Author Topic: I'll be speaking at MWAIC '10  (Read 23038 times)

Offline RBA

  • PixInsight Guru
  • ****
  • Posts: 511
    • DeepSkyColors
Re: I'll be speaking at MWAIC '10
« Reply #15 on: 2010 July 24 07:26:08 »
Also glad to hear it was well received. I'm sure you did a terrific job!
What topics did you cover?

Offline Nocturnal

  • PixInsight Jedi Council Member
  • *******
  • Posts: 2727
    • http://www.carpephoton.com
Re: I'll be speaking at MWAIC '10
« Reply #16 on: 2010 July 24 20:17:32 »
Thanks guys. Even got very nice feedback from the Adobe representative that was there as well as Ken Crawford, Warren Keller, Mike Unsold, Craig Stark and the other speakers. Generally a very good atmosphere there. I had a chance to do some 1-1 work with some folks to help them remember what I demonstrated.

I think RBA will be well received at AIC. When Jack was at AIC in '08 we were still outsiders. That's changing.

Also learned a great deal from Ken and Adam's PS demonstrations. I didn't like everything I saw but definitely could tell the appeal of the layering strategy. It's very powerful to adjust layer parameters and update your image. This doesn't work for all processes but it does for quite a few. According to Alan from Adobe that feature will be extended to include more heavy-duty processing. PI can't match that right now. You can undo and change your processing parameters but changes don't ripple forward. Of course with smart process container management it can be emulated. Conceptually PS is really complicated though. So many layering techniques. It's understanding that there is such a good market for training DVDs.
Best,

    Sander
---
Edge HD 1100
QHY-8 for imaging, IMG0H mono for guiding, video cameras for occulations
ASI224, QHY5L-IIc
HyperStar3
WO-M110ED+FR-III/TRF-2008
Takahashi EM-400
PIxInsight, DeepSkyStacker, PHD, Nebulosity

Offline Juan Conejero

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 7111
    • http://pixinsight.com/
Re: I'll be speaking at MWAIC '10
« Reply #17 on: 2010 July 25 01:22:31 »
Hey Sander,

Congratulations! I am very happy to know that your presentation was a success and everything was fine. I appreciate all of your hard work and enthusiasm to support PixInsight and everything related to our project.

Quote
It's very powerful to adjust layer parameters and update your image. This doesn't work for all processes but it does for quite a few. According to Alan from Adobe that feature will be extended to include more heavy-duty processing. PI can't match that right now.

In terms of pure image processing, there is nothing you can do with layers that can't be done better with PixelMath, masks and previews in PixInsight. PixelMath is much more powerful (in mathematical terms), accurate and versatile than PS's layering system, and has no bit-depth or data type limitations (it can work irrespectively from 8-bit integers to 64-bit floating point). I can demonstrate this when you want.

The only class of tasks where layers are better than our PixelMath is interactive painting. This includes the ability to move individual layers manually, which is something that PixelMath obviously can't do (you can move images with PixelMath only in a programmatic way). This includes also text and vector-based drawing, as well as bitmap rendering and composition operations. When I implement the LayeredComposition tool that I have on the drawing board (perhaps next Winter), I'll include most of PixelMath's operations and the full set of Porter-Duff compositing operators. The tool will be dynamic with full manual interaction.
Juan Conejero
PixInsight Development Team
http://pixinsight.com/

Offline mmirot

  • PixInsight Padawan
  • ****
  • Posts: 881
Re: I'll be speaking at MWAIC '10
« Reply #18 on: 2010 July 25 05:11:54 »
It was a great talk Sander. I enjoyed it ,so did the audience.  He drew the short straw and had to go first too :angel:


Juan, I think Sander means you can modify your processes in real time without going backward.
In PS you can an attach process to layers that effects the stack of images.  For example, you might have levels, curves, hue, saturation on a layer.
You just click on it ; it opens and you modify the settings. It immediately effects the image. It should be noted that not every function in PS can be performed this way.

I could easily imagine similar function in PI without layers. It would be great if you had a history of applied  processes and could just go back and modify a processing step non sequentially.  Perhaps, my second step is a histogram stretch, the third curves and fourth is deconvolution.  Click on the curves icon in the history and the module opens . You just change the settings which effect the image.

BTW, Layers themselves are convenient and intuitive way to combine images. As a user interface it is a powerful feature.

