Jump to content


Photo

ASI120MM vs DMK21AU618

  • Please log in to reply
14 replies to this topic

#1 mak5

mak5

    Mariner 2

  • -----
  • Posts: 276
  • Joined: 09 Mar 2009

Posted 15 February 2013 - 09:12 AM

I have borrowed from a friend his new camera, the ASI120MM, to compare it to my DMK21AU618. The target was Jupiter on some medium seeing sky. To get more or less the same sampling, I shot at f/13 with the ASI and f/22 with the DMK. I expected to get more fps with the ASI because of the smaller and brighter image at f/13, but I only managed to get 50fps at 480X320 vs 60fps with the DMK (both cases ~ 60% histogram). Gamma was 50 and gain about 53-55 with the ASI. I used a C8 and Astronomik filters in both cases.
I’ve also got some strange pixel pattern on the planet after sharpening on the ASI120MM stacks.
Anyone has an idea why with a brighter image I got less fps and what causes the pattern?

Attached Files



#2 wenjha

wenjha

    Vendor

  • -----
  • Vendors
  • Posts: 732
  • Joined: 24 Sep 2008
  • Loc: SuZhou China

Posted 15 February 2013 - 09:27 AM

you can turn the gain higher to get fater fps. 53-55 gain is too less.
the pattern usually because of very low signal to noise ratio of your image.you can choose a higher noise Robust in stack software

#3 pippo

pippo

    Explorer 1

  • -----
  • Posts: 94
  • Joined: 12 Jan 2013
  • Loc: Paris (FR), Pantelleria (IT)

Posted 15 February 2013 - 09:36 AM

In principle increasing the gain will not improve the signal/noise ratio, since with an histogram of 60% the noise due to the digitization should be smaller than that associated to the readout.

#4 PiotrM

PiotrM

    Gemini

  • -----
  • Posts: 3181
  • Joined: 03 Jan 2010
  • Loc: Poland

Posted 15 February 2013 - 11:14 AM

If you set the same image scale for both cameras then faster f-ratio for the ASI (with smaller pixels) won't give brighter image. Gamma should be at default value. With the DMK21AU618 it's easy to do 60FPS Jupiter recordings at f/20. I used 1/60 sec exposure and just lowered gain if needed (for broad filters a lot). And I usually keep histogram at 80% or more (excluding problematic channels like blue).

The rolling shutter in ASI may sometimes give artifacts if the objects wasn't stable.

#5 Kokatha man

Kokatha man

    Fly Me to the Moon

  • *****
  • Posts: 6859
  • Joined: 13 Sep 2009
  • Loc: "cooker-ta man" downunda...

Posted 15 February 2013 - 08:16 PM

A gain of "53" would seem to be far too low for a C8 imho mak - even around opposition with a C14 I employed that value for an 80% histogram so the C8 would logically demand a considerably higher gain.....which of course brings us to the question of 60% Vs 80% histograms - I think that Sam is (or should be!) referring to this factor as possibly introducing artefacts rather than "noise per se" pippo.....

Piotr keeps "rolling out" the ASI's "rolling shutter" and I keep "rolling it back to him" :lol: with the response that I have never once seen any such "artifacts" (sic) despite often running a very unstable onscreen image due to my transportable imaging activities :shocked:.....I think it is a bit of a game between him and me :grin: but to paraphrase someone else's comments "I have most probably done more imaging with the ASI120MM than just about anyone else" - so you'd think I would have encountered it if it was a possible problem! :question: :)

More importantly, if you could only get 50fps @ 480X320 using the ASI120MM (despite optimal histogram settings etc) then you haven't set Firecapture up correctly - presuming you used F/cap!

There is a little "wheel cog" on the RHS of the exposure window in F/capture which you can see in a screenshot I did for Sam on the ZW Optical page.....this must be clicked on "High Speed" to get the correct frame rates. (this is altered for the new drivers that I'm currently testing but for all drivers up to now this action applies...)

The only somewhat similar effect to the one you've posted is when heavy wavelets are applied to the image - but this is when using linked wavelets, and before applying denoise which also removes said.....here is an example of such.

Also of course you must set the exposure at the appropriate value for any particular frame rate.....I find these cameras accept variations to this but if you change the exposure markedly from "parity" with fps (ie, 100fps=10milliseconds...50fps=20mS etc) then there are very considerable drops in the frame rates attainable...

Attached Files



#6 DesertRat

DesertRat

    Fly Me to the Moon

  • *****
  • Posts: 5193
  • Joined: 18 Jun 2006
  • Loc: Valley of the Sun

Posted 15 February 2013 - 08:41 PM

Piotr keeps "rolling out" the ASI's "rolling shutter" and I keep "rolling it back to him"



Piotr has every right to bring up the point. I know you Darryl well enough by now to know you would agree with that! However I think it would be more useful to address it a little more thoroughly.

