 #...
·
1
like
|
|---|
Ruediger: I wanted to be clear on the point you make here. My flats are no where close to a bias frame. My flats are at 30% full well or above. It has nothing to do with exposure time. When I have done bias, the most I see is a few electrons per well. My last flats, at 0.03 sec, were most certainly in the linear range. I know this because I take note of the adu vs time to arrive at my desired exposure. They scale well with time. I do not believe my sensor is that different than most other CMOS cameras. After all, all of these astro camera sensors are found in the cameras we use daily in our phones, point and shoot, and SLR cameras. They are quite capable of accurately shooting over very wide isos and shutter speeds. These cameras routinely operate at shutter speeds well faster than what I am suggesting. |
18.84
#...
·
1
like
|
|---|
|
Hi Alan, sorry, but Bias frames have nothing to do with full well, They are only dependent on exposure time. Bias frame are made at the shortest possible speed to get the read out characteristics of the sensor. The shorter you make your dark flats (not flat) the closer you get to Bias frames. Please don't mix that up. An exposure of 0.03 is definitely in a critical range. Your chip is not different, but it works for you since probably the gradient introduced from your scope is super exceeding the sensor non linearity and it is not notably. DSLR do the correction in hardware. The vendor integrates the imaging processor which takes care of the compensation of non linearity even in RAW. Once again: If it works for you - fine. But this is probably simply good luck and I just want to make the OP aware that it might not work for him. He has to test. That's all. CS Rüdiger |
#...
·
1
like
|
|---|
Ruediger: I can agree with that. Not sure how that pertains to the issue here. The original question was about needing to take Flats. Not dark flats. Or flat darks, as some might say. Not sure this person cares about doing dark flats. I can assure you that I am working in a relatively linear capability of my camera. Actually both of them. And that is my direct experience. Might try faster, just to see where that breaks down. I agree, bias is a measure of dark sensor characteristics. Bias frames are done with a capped sensor because light would swamp the bias frame for its intended purpose. I am talking about flats here and only flats. Strong signal. Certainly has a bias component. But at 35,000 edu, I'll let you do the math as to how much the bias contributes to my flat frame. And for that, exposure is irrelevant. Now, if you claim some cameras will introduce artifactual signal at short exposure times, that is a different issue. Yes, I may be uniquely lucky! I don't do dark flats. Don't need too. |
3.71
#...
·
3
likes
|
|---|
|
All of this wonderful input will be passed on, certainly, folks, and on behalf of Martha, my thanks. The general consensus seems to be that, given her permanent situation, multiple layers or darker material is the key, and will work fine. For myself, I didn't think that a darker material would matter with a mono camera, so this information should verify that. She's mentioned reading articles about exposure length, proper histogram levels, etc., so this is her concern because she is having exposures from .2 to .4 seconds, and didn't feel these were long enough, based on what she's read. I've seen her Flats, which seem fine, but she was curious when I mentioned that my own flats, with an 11" Edge and QSI 683, were in the 3-5 second range, depending on filter. Thank you all again for your time and input. If the fires ever die down here in the western U.S. and skies clear, I'll get back to work . - - Steve |
|
18.84
#...
·
1
like
|
|---|
Alan Brunelle: For any CMOS it is mandatory to take absolute precise dark flats along with flats and no bias at all if you want to take advantage of your camera's full potential, since bias do not work with CMOS. If you are satisfied in any other way it is fine for sure. But for my ASI6200 it makes a huge difference between almost unusable and good stacks. |
|
18.84
#...
·
1
like
|
|---|
Steve Solon and Terry Chatterton: You are welcome! I'll keep my fingers crossed! CS Rüdiger |
15.85
#...