Max

Offline RBA

  • PixInsight Guru
  • ****
  • Posts: 511
    • DeepSkyColors
Re: I'll be speaking at MWAIC '10
« Reply #19 on: 2010 July 25 07:11:43 »
Um... Another discussion about layers?  ;D

I'll save my opinion for now, but a layering UI that carries the power of PixelMath is the one thing I haven't seen anywhere yet and I'd love to see - somewhat that's what I think the LayeredComposition tool would offer, among other things. What I don't know is whether one would be able to enjoy realtime previews as they adjust their "blending operations" and such... (of course that's my way to say "I hope it does"  >:D )

Anyway enough about that.

Sander, my presentation at AIC won't be a "PixInsight" presentation, more about "this are some of the less obvious things I do when I process my images" but of course that means there'll be a lot of PI in it. So it's more "applied PI" than anything else, but I'll also cover other things such as composition, etc. That's also why I asked you what topics you covered (if you don't mind sharing)  :)



Offline Nocturnal

  • PixInsight Jedi Council Member
  • *******
  • Posts: 2727
    • http://www.carpephoton.com
Re: I'll be speaking at MWAIC '10
« Reply #20 on: 2010 July 25 07:31:20 »
Hi Juan,

I understand at least some of the limitations of the layering system and it appears Adobe does too and wants to address them. I think Max explained it well that it's 'cool' to go to layer 2 and adjust some parameter and have the effect ripple through the entire stack without having to redo the layers above it. Personally I don't care to draw on layers and I don't care about text either. Those are not PI tasks as far as I'm concerned. I even dared tell Ken Crawford that I cringed when he was drawing on his layers to remove masking to bring out selected detail in one of his demos. That's something I would rather not do.

A long time ago I thought of a system in PI where modifying one of the processes in the history would cause (optionally) all subsequent operations to be re-applied. This would be similar to the layering system in that you can go back n steps and tweak things ala 'How would things look if I had stretched a little more'. Like I said this can be done with process containers but I can see the appeal of PS in how it deals with things. I can also see the danger :)

Hi Rogelio,

I talked about PI in general and then did a histogram stretch on an M31 crop followed by HDRWT to show how to use previews to experiment in order to find the right parameters for a process.  Of course this also showed how to regain details lost in blown out bright areas. I then showed how to use DSS imagery to find DBE sample spots. This involved registering the DSS image, lots of messing with process icons and also running the waveletextraction script to find problems in our background image. I quickly touched on erosion in the MT process together with starmask to reduce star sizes but ran out of time so I only showed the effect. Actually I knew the last demo wouldn't work, I had it in my back pocket just in case. There was still this fear I'd run out of stuff to say before my time was up :)
Best,

    Sander
---
Edge HD 1100
QHY-8 for imaging, IMG0H mono for guiding, video cameras for occulations
ASI224, QHY5L-IIc
HyperStar3
WO-M110ED+FR-III/TRF-2008
Takahashi EM-400
PIxInsight, DeepSkyStacker, PHD, Nebulosity

Offline Juan Conejero

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 7111
    • http://pixinsight.com/
Re: I'll be speaking at MWAIC '10
« Reply #21 on: 2010 July 25 07:33:52 »
Quote
you might have levels, curves, hue, saturation on a layer.

Note that only those processes that can be applied as LUT transformations (look-up tables) can be applied as layers in that implementation. That only works for integer images and is actually a trivial implementation.

You can get the same functionality with a ProcessContainer and a preview in PixInsight. The advantage is that you can use any process, not just LUT-based transformations, and any data type, including 32-bit and 64-bit floating point and 32-bit integers. The only aspect where PS is 'better' is in that you can see the changes in real time, giving the false impression of a non-sequential scheme —actually, what happens is that the whole composition is recalculated, perhaps benefiting from partial precalculated results. It is just an advantage in terms of usability, at the user interface level; in all other respects PI is largely superior.

This will change when ProcessContainer gets real-time preview functionality. That will happen somewhere between PixInsight 1.6.2 and 1.6.5. ProcessContainer will be able to capture RTP, provided that all contained instances can also be previewed in real-time. This currently includes Curves, Histogram, ColorSaturation, UnsharpMask, ATrousWaveletTransform, DDP, SCNR, ACDNR and other processes, and the list will be extended.

Anyway, in my opinion getting curves, histogram and color saturation changes in real time through layers is only of marginal importance for astrophotography. If you are processing your images seriously, then you carry out some critical procedures when the image is linear in PixInsight. You then perform an initial nonlinear stretch which conditions the rest of your processing. You cannot just change the initial stretch and wait to see what happens after a HDRWT that has been applied three steps later; that simply doesn't work at all. We have tools that are much more powerful than the typical set of masked layered curves transformations that are common in other applications. In general, curves and color saturation are final tweaks that, IMO, don't benefit from a layered composition scheme in astrophotography. A layered composition tool is nice and it will be available in PixInsight. As I see it, however, its main purpose will be implementing a set of bitmap and vector-based drawing and painting tools.
Juan Conejero
PixInsight Development Team
http://pixinsight.com/

Offline Jack Harvey

  • PTeam Member
  • PixInsight Padawan
  • ****
  • Posts: 975
    • PegasusAstronomy.com & Starshadows.com
Re: I'll be speaking at MWAIC '10
« Reply #22 on: 2010 July 25 08:18:21 »
This is why cloning the entire image is nice.  I will do all my linear stuff and then do my first histogram stretch.  Then make a clone of the stretched image and continue on with more hits, HDR or whatever.  If in 3-4 steps I am happy then the cloned image becomes the iamge I will work with and I make a second clone and proceed.  If 3-4 steps later I am unhappy I just go back to the most recent clone and start again.
Jack Harvey, PTeam Member
Team Leader, SSRO/PROMPT Imaging Team, CTIO

Offline sreilly

  • PixInsight Padawan
  • ****
  • Posts: 791
    • Imaging at Dogwood Ridge Observatory
Re: I'll be speaking at MWAIC '10
« Reply #23 on: 2010 July 25 11:21:19 »
I'm sorry to say that I missed the forum but that said, I have to say that having used both PI and Photoshop, the instructional DVDs and books are out there and have been on how to use PS on images. At this point it is a huge benefit to PS just having so much by well known imagers. An in depth processing DVD series on PI would certainly generate a lot of interests from those such as myself who see the potential benefit to processing with PI but are way to new to fully understand how. Adam Block with his DVD series as well as the newer series by Tony Hallas, which I have to admit I learned a great deal from using PS, more so than Adam's. And books written on just that subject. It's well entrenched and talked about at every astro imaging venue.

A well planed and executed DVD series using PI would go a very long way in showing potential newcomers how to take advantage of the extreme control PI gives you. This takes nothing away from those videos posted by Harry and others but it puts it into a more fluid, comprehensive format that can go into much more detail. I have to admit that on new images, I almost always look at processing the way I use to first using PS to see what the image is shaping up to. After I get that done, I'll go into PI and try what I've learned there but then again, I know more about PS that PI at this point so I'm not seeing everything possible.

Out of the total number of people who have bought PI, how many are actually comfortable with feeling like they are getting the full potential from the software? Are they using it totally for processing or are they overwhelmed by all the processes and its many settings? This is by no means an attack on PI. I own the software and look forward to its many abilities but am intimidated by the program as well. For me, at least, it takes a great deal of time experimenting with these processes and then trying to figure out what order to use the processes. Again, this is where a DVD series would certainly help. I have to think that the more people processing images with PI and presenting them on the web, the more attention would be generated toward PI itself. I have seen, if only roughly, the power of HDRWaveletTansform in bringing out great detail in galaxy cores, even when all seemed blown out in the image. Most other images of these galaxies are missing this detail. I've also seen where I get the core detail but also get artifacts in other parts of the image I don't want. I have no idea how to prevent this in PI but I do know that I can save the image and use a layer mask to get just the core detail in PS. So that's what I do for now. I would much rather take the image all the way in PI but lack the understanding at this point. As there is no help file for PI, I would think a DVD series of instruction would be most welcoming. As new features are added, additional segments can be generated. I don't use PS for much of anything else. I'd just soon get a complete package I know how to use.

Steve
Steve
www.astral-imaging.com
AP1200
OGS 12.5" RC
Tak FSQ-106ED
ST10XME/CFW8/AO8
STL-11000M/FW8/AO-L
Pyxis 3" Rotator
Baader LRGBHa Filters
PixInsight/MaxIm/ACP/Registar/Mira AP/PS CS5

Offline Nocturnal

  • PixInsight Jedi Council Member
  • *******
  • Posts: 2727
    • http://www.carpephoton.com
Re: I'll be speaking at MWAIC '10
« Reply #24 on: 2010 July 25 17:29:01 »
Hi,

I think this is a good discussion but the wrong thread :) Perhaps we should start a new one and discuss further?

