren't a patch on some of the stuff posted on here recently.

Oh well, as you say, wouldn't be fun if it was easy!

Atb.

John

[wornish]

1 hour ago, vlaiv said:

 

It looks like ASI290 does not have UV/IR cut filter, but simple AR coated window. You should be using proper UV/IR cut filter. Not sure if the one you are using is only IR cut filter but if it is - get proper UV/IR cut filter to filter both UV and IR part of spectrum - both can lower contrast and blur out features.

Thanks for the extra  info.

The filter I have is the ZWO 1.25 IR Cut not the  UV/IR Cut.  I am going have a go with it removed to see if it is contributing to the blurring. 

Depending on the outcome then will get the UV/IR cut as you suggest.

[Tommohawk]

59 minutes ago, wornish said:

The filter I have is the ZWO 1.25 IR Cut not the  UV/IR Cut.

Sounds the same as mine - it's the ZWO one marked IR cut, and I can't find the spec for this to check if it cuts UV too- it does seem to work well though.

I did find this though on the ZWO forum. It's dated 2013 so I think ours must be that version - and the graph posted shows that this version cuts UV too.

Hope that helps.

Edited September 15, 2020 by Tommohawk

[wornish]

2 hours ago, Tommohawk said:

Sounds the same as mine - it's the ZWO one marked IR cut, and I can't find the spec for this to check if it cuts UV too- it does seem to work well though.

I did find this though on the ZWO forum. It's dated 2013 so I think ours must be that version - and the graph posted shows that this version cuts UV too.

Hope that helps.

Excellent news.  Yes mine is the one called ZWO IR Cut.   

Update:  The box says IR Cut but just looked at the actual filter and printed on the side it says UV/IR Cut so that's good news.

Edited September 15, 2020 by wornish

[wornish]

Had another go last night.

This is using 317 gain and 6ms exposure best 5% of 20000 frames.

Think its a bit better but  still not what I hoped for

[wornish]

Another go tonight.  Gain 317, 6.2ms exposure.  20,000 frames stack of 25%