QHY8 - Still cant get it.

[toryglen-boy]

ok, still i am trying with the QHY8 and Maxim DL, still i cant get it. Here is a screenshot of what i get,the histogram seems like a complete puzzle, any comments welcome, before i bounce this thing of the wall

[sheeny]

hmmm... I don't have one, but I don't see anything too unusual about your histogram for what I'd expect with a 16 bit camera:shrug:... Maybe clipping both ends a bit, but that's just an adjustment.

Can you be more specific about the problem?

Al.

[toryglen-boy]

Yeah, i dont understand it, and cant take a decent image

:)

[jase]

Yes. Takes a little getting use to Maxim's screen stretch. Quite powerful and useful.

In the minimum field enter 0 (zero) and hit enter. This will place the red triangle (black point) as far left as possible. Then hit the + magnifying glass to zoom into the histogram more. As you zoom, you'll be able to raise the red triangle and lower the green triangle (white point) to display the data you want to view.

I'm not sure what you've taken an image of, but the signal is weak. Imaging is a numbers game. Despite what many believe, FITS files aren't conventional image binary files, but series of numerical values. This provides incredible flexibility allowing the imager to multiple/add/subtract values to all pixels to fill the bit space (16bit or similar).

If I was to judge your data, I'd say you need more exposure time. 3 minutes isn't going to cut it. The info window shows that the minimum pixel is 1063 counts with the average 1433 counts. This will change where ever your cursor points, but for an average to be so low is probably a good indication more exposure time is needed. The maximum of 65k will be hot pixels or saturated stars. The info window are real values i.e. you can screen stretch the data as soft or as hard as you please, but the pixel values remain the same. Screen stretch does not manipulate the image data itself, just how its presented on the screen buffer.

[toryglen-boy]

thanks for that Jase, its NGC 300, here is a single 45 mins shot (ignore the guiding going mental)

i think i might fire up the 1000d, it was sooooooo much better than this

[jase]

The data is all there.

Just need to play with the sliders. Please keep in mind that the screen stretch function is performing a linear stretch on the screen buffer data. This is the reason why you're highlights look saturated and your shadows black clipped. Obviously you need to adjust the red and green triangles to reduce the clipping. The point however is that the data is all there. linear stretching does not show the mid-tones well hence the contrasty look.

Now, take a couple more sub frames and combined them and save it as a 16bit tiff. Use photoshop or gimp to do a non-linear stretch. Or you can use MaximDL's DDP and log stretch functions to fill out the bit space.

[AlexN]

How have you set the gain and offset values for your QHY8 Duncan? This can make a big difference to the final image with a QHY8.. when I had mine, the default values in Nebulosity 2 were Gain 50, Offset 124. after working out what settings I needed by taking bias frames and adjusting the values so that I had an average of between 800~1000 ADUs in a bias frame, the settings were Gain 0, Offset 64. This always provided me with a very nice image..

As Jase said, the screen stretch is not manipulating the data in any way, just changing how its displayed on the screen, it is also not indicative of what the final stacked, processed image will be...

Are you shooting RAW or Colour? I would recommend shooting RAW..

[TheDecepticon]

Hi Duncan! Have you tried the image capture program that comes with the camera, CCDCap, just for s**ts and giggles? Just to see what you can get. MaxIm doesn't present your image as it actually looks. Can you check the image in Photoshop as a fits using fits liberator, just to see what it looks like in there?:shrug: I really feel for you:sadeyes:, it is a great camera and you should be getting some cool pics. What about just trying some of the easy, boring regular objects to get to know how the camera works? Also, some aren't gonna like this, however, I would go over to Nebulosity, it is a lot easier. If you want to, PM me, and I'll give you my home number and see if I can help a bit while your in front of your PC.:question:

[toryglen-boy]

Numbers came about through trial and error, i am shooting in RAW, although whe i debayer it, its colour ?

:question:

[dpastern]

*carefully taking notes of all of this*. (and trying to learn).

Dave

[Tandum]

Duncan,
Here's a guide to setting gain and offset - http://www.lightstorm.at/pdfs/QHY8_settings.pdf You will need a different offset for each gain setting you want to use. Start with a gain of 0% to get going.

I use the maxim sequence section to capture lights, bias, flats and darks. I use the focus section, set to continuous, to frame the shot. With binning set to 4x4, a 5 sec exposure and screen stretch set to high, I can see the target on screen to frame it.