Best,

    Sander
---
Edge HD 1100
QHY-8 for imaging, IMG0H mono for guiding, video cameras for occulations
ASI224, QHY5L-IIc
HyperStar3
WO-M110ED+FR-III/TRF-2008
Takahashi EM-400
PIxInsight, DeepSkyStacker, PHD, Nebulosity

Offline Nocturnal

  • PixInsight Jedi Council Member
  • *******
  • Posts: 2727
    • http://www.carpephoton.com
Re: I'll be speaking at MWAIC '10
« Reply #25 on: 2010 July 27 07:46:15 »
Not sure why my dad (he flew in from Holland to attend) sent this specific picture to me but here it is:

http://picasaweb.google.com/lh/photo/cHjleICbR_tZ91QqlsWAhg?feat=directlink
Best,

    Sander
---
Edge HD 1100
QHY-8 for imaging, IMG0H mono for guiding, video cameras for occulations
ASI224, QHY5L-IIc
HyperStar3
WO-M110ED+FR-III/TRF-2008
Takahashi EM-400
PIxInsight, DeepSkyStacker, PHD, Nebulosity

Offline Carlos Milovic

  • PTeam Member
  • PixInsight Jedi Master
  • ******
  • Posts: 2172
  • Join the dark side... we have cookies
    • http://www.astrophoto.cl
Re: I'll be speaking at MWAIC '10
« Reply #26 on: 2010 July 27 08:01:14 »
LOL :D
Regards,

Carlos Milovic F.
--------------------------------
PixInsight Project Developer
http://www.pixinsight.com

Offline RBA

  • PixInsight Guru
  • ****
  • Posts: 511
    • DeepSkyColors
Re: I'll be speaking at MWAIC '10
« Reply #27 on: 2010 July 27 08:27:07 »
Ok, now I must have a copy of all those slides... Pronto!!!

So PI's reputation is that of a hungry extincted beast?  ;)


Offline Niall Saunders

  • PTeam Member
  • PixInsight Jedi Knight
  • *****
  • Posts: 1456
  • We have cookies? Where ?
Re: I'll be speaking at MWAIC '10
« Reply #28 on: 2010 July 27 09:33:20 »
I would recognise that picture anywhere. It is Juan, after the 23rd "Program Compilation Error - unknown or undeclared Variable" message, at 4am, on the 'original' day (i.e. 'UTC' time, not 'JCPITC') that he planned to release the 'next' version.

Fortunately the 'beast' is easily placated - just keep telling him that you are in no rush, and that the next version will be happily received no matter when it is released, and that we don't REALLY need full documentation - not just now anyway.

Whatever you do, do NOT try and poke the beast with sharpened sticks. Truly a foolish thing to do.

Cheers,
Cheers,
Niall Saunders
Clinterty Observatories
Aberdeen, UK

Altair Astro GSO 10" f/8 Ritchey Chrétien CF OTA on EQ8 mount with homebrew 3D Balance and Pier
Moonfish ED80 APO & Celestron Omni XLT 120
QHY10 CCD & QHY5L-II Colour
9mm TS-OAG and Meade DSI-IIC

Offline Nocturnal

  • PixInsight Jedi Council Member
  • *******
  • Posts: 2727
    • http://www.carpephoton.com
Re: I'll be speaking at MWAIC '10
« Reply #29 on: 2010 July 27 10:07:03 »
When I started looking for monster pics I was surprised there wasn't a whole bunch of outrageous pictures available that I felt were suitable. This one was clearly inspired by T-rex but it has some kind of wing on it so it's more dragon like.  I actually had a 'roar' sound effect but that didn't work.

The next slide showed that PI was really a big ol' toolbox :)

Rogelio: Al will have to decide if the presentation may be distributed. I think it's his IP so I can't post them yet. Of course I'm able to re-do this presentation elsewhere so I'm a bit confused on ownership.

In any case, the beef of the preso was in the demos rather than the slides. I hope Al makes the recording available for viewing. Well, maybe I don't :) I'd rather not hear after the fact that I did it all wrong :)
Best,

    Sander
---
Edge HD 1100
QHY-8 for imaging, IMG0H mono for guiding, video cameras for occulations
ASI224, QHY5L-IIc
HyperStar3
WO-M110ED+FR-III/TRF-2008
Takahashi EM-400
PIxInsight, DeepSkyStacker, PHD, Nebulosity