Jump to content


Photo

Flea3, 16-bit?

  • Please log in to reply
11 replies to this topic

#1 bunyon

bunyon

    Gemini

  • *****
  • Posts: 3324
  • Joined: 23 Oct 2010
  • Loc: Winston-Salem, NC

Posted 24 January 2013 - 04:16 PM

Has anyone used the Flea3 in 16-bit mode? I clipped the box this afternoon imaging the sun and like my image. But I don't know if going to 8 bit would improve things or not? (I changed several things in my routine).

Anyone have any experience?


Here is the image:

Posted Image

#2 Kokatha man

Kokatha man

    Fly Me to the Moon

  • *****
  • Posts: 6995
  • Joined: 13 Sep 2009
  • Loc: "cooker-ta man" downunda...

Posted 24 January 2013 - 07:57 PM

.....I'll put my neck on the block here Paul :grin: and state that I really don't believe 16 bit offers much if anything to the outcomes obtainable within the image parameters of even the best planetary images we can capture at present as AA'ers.

In the future, who knows..? But imho the benefits of shooting at 16 bit when we edit our images in processing aren't anywhere near those obtainable in (for example) landscape etc photography where tonal variations etc and the editing/processing for shadows, highlights, gradients and their impact upon colour & detail etc is much more demanding and self-evident - wherein the benefits of RAW captures (16 bit) become far more important with these and other issues than 8 bit. :)

Others will have different opinions but why don't you do your own tests and convert one image to 8 bit and apply exactly the same processing steps to it and the 16 bit (preferably side-by-side in P/shop) and compare the results..! :)

#3 Kokatha man

Kokatha man

    Fly Me to the Moon

  • *****
  • Posts: 6995
  • Joined: 13 Sep 2009
  • Loc: "cooker-ta man" downunda...

Posted 24 January 2013 - 08:09 PM

ps: someone is bound to throw in that term "posterisation"... ;) :grin:

This is the technical term that in practice describes gaps in histogram created in processing (they become greater as processing becomes heavier in simple terms) and in a lot of imagery 256 values of tonal variation (8 bit) won't support said processing: why I mentioned gradients and colour variations etc.....but as I maintain, this isn't really an issue in planetary imaging imho! :)

#4 sfugardi

sfugardi

    Gemini

  • -----
  • Posts: 3077
  • Joined: 30 Sep 2006
  • Loc: CT

Posted 24 January 2013 - 08:15 PM

Paul, awesome solar shot! The problem with 16bit capture is the reduction in frame rate collection. My laptap barely handles 80fps on Jupiter 8bit, dropping to approx 45fps using 16bit. I plan to try it on Saturn this year since the frame rate will already be low enough. I have no idea if it will help, but it can't hurt if you have enough disk space

Regards,
Steve

#5 MvZ

MvZ

    Surveyor 1

  • *****
  • Posts: 1583
  • Joined: 03 Apr 2007
  • Loc: The Netherlands

Posted 24 January 2013 - 08:37 PM

The biggest problem with 16-bit in planetary imaging is that you will have trouble finding yourself in a condition when you need it. In 99% of the time it is a waste of disk space, processing time, and perhaps even makes you miss out on frames because the FPS is limited.

For pretty much all cameras that have the correct settings applied (most importantly a neutral gamma value!) you will find that if you see even a hint of noise visually in the images during recording, 8-bit is more than enough and 16-bit will add nothing.

Now for solar recordings you might find yourself in the 1% if and only if you hardly use any gain on your camera. For solar imaging in white light I set the gain of my camera to the minimal recommended setting (+- 3dB), and even those images had enough noise in them to not have any benefit from more than 8-bit capturing. That was with the Basler Ace, which is very similar to the Flea3. If you apply say 15% or more gain on your Flea3, I would spare the trouble of testing things out and stay away from 16-bit. Use 8-bit and higher frame-rates.

If you do decide to test things out (because that can be fun to do!) simply record in 16-bit ser, and transform that file to an 8-bit avi for example (using a reliable transformation tool). This will leave you with two identical recordings, except one of them has just 8-bit/pixel data and the other 16-bit. This is exactly what the camera also does, so it makes for the perfect test case (no influence of anything but the bit depth of the recording). If the conversion software has done a correct job (most importantly not messing about with brightness of the images!), you'll likely find the stacks are identical.

#6 Kokatha man

Kokatha man

    Fly Me to the Moon

  • *****
  • Posts: 6995
  • Joined: 13 Sep 2009
  • Loc: "cooker-ta man" downunda...

Posted 24 January 2013 - 08:45 PM

The biggest problem with 16-bit in planetary imaging is that you will have trouble finding yourself in a condition when you need it. In 99% of the time it is a waste of disk space, processing time, and perhaps even makes you miss out on frames because the FPS is limited.

For pretty much all cameras that have the correct settings applied (most importantly a neutral gamma value!) you will find that if you see even a hint of noise visually in the images during recording, 8-bit is more than enough and 16-bit will add nothing.

Now for solar recordings you might find yourself in the 1% if and only if you hardly use any gain on your camera. For solar imaging in white light I set the gain of my camera to the minimal recommended setting (+- 3dB), and even those images had enough noise in them to not have any benefit from more than 8-bit capturing. That was with the Basler Ace, which is very similar to the Flea3. If you apply say 15% or more gain on your Flea3, I would spare the trouble of testing things out and stay away from 16-bit. Use 8-bit and higher frame-rates.


.....another way of saying what I've said Emil with the added points about disk space and possible framerate reduction: noise is one of those factors I didn't articulate specifically but as you say it shouldn't come into the equation - but Paul can easily do a comparison by converting his 16 bit RAW captures to 8 bit (if he has any decent software! :grin:) and do the comparison I suggested.

But essentially you and I are in complete agreement on this! :)

edit: I see you've edited!!! :lol:

#7 bunyon

bunyon

    Gemini

  • *****
  • Posts: 3324
  • Joined: 23 Oct 2010
  • Loc: Winston-Salem, NC

Posted 24 January 2013 - 08:59 PM

Thanks guys. I was only thinking of doing this for solar imaging. Imaging at prime with the PST, I had the gain all the way down and an exposure of 0.3 ms. I couldn't see noise but it is hard on such a recording.

For what it's worth, I was able to capture at 67 fps in 16 bit.

What is good software to convert the ser to 8 bit avi?

#8 ValeryD

ValeryD

    Vendor (Aries)

  • *****
  • Vendors
  • Posts: 1147
  • Joined: 26 Nov 2005
  • Loc: Kherson, Ukraine.

Posted 25 January 2013 - 12:30 PM

My own experience tells me that 16 bit is definitely a plus when I imaging disk with prominences. Also, post processing is easier when I trying to make subtle details better visible.

I have installed a very fast hard disk (with flash memory) and I see no difference in speed recording.

I also use 2TB disk splitted on two for temporary recording. External 4Tb disk is for saving of the movies and pictures.

I do use a Flea 3 GigE mono version.

Valery.

#9 MvZ

MvZ

    Surveyor 1

  • *****
  • Posts: 1583
  • Joined: 03 Apr 2007
  • Loc: The Netherlands

Posted 25 January 2013 - 01:05 PM

Could you share your camera settings Valery? And perhaps also example images?

It would really surprise me if the difference you see are actually from using 16-bit or if there is something else going on. If you have the option, I'm also very much interested in one of your 16-bit recordings just to see if it is converted to an 8-bit recording the 16-bit really is superior.

There have been more than one 8 vs 12 vs 16bit discussions, and the conclusion is that with gain values above extremely low there is no reason the 16-bit should produce better results, and I'm very much interested to learn from that. (in optimal conditions the CCD in the camera can do only 10 bit. If you use a bit of gain and start seeing noise in the image, the data is already below 8-bit effective, capturing in 16-bit would then add nothing).

#10 Mitchell Duke

Mitchell Duke

    Gemini

  • *****
  • Posts: 3175
  • Joined: 22 Apr 2008
  • Loc: Georgia

Posted 25 January 2013 - 01:18 PM

Try PIPP https://sites.google...site/astropipp/

This is what I have used in the past to convert .SER to AVI format.

#11 Mike Phillips

Mike Phillips

    Soyuz

  • *****
  • Posts: 3523
  • Joined: 21 Feb 2006
  • Loc: Swift Creek, NC - 35.682 N, 78.743°W

Posted 25 January 2013 - 04:04 PM

I blindly use 16-bit for everything. I ROI it at capture to reduce the overhead and wasted 'black' edges. I capture to ramdisk and get nearly 100% rate of whatever I record so no issues there. Am I getting benefit over 8 bit, prerhaps not.

Surprised to see no one has mentioned it, but modern cams record a 12-bit depth in a 16-bit container, so it's not as sexy as it sounds either way!

Mike

#12 ValeryD

ValeryD

    Vendor (Aries)

  • *****
  • Vendors
  • Posts: 1147
  • Joined: 26 Nov 2005
  • Loc: Kherson, Ukraine.

Posted 26 January 2013 - 10:29 AM

Could you share your camera settings Valery? And perhaps also example images?

It would really surprise me if the difference you see are actually from using 16-bit or if there is something else going on. If you have the option, I'm also very much interested in one of your 16-bit recordings just to see if it is converted to an 8-bit recording the 16-bit really is superior.

There have been more than one 8 vs 12 vs 16bit discussions, and the conclusion is that with gain values above extremely low there is no reason the 16-bit should produce better results, and I'm very much interested to learn from that. (in optimal conditions the CCD in the camera can do only 10 bit. If you use a bit of gain and start seeing noise in the image, the data is already below 8-bit effective, capturing in 16-bit would then add nothing).



Sorry, I didn't saved my 8bit vs 16bit records. If I imaging sunspots and surrounded regions at near the center of the disk, I do use 8bit - enough.
However, when I imaging almost full disk or great part of disk + edge with prominences, I always do use 16bit.

I will start to imging when the warm season will come here. Hate a winter so much!






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics