WBPP : PixInsight : Output Pedestal for Narrowband Imagery

[@setiv2]

Wow! I just checked all my recent narrowband images and the amount of red pixels was staggering! Re calibrated everything with a 200 pedestal and all signal back in my images! Thank you for making these videos and sharing your knowledge!

[@AdamBlock]

I am glad you found the information useful! Become a member of my site... lots more useful stuff. :)

[@christophebrun-franc984]

Great presentation. I finally understand what is pedestal and what it is used for ! Thanks.

[@AdamBlock]

Thanks!

[@TheMje1963]

Kewl, Just learned something new. Since 90% of my data is Narrowband I will most likely be using this quite a bit.

[@AdamBlock]

Glad you found it useful!

[@Brisbane_Astro]

Woah. Such a simple tip. I checked a run that I have just completed and lo and behold, there was the red rash through my Sii channel. Adding the pedestal to my WBPP routine for sure.

[@zelodec]

Just finished a 9 panel mosaic of the Veil Nebula and I used the IMX571c ToupTek color camera with a dual narrowband filter. After calibration I did the PixelMath trick and oh boy the whole screen was red. I'm tempted to redo the mosaic but first I will try just 1 panel and compare the difference with added pedestal. Luckily I just got the camera. Under-rated channel. You deserve much more exposure. Thanks¨!

[@dmintz88]

Great video, Adam! This was super helpful!

[@AdamBlock]

Great! This is one of my better videos... very concise I think.

[@Thommy78]

Great instructions and explanations! Thanks a lot.

[@wandaconde3696]

Excellent video!! I will definitely go back to some of my “old” narrowband images and find out how many zero value pixels I get. Thanks for putting these videos together. I am enjoying and learning a lot from them. Thanks for making them publicly available.

[@AdamBlock]

Please tell your fan club (you have a big following...I know it!) about this KZbin channel. I have only a tiny number of subscribers. :)

[@wandaconde3696]

I don’t know about a fan club LOL LOL, but I will definitely spread out the word. The work and effort you are putting on this videos should be widely recognized. They are awesome!

[@juliobynight2495]

Thank you ! I checked my files and I discovered lots of zeros values.. processing all over now.. but for better results.

[@AdamBlock]

I am glad you found it helpful. Much more of this kind of stuff at AdamBlockStudios.com ! :)

[@yosmith1]

thanks for providing this, I had no idea what that value was for. I have a lot of data that I want to go back and try this on.

[@michaellewis5921]

Adam, What a great presentation about what is a simple concept that is frequently not understood and/or overlooked by many, including myself. Your usual clear and concise way of covering this will have me now scurrying back to some previous data to see to what extent I may have been experiencing this issue. I was overdue for subscribing to your channel as well - that oversight is now remedied :)

[@gclaytony]

Great video and timely. I had just finished a two pane mosaic of the Veil nebula with an Esprit 100ED/ZWO ASI2600MC-P+L'Extreme filter. I performed your single Pixel math test to see what it showed - and found ZERO red pixels. I zoomed to the same area in the Dark and light frame I used and found approx 150 count delta in the light above the same pixel in the dark. Not to be shy about this, I have had issues in the past (and recent past again) with 'bad' darks which caused excessive low and high end 'clipping' during the WBPP process. I hate to take the extra step, but the best quality darks I've made have been when I pulled the camera from the scope and installed the ZWO sensor cap AND covered it with foil. Narrow band monochrome frames can be very dark, so not saying that light leaks are always a contributor to the low clipping you found, but the stretched image of the dark in your video could indicate the possibility (at least based on what my 'bad' darks looked like when I started examining them). I suspect cameras like my ASI2600, with extremely low e-noise, may be more susceptible to light leaks 'contaminating' darks, but that is just a guess on my part. YMMV

[@lancesmith7821]

This absolutely blew my mind. I had been working on an image that I thought looked great, then found this tutorial. Did the trick to highlight the zero pixels and 30% of the image is red! I will be going back and trying again. Thanks Adam!