A rolling shutter generating artifacts is only possibilty at this time. Most of us have seen examples of a rotating fan or a speeding car showing this effect. Thus far I have not seen an example of it in a planetary image. Somebody should do the math.

I only have a guess that a planet would really have to be moving at an incredible speed about the frame to generate distorted elements from a rolling shutter. At the frame rates you and others have been using it does'nt sound likely especially in any seeing condition even halfway decent - but I would like to hear what others think.

This kind of thing could be addressed experimentally, mathematically treated or modeled. Anyone? I only suggest this because it would be nice to have some basis which Sam and others could address on their web pages so we don't have the question come up over and over again - like other frequently asked questions, myths and speculations do on this and other forums.

Glenn

#7 wenjha

wenjha

    Vendor

  • -----
  • Vendors
  • Posts: 732
  • Joined: 24 Sep 2008
  • Loc: SuZhou China

Posted 15 February 2013 - 09:20 PM

increasing the gain will decrease the signal/noise ratio. but lower exposure maybe freeze the seeing.and stack more images will increase the singal/noise ratio. so which point is the balance?maybe need to have more tries.from my experience, keep the gain from 60-80 is a normal range.ZuoZhao usually turn it to 100 go get a fast fps in a low transparence sky.So maybe you need to find the best way suitable for you.

In principle increasing the gain will not improve the signal/noise ratio, since with an histogram of 60% the noise due to the digitization should be smaller than that associated to the readout.



#8 Kokatha man

Kokatha man

    Fly Me to the Moon

  • *****
  • Posts: 6859
  • Joined: 13 Sep 2009
  • Loc: "cooker-ta man" downunda...

Posted 15 February 2013 - 09:40 PM

Piotr keeps "rolling out" the ASI's "rolling shutter" and I keep "rolling it back to him"



Piotr has every right to bring up the point. I know you Darryl well enough by now to know you would agree with that! However I think it would be more useful to address it a little more thoroughly.

I only have a guess that a planet would really have to be moving at an incredible speed about the frame to generate distorted elements from a rolling shutter. At the frame rates you and others have been using it does'nt sound likely especially in any seeing condition even halfway decent - but I would like to hear what others think.

This kind of thing could be addressed experimentally, mathematically treated or modeled. Anyone? I only suggest this because it would be nice to have some basis which Sam and others could address on their web pages so we don't have the question come up over and over again - like other frequently asked questions, myths and speculations do on this and other forums.
Glenn


Of course he has that right Glenn.....and within the context of my posting commentary above I think my response was "fair & reasonable" also. :)

As you know, I'm primarily a "nuts & bolts" man who applies whatever commonsense & logic I can muster on whatever I engage in/encounter.....I have commented numerous times on this "topic" and perhaps I should reiterate here one of my oft-stated experiences...

BUT - "mea culpa" :rainbow: - I have been a little loose with the truth hereon.....despite saying I have never seen any "rolling shutter" effect I cannot (within my knowledge/experiences) rule out that there was one instance relatively recently where this could have been an influence.....but I'm fairly unconvinced on this specific occassion..! :lol:

Upon loading and quality analysing a Jovian capture in AS!2 about a month or 2 back the best frame was a "half-image" which either was subsequently "spat-out" or made no lasting effect upon the stacked result. (which was quite good incidentally)

However, I tend to associate that instance to the "frame-tearing" that can occur with the PGR Flea3 occassionally which I had a little instance of one night back in Tennant Creek...ruining some very fine captures because I wasn't aware of it untill the next day - fortunately it was only a "passing pecadillo" of the camera which rectified itself the next time I used it... :)

This (the ASI120MM phenomenum) was on a very windy night and I suspect would have definitely qualified for your "moving at an incredible speed about the frame" instance - in fact it was literally disappearing from the screen on numerous occassions during the recording (an amazing testament to AutoStakket's capabilities!) and although 60/80 fps was probably high enough to "freeze out" a lot of situations I have shot as low as 40fps (to see how lower gain/noise captures compare and also give me that larger "screen" to accomodate the wild onscreen image movements in windy conditions) without finding any situations which could be interpreted as such.....