I use Deep Sky to stack all these files, maxim often won't stack qhy8 images correctly for me. In DSS, under Raw/Fits DDP settings, go to FITS tab, tick monochrome and select QHY8 from the camera list. The first attached image is what I see in PS after stacking in DSS. The second is the same image after curves and levels. I sometimes use ddp in maxim to stretch before going to PS but for me it's not always successful with qhy8 images so normally I'll stetch the image in PS.

Don't give up, they are pretty good cameras for the cash.

I've added a single 10minute sub, shown in maxim and shot from a dark site, from that image set.

[Steve]

It looks like you might have some fogging or ice on the camera window, that can cause the gradient you see in your images. I'd check it with a flashlight while you're in the field to make sure it's clear. There's a few ways to fix it if you are getting it.
Steve

[multiweb]

:lol: Duncan you're not QHY8 compatible. Pack it up and take it back to the shop. :lol: ;) :thumbsup:

[toryglen-boy]

actually, and dont hold your breath, its starting to make sense!

its my own fault, i thought it would be much easier than it was, and i should have done my own research, instead of posting absolute drivel on here ..


:lol:

[multiweb]

I'd suggest to put the camera aside for a little while and do a lot of reading. You'll save yourself a lot of time and confusion. There are absolutely stacks of info on processing or data acquisition with the QHY8 in many many forums. Have a look at the QHY8 forums, CCDware forums, Yahoo groups: stark-labs, QHYCCD. That's a good place to start from. :thumbsup:

[gbeal]

Whatever you do, don't give up, keep asking. The "get the DSLR out" idea has merit, as it will show you how easy it is to get an average shot, but the QHY8 will give you better.
First thing you really need though is the "proper" offset and gain, so work on this. I did the same as what was described in the "lightstorm" link Robin posted. In my case I ended up with Gain 1, Offset 114, but each camera is different, so trial and error it, and get this sorted first.
Then tuck into some more subs, longer being better, or if the skyglow intervenes, more of them.
Gary

[AlexN]

As others have said... Expose longer... If the skyglow is a problem, either get a Hutech IDAS LPS filter, or take plenty of shorter subs... I would say 5min is the minimum exposure time for any target out there (apart from M42 which always requires some short subs to tone it down a bit) when I'm at dark skies, I don't bother with anything less than a 20 minute exposure, unless im capturing colour information for a Ha lum or something like that. with an LPS, 10min subs should be doable from suburban skies... I would strongly recommend doing this over taking heaps of 3 min subs...

As Robin and Gary said, the gain and offset can be different for every camera, Gary's was Gain 1 Offset 114.. My QHY8 was Gain 1 Offset 66... So follow that guide that Robin linked...

Also, with a 16bit CCD, you can expect to have to do quite a few itterations of levels and curves to pull out data... Unless you're taking very long subs, the data will not be clearly visible in an unprocessed image like it is with a DSLR.... It will only be clearly visible in 15~30 minute subs without applying some stretching...

I have an unprocessed 30x20min shot of M8 and M20 on the HDD, I can upload a JPG copy of how it looks pre-processing for you to show you what I mean... even with the epic exposure times, its still not much to look at...

[Gama]

The Idas LPS filter does OK for light pollution, but you will lose a heap of sensitivity on galaxies and the like.
Here an image from the QHY-8 with my CDK-20, its a stack of 20 x 5 minute subs, with NO darks, No Flats etc, so you can get the feeling of what the image is like thru the IDAS LPS filter. I have also attached an image without the LPS, and notice the light gain.


Theo

[allan gould]

The gain and offset for my QHY8 was 1% and 65, respectively.
Its different for each camera.
Keep in there Duncan - you will finally twig to it all and then you will be off. I felt the same in a way with my flats until Marc put me on the right course.
We'll help, so don't give up.

[dugnsuz]

http://www.centralds.net/en/index.htm:whistle::whistle: (http://www.centralds.net/en/index.htm)

[Tandum]

Wow .. 5Kg cameras ...

[dugnsuz]

Astro 40D = 1020g:shrug:
Am I missing a CCD v's DSLR in-joke?

[Tandum]

The fan on the side of that thing is the size of a qhy8 doug.

And how can you compare the 2. The qhy8 is 16 bit and vitually noise free ... You can stretch the images to tomorrow and back.