[@AdamBlock]

I am glad you gave it a try. Just think about all of the other mind-blowing things you will find by becoming a member of AdamBlockStudios.com . :)

[@DSOImager]

Thank you Adam. I have never added a Pedestal before.. after coming across this video I tried it out and the difference is notable. :)

[@mars_man5619]

Adam thank you very much. After applying your formula to my 20 minute Ha (FLI PL16803) data, I was amazed.

[@trevorgreen2232]

Thanks for this Adam , another great presentation

[@z28rgr8]

The Image Calibration process has this capability. Not a user/fan of the WBPP script. Auto pedestal seems to work pretty well. Cheers

[@DrReinerHutwelker]

Dear Adam, you made my day! Wonderful! - I was so confused about the result of my first Narrowband, took new flats, ... - Is there a sweet spot for the value? And how could I find it?

[@AdamBlock]

Guess what the developers of PixInsight did...they implemented an automated routine that will discover and use the appropriate value (in WBPP). It is a feature you can check if you like.

[@DrReinerHutwelker]

@@AdamBlock Thank you so much!

[@andyweeks2216]

Another wow video Adam! Thank you!

[@ACKitsBilltheCAT]

Adam Block Star Reduction

[@Calzune]

Amazing Adam! Will this have a negative effect in I add 100 pedistals on data that has no 0 pixels? My Ha data had tons of red pixels but my OIII only had very very few of them.

[@AdamBlock]

No, not really. Add 100 to a 16-bit image isn't a significant adjustment to the range. (100 out of 65,535 is small). Also, most sensor saturate before this maximum value. So you are just shifting the values of the data with no meaningful change.

[@Calzune]

@@AdamBlock thank you Adam!

[@nihalmirza8864]

Hi Adam, thanks for the videos very helpful, will you be doing a video on post processing ie registration & integration setting and recommended values (unless I missed it). Also is there any benefit in generating drizzle data, if yes how does one use it with WBPP? Or is it a manual operation? Thanks in advance.

[@AdamBlock]

Hi Mirza - all very good questions. Have you seen my site called AdamBlockStudios.com ? This is the official site where all the answers are! On this KZbin channel I am showing you "helpful" videos... but this is only the tip of the iceberg. Please see what others are saying here: www.adamblockstudios.com/categories/Reviews

[@astrofotoperu]

Hi Adam, thank you very much for this information. Is the pedestal value also recommended for broadband images too?

[@AdamBlock]

This usually isn't an issue for broadband. The sky is much brighter- so the negative values do not occur. You can always check your calibrated data in the way I demonstrate to see.

[@astrofotoperu]

@@AdamBlock Thank you very much. It esencial to know this information. Clear skies from Lima, Perú!

[@nickambrose8606]

So I suppose in theory, if WBPP did multiple passes, it could actually calculate the pedestal by doing the subtractions and maintaining the maximum negative value, then going back over the images ? Also, maybe a minor issue but in this case you might saturate pixels that are very close to the maximum ADU value (hopefully far fewer) which is also a loss of data, correct ? Just checked and I have a ton in my L-Extreme data -- I am going to re-stack and compare the two !

[@kevinmorefield824]

Nick, I think adding 100 ADUs to a 65535 ADU range will not noticeably affect star saturation. For sure you don't want to add more than you need but the relative difference between a 65435 count star pixel and 65535 count star pixel is 0.15% while the problem with the background being solved is something like 50% of the background values being lost.

[@dawnlowry3279]

Thank you Adam~~~. As usual . ..Awesome info.... I believe in one of my acquired (long way calibrating icons sets I had a 800 pedestal). I recently removed it ( was advised to.. now i see though.. ... can it hurt to globally apply a 500 pedestal all the time.... Nb or broadband in WBPP? how does it affect the image if you ~. dont have the ZERO's?