·
1
like
|
|---|
Ruediger:Alan Brunelle: This is all good great convo! I’ve learned somethings here myself so for me thats a win situation! When I’m done with my flats, I take down my light panel and cover the scope shut all the lights off and shoot my “dark flats” at the exact same settings as my flats. Only difference is there is NO light. @Ruediger I’m embarrassed to admit that I didn’t know that I shouldn’t have been using Bias frames with my CMOS cameras. I don’t know that I ever read that, so thanks for that tidbit of info! 😊 Dale |
|
18.84
#...
·
1
like
|
|---|
Dale Penkala: My pleasure! I went through this tedious learning process when switched from CCD to CMOS and it took me more than 6 month to find out why the results were so awful and only for the trashbin. Therefore I am glad if I can prevent anyone from suffering from the same. I had learned that CMOS reacts very drastically to bias and short darks whereas a CCD just shrugs its shoulders and spits out perfect results.  |
|
15.85
#...
·
1
like
|
|---|
Ruediger:Dale Penkala: Ruediger: Ditto! That covers anything I have learned “Astronomy Related” I help others get involved in astronomy, Astrophotography and as “WE ALL” know there is a learning curve. I struggled learning many things and if I can help others to learn “easier” then I always try to do that! Believe it or not I’ve had “others” ask why do you do so much to help others, let them learn just like I had to. My answer is: If I can help someone do something they want to do without the struggles I had then I’m doing my part in the wonderful hobby we call Astronomy. Dale |
1.81
#...
·
3
likes
|
|---|
| I use a translucent white acrylic sheet panel in front of my dimmable lighted tracing panel and I am able to get nice 30K flats with 4 different mono and OSC cameras. Thicker or multiple panels should provide additional light dimming and dispersion. |
22.40
#...
·
2
likes
|
|---|
|
Beside all of good reasons already mentioned, there is another reason to be careful not to use really short exposures with some types of flat panels. Some types of light sources, such as a flat panel display or an EL panel exhibit flicker at a multiple of the drive frequency. For larger scopes, EL panels are just about the only option for a large flat panel and many produce a cosinusoidal irradiance distribution across the panel that becomes evident with really short exposures. The pattern varies randomly in phase so any exposure longer than a second will time average the pattern. The panel also does not sit in an image plane, so the spatial variation is not directly imaged in the calibration frame but it is still a good idea to time average it with a slightly longer exposure measured in seconds rather than in milli-seconds. The brightness of LED panels are typically controlled using pulse width modulation so you want the exposure to be longer than at least 100 cycles of the modulation frequency, which suggests minimum exposures on the order of a second to achieve very repeatable results. I completely agree that the most important thing is to make sure that the flat signal is high enough to be in the linear region of the sensor but remember that signal is not the same thing as exposure time. Signal is given by irradiance*responsivity*time, which means that you can use very short exposures on a very bright source to produce a signal that lies well within the linear response region of the sensor. That also means that the only way to hold signal constant while increasing exposure time is to reduce irradiance, which you can do with t-shirts, a ND filter, or by electronically "turning" down the brightness of the source. John |
 #...
·
2
likes
|
|---|
John Hayes: Thanks John Ruediger: Thanks Ruedinger And to Alan and all these very interesting and important issues that were analysed under this topic, so I guess I should also thanks to the person that rise the point in the first place, hopefully it will become part of the AB comunity. In conclusion, I have to go back to study a few of the concepts that were mention during the thread, and to put them into practice, hopefully I will succeed with my damn bad flats I got for the ZWO294 and those bl...dy NB dual filters (L-Enhance and L-Extreme). Cheers and CS to everyone. |
#...
·
2
likes
|
|---|
Claudio Tenreiro: And much thanks to you for bringing up that issue with this filter type. I have some on backorder so when or if I ever get them I will remind myself to be on the lookout for that very issue! |
0.90
#...
·
1
like
|
|---|
|
Just curious but is there such a thing as having a flat exposure too long, provided you still achieve the desired ADU? Dale |
|
1.81
#... |
|---|
Steve Solon and Terry Chatterton: Steve, Sorry if I missed this, but did you mention what camera and filters Martha was using. A few months ago, one camera I tried briefly would not yield any usable flats in SII, and OIII and Ha had some strange sensor patterns. This was with my C925Edge and I had ghost imprints of my secondary mirror which could not be phased out, no matter how short the flats exposure. Regards, John |
|
1.81
#...
·
1
like
|
|---|
Dale A Chamberlain: Not sure if there is a limit, but if a see flat taking over 30 seconds or more to obtain, something is usually wrong and they don't look very appealing after processing. Maybe someone knows a mathematical limit as to why this occurs. My broadbands are usually around a second or less. NB maybe 7-15s ballpark. |
|
18.84
#...
·
2
likes
|
|---|
|
Hello friends, If you use e.g. sky flats changing light condition may become a topic, since the sun is moving. Less significant for a single flat, but when you take many. The summarized time gets significant. When taking long flat with a constant light source over the complete time e.g. Flat Panel, the topic "darks" become also an important point because you sum up a significant amount of dark current / thermal noise. But this is well controllable if you take proper darks. Also compared to the signal it is rather low. So actually not an issue, but needs to be payed attention to. Very long flats for UNB Filters (like 3,5nm) are not so uncommon e.g. I get to up 120s but I am using a Gerd Neumann flat and this woks fine. Or, if I am not patient enough, I stack Acrylic plates on top at day and try to get around 10s. But I have to admit: sky flat always have an slight gradient. But sometimes this quick and dirty solution is good enough and saves a lot of time. It is not always required to be perfect.  BTW: Just consider If a Lum flter needs 1s and covers the complete spectrum and now you punch out such a small window like a 3,5 UNB does you are quickly having extream long multipliers. There you need really very bright flats, or you have to increase exposure time heavily. Strange flats with UNB are also well known, since you identify very quickly any problem with the coating due to the very close HBW. I had an (defective) UNB Ha 3,5nm which produced very wired cloudy flats. Completely strange. First I though the flat has an issue in Ha frequency, but the culprit was the filter coating. I have got a new one (same type) and the problem was much better. The new versions are now absolutely homogeneous. CS Rüdiger Defective Ha 3,5nm UNB:  |
|
18.84
#...
·
2
likes
|
|---|
|
Just another interesting experiment for cold winter nights: Many of us have calibrated monitors and use a colorimeter like a Spyder to adjust it. These colorimeter can precisely measure the luminescence and color. Try that on your flat I did that on some and immediately returned them to the vendor. It turned out that many of them had huge surface luminescence deviations. Some up to 15%. Additionally very often unevenly distributed, so it matters in which position you use it. One was even so bad, the gradient of the flat panel was worse than the vignetting of the scope. So asked them: How should I calibrate my images, if the calibration device itself has a 15% deviation over the area? They all accepted the returned devices without any discussion. My conclusion: The quality of the flat devices is really ranging from unbelievable bad and useless to acceptable. Perfect was non of them. Either uneven, or missing (resp very weak) the frequencies required for NB filters (e.g. Red in Ha), which is a consequence of using LEDs with discrete spectrum. Conclusion: Try your colorimeter on your flat (but only if you are still in return period). This avoids frustration.  Cheers Rüdiger |
1.20
#...
·
1
like
|
|---|
|
Didn't read the whole thread, but you can get a tracing board fairly cheap. Adjustable brightness, some has its own battery others you just plug into a USB port. There are small/medium/large. If still need to defuse it more, you can tape sheets of paper on it. Just search amazon for tracing board. |
|
0.90
#...
·
2
likes
|
|---|
Not sure if there is a limit, but if a see flat taking over 30 seconds or more to obtain, something is usually wrong and they don't look very appealing after processing. Maybe someone knows a mathematical limit as to why this occurs. My broadbands are usually around a second or less. NB maybe 7-15s ballpark. That has been my experience as well. Sky flats haven't worked out for me on narrowband because when the sun sets there isn't enough time to get the same amount of light for narrowband filters. |
