DeepSkyImager FAQ
From SkyInsight
|
Guiding, Tracking, and Guiding while imaging
From: Chuck Reese (chuckr54)
Date: Oct 6, 2005
Subject: Re: Help Track and Image at same time
Perhaps the most confusing set of terms & features with the DSI :-)
- "Can Guide" means the system will guide using the GPS command set.
- "Can Track" means the system will guide using the Classic command set.
"Guide and Image with the same camera" does work. However, guide commands are issued *between exposures*, not *during exposures*. This is useful for keeping an object centered on the chip over an imaging session if our sidereal tracking is a bit off, or our alignment is less than perfect. It will *not* correct PE during the exposure, which is what we usually think of when we use the term "autoguiding".
The StarMate (LX200 Emulator) is reported to work just fine with all the above features using the Classic command set ("Can Track").
Guiding the LX200 Classic and GPS
From: Chuck Reese(chuckr54)
Date: Tue Jul 19, 2005
Subject: Re: Autoguiding with DSI, question
Hi, Guys - YES, it will guide the Classic.
The DSI will guide both scopes - it guides the Classic with the Classic command set, and it guides the GPS with the enhanced command set.
It guides both scopes with the appropriate command set for the scope. In the case of the GPS scopes, it calls it "autoguiding". In the case of the Classic scopes, it calls it "autotracking".
The difference is the GPS guide commands are a move/duration command, while the Classic uses a move/stop command.
Those who have used both say they can't tell a difference. I do not have direct experience, but I do believe them.
HTH,
ChuckR
From: Matt Taylor
Date: Tue Jul 19, 2005 1:13 pm
Subject: Autoguiding with DSI
No doubt about it, the DSI/DSI Pro will guide the Classic as well as the LXD650/750 which also use the classic command set. I use it to autoguide the LXD650 & it works fine.
MattT
GuideDog and the DSI
From: Steve Barkes
Date: 7/19/05
It's currently in development. I hope to have a beta for test in a couple of weeks.
K3CCDTools and the DSI
From: Peter Katreniak
Date: Sat Jul 16, 2005
Hi,
I can confirm that I received the needed code for DSI from Meade's senior engineer.
I am already working on Meade DSI implementation in K3CCDTools. It's quite a lot of work, because I need to implement 16-bit image processing to the realtime capture part of K3CCDTools. Furthermore I found some problems with the current Meade DSI interface library:
- currently it is not possible to cancel the current exposure (i.e., e.g. when you set 120s exposure inadvertently, then you must wait 120s until the next change)
- it has problems to detect whether camera is used aready - this will cause interface crashing when you try to connect to camera which is already connected (e.g. by other software)
- the interface library doesn't provide the exactly same RAW image as it is captured by Meade's software - it's a bit shifted and furthermore it contains one black line.
I wrote a detailed report to the Meade's senior engineer at the end of May, but I haven't received a reply so far (he wrote me, that he would have a busy start of June).
Who is author of the interface and who can I contact to discuss these technical problems?
Best regards,
Peter Katreniak
http://www.pk3.org/Astro/
http://www.pk3.org/K3CCDTools/
Meade Telescopes
Celestron Telescopes
The DSI can be used with all telescopes, including Celestron telescopes, for image capture and processing. The DSI cannot control non-Meade telescopes.
Guiding a Celestron telescope
- Star-Mate (third party LX200 emulator) http://www.star-mate.com/lx200.htm
- You will be able to autoguide & autotrack with star-mate & DSI using the Envisage (autostar DSI) software. However the Goto functions of the autostar suite are not supported. You need a Meade mount to do that.
Problems
No Imagers error message
I had the same problem when I bought my new laptop with Windows XP. I got exactly the same message "No Imagers". I called Meade Tech Support. They seemed to know just what the problem was and gave me these instructions which worked.
- Uninstall Autostar Suite.
- Go to www.microsoft.com
- Download ALL windows updates. (This may take some time so be patient.)
- Go to www.meade.com
- Install Autostar Sute update
- Install your disk and go though the Autostar Suite install
Horizontal Lines
- Make sure you are using USB v2.
DSI won't allow Mono checkbox to be unchecked
Problem: The DSI-Color Camera (not the Pro) has the Mono check box checked and it is grayed out so the user is unable to set the camera back to color:
Reason: The DSI-C has an EPROM inside of it that stores information about the capabilities of the CCD chip. Due to some issue such as a power surge or some other reason, the EPROM has been cleared and no longer contains the data. Due to this, the software does not know this is a color camera and as such, forces it into a Mono Mode.
Solution: Perform the following steps exactly as they are detailed here. Once complete, you should be able to check and uncheck the Mono check box once again: Steve Hamilton's fix
Binning
[Kurt Massey] Binning is when you combine groups of pixles on the chip. It effectivly decreases your resolution but increases you sensitivity. If the DSI on you scope Gave you a 1 arc sec/pixel. Binning 2x2 would take two vertical and two horizontal pixles and combine them, you'd now effectivly have a 2 arc sec/pixle ratio.
LRGB Components
[Kurt Massey] The easyest way to understand LRGB is to look at it this way. You have a light box, on which you put a B/W positive image, This is the Luminous channel. The RGB would be pieces of colored film added on top ( like when they colorize an old B/w movie). The L determins the Shade of the pixle, RGB determins only the color. By doing this you can use much shorter RGB times and over weight them ( 1.6,2,2 instead of .8,1,1), they just determin color. The Luminous channel now carries the actual image.
[Gabriel Marcelo] To complement what Kurt said, it is not necessary to have a "complete" color information when you are taking LRGB. The most important aspect is to obtain the best Luminance information on your LRGB photos as possible. For the additional color information you can take your pictures binned (2x2 or 3x3). With this, you use all the ccd to take the color data only excluding some "shape" information. It is normal, take Lumi at full resolution and RGB 2x2 binned for example.
Choosing between the DSI (Color) and DSI Pro
Q: I'm a new guy here ...Just a few questions please as I am going to purchase either the Meade Deep Sky Imager or the newer Deep Sky Imager Pro. Is it worth the extra money to buy the new Pro Camera ??
A [Steve Bell]: You say that you are a 'new guy'...new to the DSI, or new to imaging in general?
If you are new to imaging, I would suggest the DSI-Color, because you will get good color images from the get-go. If you have some imaging experience, I would recommend the Pro with the color filters. Better resolution, better sensitivity, and more flexibility. You can't beat the price on the color filter set, either, even if there are some issues.
Your choice also depends on the telescope you have. If you have a newer Meade scope, you will be able to use all of the functionallity of either camera and software. If you are able to track accurately with your mount, Meade or not, either one will still work well, with some perseverence. If your scope is more archaic (like mine, no GOTO) you will have to work a lot harder...and all of the software functions will not be available to you, at least not easily.
- Beginner, DSI-C
- More advanced, or up for a challenge...Pro w/color filters.
A [Nick Howes]: The C uses a Bayer matrix on the CCD, which gives it the one shot colour capability, but does drop the theoretical resolution etc. Also the sensitivity on the DSI-PRO is over 2x better, so great for luminance shots (see some of the amazing images in the Sky Insight Image Gallary.
The one issue with the Pro I can see is that for a beginner it is a mono single shot cam, so to get nice colour images, you must also get the filter set, and be patient with your image stacking (more so than the colour DSI), and imaging times need to be properly timed.
USB 2.0 vs 1.1
While the DSI can be run using USB 1.1, you'll more than likely experience a variety of problems. The DSI and DSI Pro should be used with USB 2.0. Also, if you connect a USB 1.1 device to a USB 2.0 hub, all of the devices on the hub will run at the slower USB 1.1 speed.
Determining if your PC is USB v1.1 or v2.0
Windows® 2000/XP
- Right-click the My Computer icon and select Manage.
- Note: Windows XP users may need to click Start to access the My Computer icon.
- Click the Device Manager icon.
- Click the plus (+) sign next to Universal Serial Bus Controllers.
Note: If Universal Serial Bus Controllers is not listed, there is a problem with the USB controller or the USB controller is not enabled in the computer's BIOS. Please contact the computer or card manufacturer for updated drivers and/or for information on how to fix or enable the USB controller.
- If Enhanced Host Controller and/or USB 2.0 Root Hub Device is listed, then the computer has USB 2.0 ports.
- If Universal Host Controller' and/or Open Host Controller is listed, then the computer has USB 1.1 ports.