[@AdamBlock]

You are just eating into your "dyanmic range" of values. When you add a constant to every pixel... You have the bit-depth of your camera (16-bits, 65353) - the pedestal. These are a few hundred counts out of many thousands shouldn't be an issue at all. Some cameras have bias levels of 1,000-2,000 counts (16-bit). But the oversubtraction is really at the level of the readnoise and really only happens for very dark sky as seen with NB imagery. So I would think just 100 or 200 or whatever counts should do the trick (again, 16-bit... lower for smaller bit depth) and only for NB images.

[@gregmckay666]

I'm curious and just learning P.I. and astro processing but is this something that is preventable by setting the proper offset level in the capture software during image acquisition? Edit: Nevermind, I see in the other comments this question has been answered.

[@AdamBlock]

Not if the offset is in both the light data and the calibration data right? (when you calibrate..the offset would be removed?)

[@claude77573]

I got an error in WBPP because of calibration overcorrection. To address this, I have used the Pedistal value explained in this great video. However, to estimate the Pedistal value, I set the PI workspace pointer to IntegerRange to see the DN (Data Number) read by the pixels of my camera's 16-bit ADC sensor (i.e. in the range of 0 to 65,535, what level did the pixels register?). I searched for a rough estimate of the lowest DN in one of the light frames and the highest DN in a Dark frame. (If the DN in the light frame is lower than the DN in the dark frame, you have overcorrection.) Subtract your estimate of the lowest Light DN from the highest Dark DN to compute a Pedistal Value (defined as "the quantity added to the result of an image calibration to ensure positive pixel values") and enter that value in the Output pedestal (DN) field on the WBPP Calibration tab. While not a scientifically precise calculation, it might be a better estimate than a "try this or that number" approach.

[@AdamBlock]

Everything you say is correct. But the important idea is that the pedestal is *so much smaller* than the total range (65535)- just picking a value that you know exceeds the oversubtraction- even without looking- is OK. There isn't any damage on the low end... and at the top end there is usually plenty of room. Very few sensors saturation at 65535... more common is 40,000. So you have many thousands of DN to work with. Just add 200-300 and call it good. This issue with NB data is a function is NOISE... which necessarily has to be small valued. If you are correcting for large valued oversubtraction- you need to look for problems in the equipment or calibration data.

[@claude77573]

@@AdamBlock Thank you for your feedback. I am new to this, so I don't know what I'm talking about. Fortunately, I think I can spot someone who does, and you are a real Jewel to the astrophotography community!

[@andreaswohlt848]

Hallo Adam, thanks a lot. I‘ve made some pictures of M92 with 1000 mm fl and only 90 sec Exp time, reached a Mean of 800 ADU and this is fairly the same as in my darks 😖. Didn’t recognized it and now some of my cal files have a lot of zero values. For the best cal/int result, Is it better to use the pedestal +2xx in WBPP for all raw, or just PixelMath with +2xx on the detected raw with this low Mean value?

[@AdamBlock]

Sorry..I didn't quite understand the question. You can let WBPP calculate the pedestal for you to take care of all raw frames you are calibrating- or you can put in a number. The automatic method should work just fine really. I didn't understand your PixelMath question- the pedestal is applied internally during the calibration subtraction.

[@andreaswohlt848]

@@AdamBlock Sorry Adam... perhaps its my bad english and I couldn´t form the right words in the right order ;( Some of my lights have a median value above my masterdark (>100) caused by a little lightpollution by the beginning and the end of the session - this frames are well calibrated in nearly all pixels above 0. But the median in some lights (middle of the night) is nearly the same as in my masterdark and this creates a lot of 0 pixels. The question is, is it better to add some 100 ADU manually with PixelMath only to this low-median raws before calibrating all of them with masterdark/masterflat, or does WBPP does this job very well for me in the auto-pedestal mode. If the auto-pedestal is fine, does it add the same pedestal to all raw, or only to the detected raws with too low median? Hope that its now a little bit better to understand, thanks a lot, Andreas

[@AdamBlock]

@@andreaswohlt848 Thank you, I understand. I believe in the Auto mode- WBPP will find the numeric pedestal that is appropriate for each frame. If you input literal value- the same value will be able to every frame (like your PixelMath idea).

[@Aerostar509]

Would this apply to my SX-694 mono CCD?

[@billtinsley5385]

Adam, Great video !!! Very useful tool. I modified your expression to find saturated pixels to aid on determine exposures times (see below). This is something that I straggled as newcomer to astrophotography. Perhaps you could demonstrate this on your videos to the benefit of your viewers. Cheers, Bill R Ch- iif($T>=0.9990,1,$T) G Ch- iif($T>=0.9990,0,$T) B Ch- iif($T>=0.9990,0,$T)

[@davidmisisco]

Would adding an offset at the time of imaging do the same as adding this pedestal or are they serving different purposes?

[@AdamBlock]

I can tell you what would not work- if you have a CMOS camera- adding the offset to both the lights and calibration data would not do anything. :)

[@davidmisisco]

@@AdamBlock thanks. Either, or, but not both.

[@AnakChan]

I used the method above to test a recent Prawn nebula Ha I took back in May and was shocked by how many pixels I had with 0 value. The thing is I don't use WBPP but I calibrate manually. I see Pedestal Mode in ImageCalibration with defaults to Default FITS Keyword (Pedestal) which my FITS header for my sub does not have. So presumably 0. I can use Literal Value instead to add the pedestal. However I've got a fundamental question. How is the pedestal different from sensor offset? Both conceptually seems to be doing the same thing. Could I just use the Custom FITS keyword Offset instead? Or are the two fundamentally too different that I shouldn't use one for the other?

[@AnakChan]

Argh...think I can answer my own question. The darks would have the same offset so net 0. I need more morning coffee.

[@AdamBlock]

The short answer is that i am not certain. I would think that an electronic offset imposed on the electronics would exist in both your light frame data and darks- so it would subtract out and you would still be left with the issue?

[@AnakChan]

Yes and I think that's why the net is 0 'cos both the lights and the darks have the same offset. I could do a quick test by creating a small set of master darks with no offset (i.e. mismatch from the lights). Naturally not ideal for real world use as I expect other side-effects, but an interesting test nevertheless.

[@jimwaters304]

Would this also apply to OSC camera's that use L-eNhance or L-eXtreme filters?

[@AdamBlock]

I am not familiar with how dark the sky/background is in those filters- but any data in which the sky is sufficiently dark would be the thing to look at. As i hope you saw in the video...it is easier enough to check once...

[@jimwaters304]

@@AdamBlock Thanks Adam. I am checking my lights now.

[@desmcmorrow2978]

@@jimwaters304 I’m interested in the answer to this question so please let me know how you get on.

[@desmcmorrow2978]

I had a quick look at some of my L extreme subs. For 5 mins subs I didn't see a problem but for 3 mins subs there were a few pixels that were zero after dark subtraction so I will be adding a pedestal when using the L extreme. Thanks Adam and Jim.

[@AdamBlock]

@@desmcmorrow2978 A few pixels probably isn't a big issue... l

[@ibnulhussaini3791]

Does it matter if these negative pixels are over dark nebulae?

[@AdamBlock]

Yes... that is bad. :)

[@ibnulhussaini3791]

@@AdamBlock Just recently acquired Ha & OIII data of the Carina region. Interesting that those pixels were only in that V shaped dark region. Used an ASI2600MM @ the low read noise default settings of 100 gain & 50 offset

[@AdamBlock]

@@ibnulhussaini3791 Yes, the issue will show up in the darkest parts of the image...the low signal parts.

[@jrlpereira]

when will WBPP 2.1 be available?

[@AdamBlock]

It will be distributed with the update of PixInsight to 1.8.8... so we are waiting on Juan Conejero to button things up based on the findings of beta testers. Soon? Imminent? lol