[AlexN]

Both cameras have their merits... The cooled DSLR's offer fantastic results and the ease of use that keeps most people imaging with a DSLR. As robin mentioned, the CCD's are 16bit.. thats just for starters... cooling in a CCD is far more serious than the cooling in the cooled DSLRs.. Couple that with the price factor... why buy a DSLR that has been heavily modified, and had its warranty voided so that its capable of reasonable quality astro images, when for a lower price you can buy a dedicated astro imaging camera that is designed and built with astro imaging in mind... It amazes me that Central DS sell cameras to be honest.. Especially due to the fact that now you can get an 8.3mp SBIG colour or monochrome ccd for about $400 USD less, the QHY8 is cheaper, the Orion Starshoot Pro is cheaper... You'd be flippin mad to pay 2K USD for a cooled DSLR in my opinion... The only reason you would ever consider it is if all your imaging (or most of your imaging) was done through Canon EF series lenses. Then you would simply for convenience... I have a EOS Lens adapter here to connect EF lenses to my SBIG... it takes more time to set it up, but again, 8.3mp mono ccd, much more sensitive, lighter, designed for the task and cooled much more efficiently with much less noise, at less cost..

Perhaps we are the ones missing the CCD v DSLR joke?

[Octane]

DSLRs produce results. So do CCDs.

It's all up to how much time the operator wishes to invest in capturing the light, calibrating their images correctly and processing properly.

Note: I am not inferring that DSLRs are equivalent to dedicated CCDs.

[toryglen-boy]

anyway guys, its finally coming together, and i am getting there.


:)

[marki]

Thats the way Duncan, keep at it. Once you knock out a few images it will all make sense :thumbsup:.

Mark

[AlexN]

As they say, Practise makes Perfect.. Resist the urge to fall back to your DSLR, stick it out with the QHY8 and the results will improve! Good to hear you're making some headway.

[dugnsuz]

Central DS website quotes camera+cooler weight as 1020g. Not too sure how accurate that is - cooler unit must only weigh around 300-400g for that figure to be accurate:shrug:
But Robin I have to concede that with the high ambient imaging temperatures I experienced last week, an ultra-cooled CCD does seem very appealing. My SMC had lots of blue and red smaller stars - not pretty.

[dugnsuz]

Must admit Alex, I'm slowly coming round to your way of thinking.
The 40D was originally purchased with an equal mix of terrestrial and astro work in mind. In reality, the DSLR has stayed fixed to the mount pretty much permanently! So this fact in conjunction with the market shift toward more affordable high quality CCDs becoming available presents an (almost predictable in astrophotography!) dilema.

[Octane]

Doug,

I see no problem in saving your money and mastering the DSLR. You are getting fabulous results. And, besides, with every image you capture, you hone in on your processing skills. Capturing the data is only 50% of the problem. It'd still be the same whether you continue to use a DSLR or a CCD camera.

I, too, am in two minds about jumping ship, but, then think what I might be able to do with the 5D Mark II...

Regards,
Humayun

[leinad]

Becoming a new owner of a QHY8 myself, when its mentioned the maximum pixel value should be between 400-1000, is this best to be closer to 400 as can be achieved ?

I followed the instructions and was able to achieve Gain 2 Offset 109 of what I thought was reasonable.
See screenshot...

I noticed I have a few dead pixels? White dots seen when the image is dark.

[multiweb]

Get your gain to 0 or 1. The offset will determine the position of your histogram in respect to the left side. Varies on cameras and sky conditions. I use 60 to 70, that gives me a sky background of 900-1000ADU typically. You want that margin on the left, no more no less. 1000ADU is comfortable to play with. No clipping.

[leinad]

Ok thanks, I'll try again tonight.

[allan gould]

Managed to purge the TVDE QHY8 drivers from my system/laptop and load the straight drivers from QHY. The reason that I did this was in Nebulosity you could only get 1x1 or 2x2 binning. I wanted to use the 4x4 binning mode but the TVDE drivers caused Nebulosity to crash. Now have all 3 binning modes for the QHY8 and am a happy chappy. Really loving this camera

[Gama]

You will always get hot pixels in any CCD, this is normal, no need to pack up and run to the hills:P. The Sony sensors have very low hot pixel counts..