In the words of that song "more than this I cannot say" :)
but I don't think many of those myths are likely to go away terribly soon.....I know of one fella on another forum who is certain that recordings of more than 30 seconds per channel at his (exceedingly small) image scale will introduce rotational blurring despite much articulation from several others contrary to his beliefs on this subject :shocked: :foreheadslap:

I'm happy to stand corrected on anything but don't feel I respond with much less than a fairly comprehensive reply most of the time - possibly many would say far too comprehensive :lol: - but you can only say what you know or can deduce from, taking the situations into account..... :question: :)

#9 mak5

mak5

    Mariner 2

  • -----
  • Posts: 276
  • Joined: 09 Mar 2009

Posted 16 February 2013 - 05:26 AM

Well, I should have begun with the technical details, sorry. I used Fire Capture 2.2 build #37, from what I remember, I did the "high speed trick" for every filter and the R, G and B avis all have 6000 frames each. I used 1500 frames for each colour.
I stopped with the gain at 55 because I saw the noise was begining to increase. I will try next time with a higher gain to see what effect it has (and hoping the noise will go away after processing).
I still don't understand why a f/13 image is not brighter than a f/22 one on a sensor more sensitive in green light on the f/13 telescope. Can anyone explain this?
I would also greatly appreciate if Darryl and other ASI users will post some typical gain and exposure values to have an idea where to start. Thanks in advance.
Next target will be Saturn for me so things will be even tougher...

#10 PiotrM

PiotrM

    Gemini

  • -----
  • Posts: 3181
  • Joined: 03 Jan 2010
  • Loc: Poland

Posted 16 February 2013 - 05:58 AM

You don't need a fan blade or car speed to create artifacts from rolling shutter - bad seeing will do too. And it won't be anything obvious, just a blur/distortion by pixel or two.

Good case would be Sun white light imaging with a big scope ;)

#11 wenjha

wenjha

    Vendor

  • -----
  • Vendors
  • Posts: 732
  • Joined: 24 Sep 2008
  • Loc: SuZhou China

Posted 16 February 2013 - 08:40 AM

compare 2 different camera is no easy because they have different pixel size and gain setting.
here is also a very detail compare thread between ASI120MM and DMK618
http://translate.goo...&prev=_t&amp...

#12 wenjha

wenjha

    Vendor

  • -----
  • Vendors
  • Posts: 732
  • Joined: 24 Sep 2008
  • Loc: SuZhou China

Posted 16 February 2013 - 08:44 AM

this depends on exposure time and how fast is the rolling shutter and how big is the sensor's resolutions.
I will get more data and reply later.because I want to know the detail too although I feel no different from my experience :)

#13 mak5

mak5

    Mariner 2

  • -----
  • Posts: 276
  • Joined: 09 Mar 2009

Posted 16 February 2013 - 10:07 AM

I know it's not easy comparing two cameras with different sensor and pixel dimensions, this is why I tried to use the same sampling. From what I understand, ASI120MM should have had an advantage because at the same sampling it' s more sensitive in green and blue and the image is brighter because of the smaller focal ratio. I was especially interested for Saturn because with the DMK I struggle to get decent frames.
And to make things clear, I'm not trying to prove a camera is better than the other, I'm just trying to understand why I didn't have enough light for higher fps, if I'm doing something wrong and if so how to correct it.

#14 yock1960

yock1960

    Gemini

  • *****
  • Posts: 3104
  • Joined: 22 Jun 2008
  • Loc: (Mostly Cloudy) Ohio, USA

Posted 16 February 2013 - 10:09 AM

I should probably stay in the shallow end...but...

I don't have the experience or the expertise to comment on the pros & cons of the two cameras and I own neither (I have the MC, not MM). But the artifacts appear to be localized, which makes me think that they could be due to some stacking issue or at least some part of the post-processing of the data. In my relatively short time doing planetary imaging/processing, I have seen some...odd (to me) occurrences; some that only happened when the original video was cropped on a different machine from which they were originally captured.

I don't know if that is the case here, but it seems to me that if the issue was the camera, ie. rolling shutter, then the artifacts would cover the entire image.

There is a veritable cornucopia of things that can happen when processing the data!

Steve

#15 RedLionNJ

RedLionNJ

    Viking 1

  • -----
  • Posts: 893
  • Joined: 29 Dec 2009
  • Loc: Red Lion, NJ, USA

Posted 16 February 2013 - 11:09 AM

The frame right definitely sounds "off" for the ASI120, not to mention the 60% histogram. I usually aim for 80-90% histogram with all my cams (DMKAu618, ASI120MC, FireFlyMV, ToUCam Pro). If the seeing or transparency is too poor to support that level of histogram, I figure I shouldn't be trying hi-res imaging on that particular night.

That being said, I can comfortably get the frame rate up to 120fps with the ASI120MC (I don't have the MM) with a slighter larger ROI than the one quoted above. This is with FireCapture 2.2b37 as well.

I do have instability issues which seem to follow me from computer to computer - trying to work through them with Sam right now. With the DMK, my system is rock solid. With the ASI, it's as if somehow the initialization can get confused, leading to a blue screen (on two different computers, going to try a third this weekend). I also want to try the new build of FireCapture (39, I believe?) in conjunction with Sam's 0204 ASI driver...

Grant






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics