Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
08-05-2020, 01:50 AM   #1
Junior Member




Join Date: Apr 2017
Posts: 56
Samyang 24mm ED AS UMC milky way?

Hi all,

is there any K3 owner of this lens that can share a milky way shot , please ? I'd like to have a straight sense of how big is the milky way in the frame. Any additional info about the lens, with reference to the astrophotography in particular, is really appreciated.

The reason is because I own the 10mm and it is a little too wide for my idea and I found a quite cheap vendor for the 24mm that I can afford.

Really thanks
Danilo

08-05-2020, 08:21 AM - 3 Likes   #2
Site Supporter
Site Supporter
MossyRocks's Avatar

Join Date: Nov 2017
Location: Minnesota
Photos: Gallery | Albums
Posts: 1,571
I don't own either but when you say a little too wide you would find the 24mm to be a lot narrower giving a field of view less than half that of your 10mm. So while the 10mm may require cropping if you are cropping substantially less than 50% in each direction the 10mm would likely be a better fit. Some other things to consider are if you have the ability to use astrotracer since astrotracer would likely work fairly well with the 24mm lens but with the 10mm one expect sometrailing in the corners and edges. Now if you are cropping some and have the main image more centered that won't be a problem. If you don't have astrotracer the 1.5x exposure time you can get with the 10mm over the 24mm would greatly benefit the image all else being equal but it looks like the 24mm is 2 stops faster, so it would appear that in the end exposure would end up being a wash. If you don't have astrotracer expect to take shots of about 20 seconds with the 10mm and with the 24mm lens take shots at about 8 seconds to produce nice pinpoint stars. From there push the ISO up enough to where you clearly see the Milky Way and get out early so you can get your test shots and get everything set before you need to actually start shooting.

Sorry I couldn't be more help but hopefully I at least managed to better focus your thoughts. Also look into stacking shots and using Sequator. If you are looking to get the Milky Way in a specific location when stacking start taking pictures several minutes before it gets there and continue for the same amount of time after. Then in Sequator choose your reference frame as the one in the middle that has the best positioning. I assume you have a foreground object in mind so do careful sky selection and use the freeze ground option so you get a nice frozen foreground and nice frozen sky. By having a frame somewhere in the middle be the reference frame you will have good noise reduction from stacking across the entire image. Then just take lots of shots, stack, and edit until you are pleased.
08-05-2020, 08:53 AM - 1 Like   #3
Junior Member




Join Date: Apr 2017
Posts: 56
Original Poster
QuoteOriginally posted by MossyRocks Quote
If you don't have astrotracer the 1.5x exposure time you can get with the 10mm over the 24mm would greatly benefit the image all else being equal but it looks like the 24mm is 2 stops faster, so it would appear that in the end exposure would end up being a wash.
first of all thanks a lot. I'm pretty much with you about everything you wrote . Just the sentence above is not too clear: what do you mean for "being a wash"? I am expecting to use OGPS-1 (I own it) using that big aperture to decrease ISO and noise . I do not expect that aperture is a problem in the night. The image below is a 17 + 3 sequator stacking , 17 for the sky and 3 for the foreground , and it has been cropped . Samyang 10mm is the lens.

I assume the 24mm on apsc could give me quite a good view in portrait mode but i guess - so here is the question - on landscape it will give ... artistic view only (tree silhouette, far mountains peak detail, ancient building etc)
Attached Images
 
08-05-2020, 09:54 AM - 1 Like   #4
Otis Memorial Pentaxian
stevebrot's Avatar

Join Date: Mar 2007
Location: Vancouver (USA)
Photos: Gallery | Albums
Posts: 36,918
QuoteOriginally posted by mctaveck Quote
I assume the 24mm on apsc could give me quite a good view in portrait mode but i guess - so here is the question - on landscape it will give ... artistic view only (tree silhouette, far mountains peak detail, ancient building etc)
24mm on APS-C is not particularly wide, about 53 vertical and 38 horizontal in portrait orientation. Whether that meets your needs is hard to say without actually being there.


Steve

08-05-2020, 10:21 AM - 1 Like   #5
Site Supporter
Site Supporter
MossyRocks's Avatar

Join Date: Nov 2017
Location: Minnesota
Photos: Gallery | Albums
Posts: 1,571
QuoteOriginally posted by mctaveck Quote
Just the sentence above is not too clear: what do you mean for "being a wash"? I am expecting to use OGPS-1 (I own it) using that big aperture to decrease ISO and noise . I do not expect that aperture is a problem in the night.
What I meant by being a wash is that if you ran both lenses wide open or even stopped down 1 stop (to clean up chromatic abberations) that the difference between the 24mm f/1.4 and the 10mm f/2.8 would be negligible as the 24mm would let it 2 stops more light but you can do an exposure 1.5 times as long (1.5 stops) with the 10mm resulting in about a .5 stop difference in exposure so basically a wash and you would end up with basically the same amount of noise in each shot, well slightly more in the 10mm shot but not a lot.

Aperture still matters at night but not really for depth of field. It is mostly to let in more of those sparse photons and control chromatic aberrations. In looking at the 24mm and some sample shots it looks like it likely can be run more open than the 10mm so that right there would make a bigger difference as you instead of a .5 difference in exposure you would likely be at 1.5 or even 2 stops difference. The fact that you have astrotracer would now make the 24mm a lot more appealing as it will have a lot fewer problems in the corners and edges with trails. If I look at your image above with the 10mm there is trailing in the corners which is actually really common with astro tracer and ultrawides. With a good calibration you should be able to run a good 2 minutes with the 24mm doing astrotracer and still have excellent corners. Add in f/1.4 and running at f/2 you might be able to get away with ISO 400 and maybe ISO 200 shots of the sky for some real low noise shots with a lot of room to work in post after stacking. In that case I would look to dong a stitched panorama because that also would give you a benefit of having more pixels and thus being able to shrink any chromatic abberations and trails in the final image thus giving a better end result. You may want to look into using a program like Hugin for creating the panorama as it will stitch a mosaic from all the frames and then blend in a foreground shot for nice sharpness in the correct location. So I would expect to shoot basically 4 stacks (5 to 10 shots per stack) for the sky with something like a 25% to 33% overlap (one for each corner), do the alignment in Hugin and then use the command line tool for combining the images (use the Sigma method for averaging as it will remove satellites, planes and other transient things). For the foreground shoot images (bottom left and bottom right) with astro tracer off in bulb mode and combine again in Hugin. Then edit the sky image so it looks nice and then blend in the long exposure foreground images.
08-05-2020, 11:05 AM   #6
Junior Member




Join Date: Apr 2017
Posts: 56
Original Poster
QuoteOriginally posted by MossyRocks Quote
1) In looking at the 24mm and some sample shots it looks like it likely can be run more open than the 10mm so that right there would make a bigger difference as you instead of a .5 difference in exposure you would likely be at 1.5 or even 2 stops difference.

2) The fact that you have astrotracer would now make the 24mm a lot more appealing as it will have a lot fewer problems in the corners and edges with trails.

3) If I look at your image above with the 10mm there is trailing in the corners which is actually really common with astro tracer and ultrawides.

4) In that case I would look to dong a stitched panorama because that also would give you a benefit of having more pixels and thus being able to shrink any chromatic abberations and trails in the final image thus giving a better end result.
.
Ok got the sense of "washed", thanks. So the 4 points above are the main reasons why I'm buying the 24mm; very very interesting to me I end up to these same assumption of you. I guess the portrait images can be nice even without panorama but it is just a guess. In case landscape is the ideal shot I can use the 10mm or have a pano with the 24mm : camera in portrait and taking 3 or 4 shots.

well, I know-ish about pano, I did some mw core with the 50mm (plastic fantastic) but with no enough noise control despite a low ISO used, sure a matter of bad Post process. Now I learn some trick to reduce noise and increase sharpening at the same time . Also I'll follow your suggestions and I'll use that tool ! Very very appreciated your help (below the pano, f1.8, 80 seconds per shot 500ISO, NR long exposure ON. For sure I should reprocess that raws)

Thanks a lot!
Danilo
Attached Images
View Picture EXIF
PENTAX K-3  Photo 
08-05-2020, 11:41 AM   #7
Site Supporter
Site Supporter
MossyRocks's Avatar

Join Date: Nov 2017
Location: Minnesota
Photos: Gallery | Albums
Posts: 1,571
QuoteOriginally posted by mctaveck Quote
Now I learn some trick to reduce noise and increase sharpening at the same time . Also I'll follow your suggestions and I'll use that tool ! Very very appreciated your help (below the pano, f1.8, 80 seconds per shot 500ISO, NR long exposure ON. For sure I should reprocess that raws)
For sharpening without enhancing noise you want to play around with the high pass tool. In this thread Pete_XL pointed me to it and a tutorial on using it. Best results can be had if you remove the stars first. To do that I use StarNet++ to create a starless image and then subtract that to have an image with just the stars in separate layers that I can work on individually. Work with high pass on the starless layer to bring out the details.

I actually advise against using the in camera long exposure noise reduction. You can get better results by taking your own dark frames (10 to 20 is usually pretty good) at the end of your shooting session as most astro stacking programs will combine them into a master dark and then use that image to perform the corrections on each of the light frames. This removes a lot of the random noise that would otherwise eat some of the limited signal you get in each light shot by using the in camera option. Also shooting at a lower ISO means that the need for dark frames is lessened especially if dealing with drifting objects (the night sky) so that the error isn't in the same location each time.

Last edited by MossyRocks; 08-06-2020 at 07:38 AM. Reason: Forgot to put in the link
08-05-2020, 03:26 PM   #8
Junior Member




Join Date: Apr 2017
Posts: 56
Original Poster
QuoteOriginally posted by MossyRocks Quote
For sharpening without enhancing noise you want to play around with the high pass tool. In this thread Pete_XL pointed me to it and a tutorial on using it. (CUT)
Thank you so much about all these tons of suggestions ! I'm not understanding what thread contains the tutorial on using pass tool, maybe a link missing?
To remove noise I build a mask using "find edge" on a channel (the most contrasted) that I think cover the same logic. Anyway I'll check Pete_-XL's post for sure.
As per the NR:. do you think stacking 10 Images with NR ON gives a worst result compared with stacking 10 images with NR OFF + 3 or 4 dark frame? I thought the dark frame that the K3 is subtracting is the closer possible noise sensor image, because of the sensor temprature etc. Also I thought PENTAX knows how to deal its own raw file compared to a 3P software as Sequator , don't you agree?

Thanks again !

Danilo

08-06-2020, 09:06 AM   #9
Site Supporter
Site Supporter
MossyRocks's Avatar

Join Date: Nov 2017
Location: Minnesota
Photos: Gallery | Albums
Posts: 1,571
QuoteOriginally posted by mctaveck Quote
I'm not understanding what thread contains the tutorial on using pass tool, maybe a link missing?
That was my fault I did forget to put the link in. I updated my previous post.

QuoteOriginally posted by mctaveck Quote
As per the NR:. do you think stacking 10 Images with NR ON gives a worst result compared with stacking 10 images with NR OFF + 3 or 4 dark frame?
In the scenario you describe the difference will be marginal. The problem with in camera noise reduction is that with a single dark frame true random thermal/quantum noise still tends to be pretty large. Because it is truly random that error happens anywhere so the dominating random portion, not the systematic error you are trying to remove, is not the same from shot to shot. So now when the camera does dark frame subtraction not only is it subtracting some of the systematic error it is also subtracting random noise that is not the same random noise as what is in the shot. So when that subtraction happens it eats a bit of the signal or maybe misses removing some random noise from the light frame or maybe actually gets lucky and removes some portion of the random noise. At best subtracting random from Signal+random gives signal+random but because there are so many random bits and the probability that doing the random subtraction improves things is actually less than 50% the image is degraded some. However you did remove some of the systematic error, not all of it for the exact same reason as above, which does improve the overall image.

Now instead if you take a bunch of dark frames at the end of your session (the temperature doesn't change that much unless you are out from dusk to dawn assuming you acclimated your gear for a bit before you started shooting) and are not using the in camera dark frame subtraction. Now using some rather simple math the stacker program creates a master dark frame from the pile of darks you took at the end. These are all averaged together which reduces the random noise in the master dark to 1/sqrt(N) where N is the number of dark frames you took. So if you took 4 dark frames manually the random noise would be about 1/2 what it would be from a single shot. So with 4 shots the noise floor got pushed down 1 bit in the data. Now 4 dark frames isn't a whole lot which is why I typically suggest at least 10 to 20 as this will now really start to push down the random noise floor making the systematic error more visible. So at 16 shots the random noise in the master dark is about 1/4 that of the random noise in a single dark frame and the noise floor has been pushed down about 2 bits. I keep using about as even 16 shots is still fairly small so as things average out there is still a good amount of variation but once you get to 20 shots the likely hood that you have not pushed the noise floor down 2 bits is very slim. So now with our master dark we have a much clearer picture of the systematic error in the system that we then subtract from each light frame. The results in subtracting less of the actual signal due to randomness as we have driven it down by 1 to 2 bits. So we still have lost some signal but a lot less.

QuoteOriginally posted by mctaveck Quote
Also I thought PENTAX knows how to deal its own raw file compared to a 3P software as Sequator , don't you agree?
Noise reduction using dark frames is all about subtracting the dark frame. The camera only takes one dark frame so there isn't much it can do other than just subtract it. And while a good portion of the systematic error gets subtracted it also eats some signal. In the best case using the 1/sqrt(N) formula with one dark frame we get 1 which means no change but since it is randomness and we are just subtracting it means at best no change which means that we will be degrading the image. By doing it in post processing we have more information and can do more complex math to get a better result and the more data you have the better that result will be. There still is the benefit of subtracting out the systematic error that wasn't destroyed by the noise so it is sub optimal. The only way to deal with the random thermal/quantum noise and shot noise in the light frames is to take more light frames which means turning off the in camera noise reduction to increase the number of shots you can take in a session, see below for a discussion about this.

Another point on why I suggest shooting your own dark frames is that you have a limited amount of time when shooting. Lets say 30 minutes (it makes the math easy) and you are taking 30 second shots. If you are doing in camera dark frame subtraction you get a total of 30 light frames from your session and then you pack up your gear and go home. Now if instead you turn off the in camera noise reduction you end up with 60 light frames. This is an entire stops worth of signal to be stacked. Also what I do at this point is put the camera aside with the lens cap on and view finder cover on and let it shoot a bunch of dark frames while I pack everything up. If you are like me it takes 10 to 15 minutes (I have a huge heavy tripod that needs tools to take down) so by the time the camera makes it into the car it has captured 20 to 30 dark frames. So by the time I leave I have now managed to collect double the light frames meaning I have an additional stop of signal to work with and when I do my dark frame subtraction on each of those light frames I will have pushed the noise floor down more than 2 bits so will be preserving 2 additional bits of signal data in each shot. We both spent the same amount of time out in the field but I have more signal data and more correction data meaning I have a much better base to start my processing and editing from.

The only time I would use in camera dark frame subtraction is when I am taking a one off shot but most of the time when there would be a noticeable benefit of doing so I am not taking 1 shot and likely over 100 so I prefer to gather those additional light frames to get an additional stop of exposure (signal) which will provide a big benefit and then gather the darks at the end while I pack up. 6111666969248776
In doing so I can capture a shot like this from what many astrophotographers consider a fairly light polluted spot (bortle class 5). Granted Andromeda is fairly bright but in that shot I was running the camera at ISO 3200 and that was at the darkest place in the Twin Cities metro area (I'm in Minnesota) and to get darker I have to go on a 2 hour drive instead of a 40 minute one.

Also I get the feeling I am helping out my competition for this months photo contest of the night sky.
08-07-2020, 02:33 AM   #10
Junior Member




Join Date: Apr 2017
Posts: 56
Original Poster
QuoteOriginally posted by MossyRocks Quote
Also I get the feeling I am helping out my competition for this months photo contest of the night sky.
A-HA ! that's not my case since I've already post my image. Then, man, what can I say. There is a huge amount of information in your posts and in your link that I'm copying it in my personal folder "manuals & tutorial" . I can only say thanks once again even if now ... I have to study !
QuoteOriginally posted by MossyRocks Quote
Another point on why I suggest shooting your own dark frames is that you have a limited amount of time when shooting. Lets say 30 minutes (it makes the math easy) and you are taking 30 second shots. If you are doing in camera dark frame subtraction you get a total of 30 light frames from your session and then you pack up your gear and go home.
That's interesting. I found boring to take shots with NR on, of course , but there was a reason for it. Now this suggestion, along with the other one, let me change my mind. I also guess for another reason: let say I'll take 10 images with a lens (10mm) and other ten images with another lens (24mm) with different subject or foreground; each shot is 30 seconds. I'll have 600 seconds extra time. Following your suggestion i can only take 10 dark frame, while having a beer, for 300 seconds and reuse the dark frames both for the 10mm and the 24mm post processing isn't it?
08-07-2020, 06:40 AM - 1 Like   #11
Site Supporter
Site Supporter
MossyRocks's Avatar

Join Date: Nov 2017
Location: Minnesota
Photos: Gallery | Albums
Posts: 1,571
QuoteOriginally posted by mctaveck Quote
Following your suggestion i can only take 10 dark frame, while having a beer, for 300 seconds and reuse the dark frames both for the 10mm and the 24mm post processing isn't it?
Assuming that the shots from the 10mm and 24mm lens were shot at the same shutter speed and also same ISO yes you can reuse the dark frames for both since they have nothing to do with focal length.

QuoteOriginally posted by mctaveck Quote
A-HA ! that's not my case since I've already post my image. Then, man, what can I say.
Even if you were I would still help out and I was just joking around some.
08-08-2020, 12:02 PM   #12
Loyal Site Supporter
Loyal Site Supporter




Join Date: May 2007
Location: Flagstaff, Arizona
Posts: 690
Here are a couple of 24 mm shots on a K-3, taken last night (SooC jpgs). Lens used was Pentax FA 20-35mm F4 AL set to 24 mm. These were taken with the camera on a tripod. The first shot is parallel to the horizon, for the second one, I tipped the camera so the long axis of the frame was along the Milky Way. These are downsampled to 20% of original to meet forum image size guidlelines.

I guess these should be a very similar FoV to what the Samyang 24 mm would give you.
Attached Images
View Picture EXIF
PENTAX K-3  Photo 
View Picture EXIF
PENTAX K-3  Photo 
08-08-2020, 04:03 PM   #13
Junior Member




Join Date: Apr 2017
Posts: 56
Original Poster
Really Thanks! With this sample I think portrait will be perfect, at least where I live; landscape will be tricky but I'd give a try. Thanks again to share these to me

Danilo
08-08-2020, 04:32 PM   #14
Pentaxian




Join Date: May 2015
Photos: Gallery
Posts: 532
QuoteOriginally posted by MossyRocks Quote
What I meant by being a wash is that if you ran both lenses wide open or even stopped down 1 stop (to clean up chromatic abberations) that the difference between the 24mm f/1.4 and the 10mm f/2.8 would be negligible as the 24mm would let it 2 stops more light but you can do an exposure 1.5 times as long (1.5 stops) with the 10mm resulting in about a .5 stop difference in exposure so basically a wash
I am sure you didn't mean to say that an exposure 1.5 times as long was 1.5 stops.

08-09-2020, 08:40 AM   #15
Site Supporter
Site Supporter
MossyRocks's Avatar

Join Date: Nov 2017
Location: Minnesota
Photos: Gallery | Albums
Posts: 1,571
QuoteOriginally posted by slartibartfast01 Quote
I am sure you didn't mean to say that an exposure 1.5 times as long was 1.5 stops.
That was my mind getting ahead of myself and not writing it correctly. The additional exposure time with the 10mm woudl be an additional 1.5 times that of the 24mm so if you got good results at 8 seconds with the 24mm you would get good results at 20 seconds with the 10mm.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
24mm, dslr, k-3, k3, lens, pentax k-3, samyang 24mm
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
For Sale - Sold: 14mm 1:2.8 Samyang ED AS IF UMC lens (Price Reduced) $150 BarrieKM Sold Items 3 05-15-2019 08:31 AM
Samyang 135mm f/2.0 ED UMC - Lowest ever price on Amazon jab2980 Pentax Price Watch 4 04-23-2019 04:54 PM
Samyang 14mm F2.8 ED AS IF UMC auto-aperture problem sky_light Troubleshooting and Beginner Help 7 12-22-2018 07:14 AM
For Sale - Sold: Samyang 24mm F1.4 ED AS IF UMC wide angle lens ($349) Lou Kraus Sold Items 2 05-25-2018 11:07 AM
For Sale - Sold: *PRICE LOWERED* Bower 35mm F1.4 AS UMC (Same as Samyang /Rokinon 35mm F1.4 AS UMC) mr.b.snowden Sold Items 2 07-23-2015 09:47 AM



All times are GMT -7. The time now is 10:20 PM. | See also: NikonForums.com, CanonForums.com part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top