Windows98/Me
- Right-click the My Computer icon and select Properties.
- Click the Device Manager tab.
- Click the plus (+) sign next to Universal Serial Bus Controllers.
Note: If Universal Serial Bus Controllers is not listed, there is a problem with the USB controller or the USB controller is not enabled in the computer's BIOS. Please contact the computer or card manufacturer for updated drivers and/or for information on how to fix or enable the USB controller.
- If Enhanced Host Controller and/or USB 2.0 Root Hub Device is listed, then the computer has USB 2.0 ports.
- If Universal Host Controller and/or Open Host Controller is listed, then the computer has USB 1.1 ports.
USB Cable Maximum Length
The maximum length of a USB cable is 16 feet or about 3 meters. While it may be possible to go beyond that length, the signal will degrade. This length limit can be overcome by using a repeater, most commonly known as a USB Hub. The most important thing is to get a USB hub with a power connection, otherwise it too will suffer signal degradation. Adding a USB hub allows you to extend the cable length another 16 feet (15 meters). USB hubs come with various number of ports and the number you'll need depends on what other devices you need to connect at the same time.
Luminance
Q: Can someone explain, in more detail, exactly what you are explaining about luminance?
A [Matt Taylor]: I'll give it a shot.
What they are talking about is imaging in LRGB, Luminance, Red, Green, & Blue. This is a method of building a CCD image that replaces the interior brightness and vibrance of the RGB image that is lost due to many factors. In addition to this it also helps to bring forward the fainter details of the imaged target.
In this case the discussion is about using the color DSI to shoot both the RGB color photo, then clicking the MONO button and shooting a black and white luminance photo. Later during post processing the luminance photo is layered on top of the RGB photo is such as way as to accomplish the goals stated above. Once layered the RGB photo recieves a slight blur so that it provides only the color and hardly any detail to the resulting photo, as details are contained in the luminance layer.
This method of imaging was developed specifically to address the lower resolution of full color CCD images as compared to their B&W higher resolution versions. For example, the color DSI uses 4 pixels to make 1 color site in the photo, the DSI Pro uses only 1, so the difference in resolution (visual) can be said to be 4 to 1 in favor of the Pro. Using this method with the color DSI narrows that difference in theory. The preferred method would be to shoot the color with the color camera and the luminance with the DSI Pro, this provides a simi quantum leap in detail and resolution, and also cuts the imaging time down because the RGB comes all at once rather than through RGB filters. Using this LRGB method also has a hand in increasing the photo's overall signal to noise ratio, but it comes at a price, I'll explain.
Using the LRGB method, be it in this manner with a color camera, or the other I explained above, reduces the vibrantcy and color saturation in an image. You can see this best while building an LRGB image in the stars within the photo. Once the luminance is layered the apparent color saturation of stars seems to fade and they are left mostly white. Color saturation of the target seems to fade as well. There are various hacks some try to bring that saturation back, Photoshop's saturation tool comes to mind, but these desperate measures fail to provide the vibrance and saturation that one seeks, save one.
- If the RGB ratios are computed for a given pixel, this ratio will be independant of the brightness of that pixel.
- Then the luminance exposure value for that pixel can be multiplied by the individual R G & B ratios, thereby scaling the RGB ratio by the brightness of the luminance pixel.
And really thats all there is too it. In order to achieve a proper color saturation and take advantage of the luminance layers high quality and detail the RGB ratio values per pixel & per color channel must be scaled by the luminance value for that same pixel, which comes from the luminance layer itself. Considering that the DSI has 250,000 pixels this could take a very very long time, lol!! But there is a simple method using a common software like photoshop.
- Duplicate the RGB image
- Convert the duplicate RGB image to grayscale
- Invert the grayscale copy
- Copy & paste this inverted grayscale image into the RGB image from which it came.
- Set the combine method for this layer to "Color Dodge"
- Now paste the luminance image onto this stack
- Set the luminance layer's combine method to "Multiply"
- Flatten the layers, your all done.
This method of building the LRGB photo will scale the original RGB photo's vibrance and color saturation (RGB to brightness ratio, per pixel, per channel) and will lead to a better overall photo visually, assuming one starts with good data to begin with.
Act II, Scene 2......
It's important to understand the reasoning behind the LRGB method. The mindset of users of this method is that color is simply color, it holds little detail other than overall major features of the imaged target. For this reason many who use this method give the color layer of the photo a blur, so we acually destroy any minor or intricate detail that's held by the color layer. All details, intricate or otherwise are therefore held within the high quality luminance layer, with this in mind it is easy to understand why it's best to shoot the luminance layer with a camera that provides greater resolution than the camera that shot the color, unless of course the camera that's taking the color is also taking the luminance which is the case if using only the DSI Pro for this method. Still, the RGB layers, once built into the RGB layer, is blurred and all detail comes from the luminance layer.
If one has both cameras, the DSI and DSI Pro, it's a no brainer to take the RGB color photo with the color DSI then take the high quality luminance with the DSI Pro. It saves loads of imaging time not to mention being less of a post capture calibration monster when compared to doing it all with the DSI Pro, and really, since I'm going to blur the color anyway, why use the Pro to take the color in the first place, right? Right.
So, there you have it. We've discussed what it is, how to do it, the only problem with it, and how to cure that problem. Now, go forth and multiply by the inverted RGB grayscale layer, lol.
Best Regards,
Matt
Imaging Planets
[Chuck Reese] Everything else being equal, the ToUcam will beat it every time. The DSI was not designed for planetary imaging (the LPI was), and it's greater sensitivity actually works against it.
Things you can do it maximize your success with the DSI on mars:
- Critical focus is critical to success - use a diffraction spike mask & focus on a bright nearby star, then move to Mars.
- Collimation matters (not the issue on your fine refractor, but SCT & newt owners should take the time to tweak collimation). Collimate with the same optical train you use to image (i.e., with any barlows, diagonals, etc. in place)
- Manage the histogram - set the white point slider most (probably not quite all) of the way to the right, then tweak the exposure time to eliminate clipping (burnout) and maximize detail. You may have to play with the exposure & white point settings a little to find the optimum tradeoff. You want to try to fill the histogram without killing the contrast.
- Use Normal mode, and at least bmp or tif formats, fits is better (but "some assembly required" for fits afterwards).
- Set the minimum quality % at least to 50, 70% is better if the seeing will support it. I usually set the eval count to 10...not sure what effect lower or higher settings will have. Once you cross the eval threshold, if the image is not off to a good start, kill the session & start again - you need a good base eval image for a good final result.
- Stack as many sub-frames as you can - let the capture session run at least 3 minutes...you can safely go to 4 or even 5 minutes with Mars to generate a stack of at least 100 frames (300 would be a good goal).
- Tweak the kernel settings - this is a user preference, I usually like the "Edge enhance medium" setting.
- Watch the preview image for periods of better seeing, try to time the capture sessions to take advantage of steadier viewing.
- Afterwards, post processing with wavelets (Registax, AIP4WIN, others) is good, careful use of unsharp mask & high-pass filtering is also effective (try alternating between them - a moderate application of USM, then a moderate applicaiton of high pass, etc.).
I have seen some *very good* images of mars done with the DSI. Not quite as good as with the webcams, but certainly rewarding.
Actually, I left out one suggestion - try using a lower gain setting. Dropping back even to 90 can make a big difference. I have seen reports from some that using a gain of 0 works well for them.
Good luck - hope your skies clear soon! ChuckR
Drizzle Software
Q: Will the Deep Sky Imager (older one) work with the Drizzle software at all?
A [Nick Howes]: DSI-C (the older one) works with Drizzle as Meade updated the software to support both models.
Q: In your opinion ...if the Drizzle will work with the old or not (only with Pro cam.) is there a big difference in the quality of the images from the older Deep Sky Imager to the newer Deep Sky Imager Pro level???
A [Steve Bell]: As for the Drizzle feature, I am becoming a raving fanatic! I use it in post-processing, and it has transformed my 0.6 megapixel camera (the Pro) into a 1.4 (potentially 2.4) megapixel monster!
FITS vs FITS3P
Q: What is the difference between FITS and FITS3P?
A [Kurt Massey]: FITS is an acronym for Flexible Image Transport System, the only difference between FITS and FITS3P is that in FITS each color is saved as a separate file, FITS3P is the same 3 FITS files, just saved in one file to simplify processing. MaxDL and the Autostar Envisage will process either.