Theo

[leinad]

:reindeer:

[Manav]

Where? tell me tell me :D

[leinad]

A quick question of gain/offset

In Nebulosity I found that
Gain 0 Offset 50 I get a Max value just on 1000. I believe thats right from what Craig Stark mentioned.

At bias 0.001 secs, this puts the histogram right to the left. Pity you cant zoom to view the histogram like MaximDL does.

http://i611.photobucket.com/albums/tt197/Leinads_pics/qhy8_orion_test.jpg

Above was a quick stack (flat/bias calibrated and stretched) with a couple of 2min, 5min (Gain 0 Offset 55) and 1 10min (Gain 5 Offset 55).

I seemed to get similar results as well at Gain 1 Offset 110 compared against the Gain 5 Offset 55, though this wasnt good with the bias test shots.

If I understand this all correctly, I should find the right Offset with Bias frames then use that, and just boost the Gain with lights to get the best results with exp time.

Feedback welcome, thanks.

[Astrod00d]

That's a great Orion, especially for a 'quick stack'!! :eyepop:
Interesting that you mentioned the lack of histogram zoom in Neb compared to Maxim. I like and use Neb, however I wish it would allow me to zoom the histogram. Perhaps a quick email to Craig...

Cheers,

Rob

[leinad]

While you're at it. A stopwatch feature would be handy too. :D
But yeah if he could do something similar to what Maxim does that would be pretty cool. Unless they have the rights to that kinda feature?

[Hagar]

The idea and method Craig used was for the MINIMUM value to be around 1000 and not the maximum. This gives a bit of a buffer where at no point the black point will go to zero and thus loose data.
I always ran my settings in Nebulosity2 at a gain of 1 or 2 and an offset of 65. Black point clipping is the main concern as data that is clipped is lost forever. The black point can be adjusted during processing and some processing tasks such as contrast enhancement and shadows and highlights requires a bit of space in the black end.

Hope this helps a bit.

[pmrid]

Hi Allan, please tell me how you managed to get rid of the Tom Van den Eade drivers please. They seem very resistent to myt attempts - sdame reason as yours.
Peter

[mithrandir]

It's a while since I switched, but as I recall it, download the uninstaller from Tom's page http://www.astrosoft.be/ (near the top of the page).

Download the latest QHY8 drivers http://qhyccd.com/files/QHY8-0906.rar unpack them, plug the camera in, and point the new hardware wizard at the directory.

[allan gould]

Beat me to it, its as given above

[leinad]

Thanks guys. :thumbsup:

[multiweb]

The only reason I have ever used the TVDE QHY8 drivers was the noise wihile downloading the pictures from the camera via USB to the laptop. But that was a long long time ago. Has this been addressed and are the stock QHY8 drivers sorted out now? I'm asking coz I'm considering going back to the main stream drivers now. Neb2 seems to chuck a fit in the very latest versions with the old drivers :shrug:.

[allan gould]

A while ago I purged the tom drivers and added the new qhy8 drivers, but with the latest Nebulosity I had noisy downloads, so I uninstalled again and loaded the drivers as above to get 1x1, 2x2 & 4x4 binning with no noise in the down loads. I just wish all these sites with all the different versions of these damned drivers would purge them all or at least give good and clear directions for which should be loaded and under which circumstances. It confuses people like me.

[allan gould]

Have a question to ask of you fellow QHY8 users. Just recently I had a go at imaging the HorseHead nebula with the qhy8. I was using the UV?IR filter in the camera and taking 1x1, 5min subs. I was disappointed as the image which resulted was not worth the effort. My modded 20D gave far better shots in 5 min than the QHY8. Do I need to change to the clear filter for this type of shot - I wouldn't have thought so looking at the spectrum of the IR filter which comes with the QHY8.

[Tandum]

What was wrong with it? This one is a mix of 5 and 10 minute subs. Apart from the halo it looks ok to me.

[allan gould]

Nothing wrong with your photo but were you using the UV?IR filter and how many subs total to give your image. The attached photo was with a 20d modded at 4x5min subs. With the QHY8 I got nowhere near this.

[Tandum]

The filter is embedded in the nosepiece, so yes.
In the raw data I see 8x300, 6x600 and 6x180seconds but I don't know how many made it into the processing, it's a while ago now :) I think more rather that less subs is always better.

[pmrid]

Allan, I'd be interested to see what your rejects looked like. The deterioration of my own QHY8 first became apparent to me imaging the same subject and, like you, I dumped a night's work because the results were so unusable. Got any of your subs left that we can have a look at?
Peter

[Tandum]

Just for interest, here is a single sub from mine which we then swapped out for peter's faulty camera and a single sub from his. Both 300sec with a simple ddp stretch.

[allan gould]

Here is a stretched demosaiced 300 sec shot from Nebulosity

[Tandum]

That's almost empty? Do you have maxim allan? What do your bias shots look like in maxim? Anything like this? The numbers are important rather than the image.

[leinad]

Whats your gain/offset in that shot ?

[allan gould]

Gain set at 1 and offset set to 65
The bias frames give a mean of 1824; max 1882 and min 1765 - taken at 0.001 secs

[allan gould]

Ill have another fiddle with it as I have recently changed the drivers for this camera???????!!!!! Was going fine until then

[Tandum]

Stick with zero percent gain to start with so the question is what's the offset and it's really camera dependant but when I was checking peters camera, maxim came up with a different offset than he got from nebulosity?

I think my camera was 125 in maxim.

[Gama]

It does look like an offsett. Remember that Nebulosity and Maxim use different figures to represent the offsett and gain.
A gain of 1 and offsett of 65 is NOT the same in maxim. So you cant just enter the same numbers..
Can you post a single sub showing the Histogram ?.


Theo.

[allan gould]

Image capture was in Nebulosity and I used the tutorial from Neb to determine the gain and offset for my camera. Ive attached a sreenshot of a single frame before and after de-bayering in Nebulosity.
Left - not debayered right - debayered

[Gama]

Allan, can you send me a Crop of 1/4 the image in a FIT format to my email.
Want to check a few things.
Make sure its NOT converted to color.

Theo

[CometGuy]

Are sure this isn't condensation on the CCD coverglass?

Terry

[allan gould]

Absolutely sure as I run a heating element in the camera (new for QHY8 from Theo) and previously I always used a 2" dew heater strip around the front of the camera.
I only use Nebulosity to capture with the QHY8 - dont have Maxim

[allan gould]

Theo - file sent to your email. Not colour converted.
Allan

[Gama]

I checked your file Allan, and under Maxim, it shows a clipped image.
I thought at first, maybe its because you cropped the image, and thus has changed the Histogram position, but i also cropped one of my own images, and the Histogram barely moved..
Attached is the image i got using Maxim to display the image..
Very strange.

Theo

[TheDecepticon]

One of the images is very noisy. I know there are some capture problems if the TEC is not working well. Is the camera getting warm at the heat sink and the fan blowing warm air? Have you checked that all components are getting their respective power. Why not check out the QHYCCD Yahoo group? You might get some info from there.:thumbsup:

[allan gould]

Found the problem. Originally I was using TVDEE drivers but changed to QHY drivers to get the 4x4 binning. Now I find with this installation Im getting a lot of read out noise and the offset settings have to be increased to ~120 to get a pixel reading of ~1000. Previously my offset was 50-65 to get the SAME value with TVDEE drivers. I have gone back to the TVDEE drivers as these give absolutely noise free images.
When I upgraded to the QHY drivers I assumed that the offset settings would be the same- not so at those settings for the TVDEE drivers 65 offset, the ADU reading was 0.
Find it interesting that now Im getting download noise with the qhy drivers and not TVDEE drivers????

[Gama]

The offsett numbers are not identical, or should i say, they are not absolute, but rather relative to its own formulars/calulations.
The same goes with the gain, its all relative to its own programming.
I use the original drivers from QHY, and i have no noise, and images are fine..
But im glad your back on the wagon..

Theo.

[allan gould]

Theo and others, thanks for your valiant efforts. Ive tried all the available drivers and settled on QHY drivers which I can now use noise free and with 1% Gain and an Offset of 111. When binning from 1x1 to 2x2 and 4x4 the ADU goes from ~350 to ~750 to ~1200 and so this looks like the settings for my camera. With the TVDEE it was 1%Gain and offset of 50. Guess that its a lesson for all to check it if you change your parameters. Noise may have been due to partially seated s video cable not being all the way in.

Allan

[multiweb]

Wow. Pretty amazing how significantly drivers can change the behaviour of a camera. I reckon I'll stick with what I've got now. Seems to be working so far. Glad it's all sorted out for you now. Back to imaging :)