Dark Streaks
Problem:
40 minutes luminance
lx200 10 inch at f/3
I had a lot of problems with dark streaks from dark removal and stacking as the camera ran warm last night because of outside temps.
Solution [Matt Taylor]: Before processing the file open it in AS-IP and apply the "Fix Cold Pixels" command with a setting of 1.0, then right above that filter you'll find the "Median" filter, run that as well. Then process the file and I think you'll find a better starting point.
Post Processing RGB
By Chuck Reese
These are the steps you use to combine and align in Meades Autostar Suite. Assuming you use fits (not 3P) & save all uncombined images:
- Group -> New -> select all your "R" images
- Group -> Examine -> page through the files & delete any bad ones
- Group -> Align -> "one star" if you are polar aligned, "two star" if you are alt-az
- Group -> Combine -> Usually average or median
- File -> Save fits image -> name it something_master_R.fits
- Repeat 1-5 for G
- Repeat 1-5 for B
- Group -> New -> select your R, G, & B master files (3 files only)
- Group -> Align -> "two star" unless polar aligned
- Color -> RGB Merge -> select each of the master R, G, & B files using the browse buttons (Tip: using the updated AS software, you will be presented with a histogram to prescale each file. I have found the default setting works well for me so far)
- Color -> RGB Merge -> Preview -> Study the color alignment in a brighter star - if you see color fringing, use the RGB X/Y tools to move the offending plane(s) into alignment. When you are happy, click "OK". The files are now merged, and you have a full color image.
- File -> Save fits image - call it something_master_RGB.fits You can now use the various image processing tools to adjust the histogram, color levels, sharpness, etc. (You might want to work on a copy of your master file instead of the original)
- (Optional) If you want to further adjust the image using another software package, you will likely need to save it in a different format - File -> Export Display As -> choose appropriate file type for your software.
Manual Star Bloat reducing in PS
By Neil Fleming
Date: Sun Jun 19, 2005 5:12 pm
Star Bloat routine - I do it in Photoshop:
- Make sure you've selected the image layer, rather than one of your adjustment layers
- Use the eyedropper tool to select the color one of the stars that you wish to work on
- Use "Select/Color Range" to select all similar stars in the picture. You can set the degree over which the color range is selected
- Use "Select/Modify/Expand" to increase the selection by 2 pixels
- I then use "Select/Feather" to "blur" the area of the selection by 1 pixel
- Go under "Filter/Other/Minimum", set the range as low as it can go - 1 pixel, and apply this filter to the selected stars
- Generally, the effect is to extreme, so IMMEDIATELY go under "Edit/Fade Minimum Filter" to adjust the amount of this last edit to a more appropriate level. The example above was probably done at 50% of the original "minimum" filter amount.
- Hit CTL+D to deselect the stars, and you can use the History window to view the various states of your work.
And there you go!
Layering the Trapezium
By Matt Taylor
OK, we're gonna cheat nature. Here's how.
- Combine all your M42 images & stretch all the faint stuff to where you want it. This will of course blow out the trap, but all the faint stuff will be nice & pretty.
- Combine about 1/2 your images into another photo, the goal for this one is resolving the trap as best it can be. You can even do a 3rd photo that resolves just the trap stars using about 1/8 of the exposures. Now remember, we're cheating, so this might not be as good as if you had short all of these in one night at different settings and as different photos then combined them later, but this will work.
Once you have your 2 (or 3) stacked photos you can follow the tutorial here:
http://www.autostarsuite.net/blogs/tutorials/archive/2005/02/10/205.aspx
but below is a brief(er) version in case you have been through it already and just need a reminder.
- In PhotoShop, open the bright photo and make a new layer, open the trap photo and copy it into the layer
- Make your mask in the trap photo, copy the bright photo into it (it will be grayscale)
- Click VIEW -> NEW VIEW
- now blur the grayscale mask to taste (watch the NEW VIEW photo, it will change as you blur)
- Now stretch the grayscale to taste.
When the NEW VIEW looks good to you, close it, then click LAYERS->FLATTEN & you all done. You should have a good photo of M42 with a resolved trap.
Now if you want to go further just use this current photo as the bright and your 3rd stack (resolved stars) as your faint & do it all over again. No, this won't be a perfect end result, but I bet you can get a really nice photo out of it!
Can't wait to see it!!
Matt Taylor
