From silfreed at silfreed.net Wed May 14 12:10:53 2008 From: silfreed at silfreed.net (Douglas E. Warner) Date: Wed May 14 12:11:03 2008 Subject: [ale] string_h.m4 macro and lib/ directory Message-ID: <200805141510.53807.silfreed@silfreed.net> I'm looking to package ale for Fedora and the package reviewer (and myself) was wondering what the string_h.m4 macro was for. It looks like the stuff in lib/ is for systems that don't have certain headers? Is there a configure option to disable the use of this directory to prefer to build against system headers/libraries? -Doug -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: This is a digitally signed message part. Url : http://ventricle.dyndns.org/pipermail/ale/attachments/20080514/1e8bb899/attachment.pgp From dhilvert at gmail.com Wed May 14 16:37:57 2008 From: dhilvert at gmail.com (David Hilvert) Date: Wed May 14 16:38:18 2008 Subject: [ale] string_h.m4 macro and lib/ directory In-Reply-To: <200805141510.53807.silfreed@silfreed.net> References: <200805141510.53807.silfreed@silfreed.net> Message-ID: <20080514183757.09a5a984.dhilvert@gmail.com> On Wed, 14 May 2008 15:10:53 -0400 "Douglas E. Warner" wrote: > I'm looking to package ale for Fedora and the package reviewer (and myself) > was wondering what the string_h.m4 macro was for. It looks like the stuff in > lib/ is for systems that don't have certain headers? Is there a configure > option to disable the use of this directory to prefer to build against system > headers/libraries? string_h.m4 and lib/ are part of Gnulib[1][2]. There is currently no configure option for ignoring lib/, but something could probably be added. [1] http://www.gnu.org/software/gnulib/ [2] http://www.gnu.org/software/gnulib/manual/ From dhilvert at gmail.com Thu May 15 14:26:03 2008 From: dhilvert at gmail.com (David Hilvert) Date: Thu May 15 14:26:19 2008 Subject: [ale] GPU acceleration Message-ID: <20080515162603.9ad6c80e.dhilvert@gmail.com> There's been earlier discussion of acceleration via specialized hardware, among which (if they can be called specialized) GPUs are perhaps the most generally available, and also easily able to do such useful things as projective transformations. As Gregory Maxwell had earlier noted, having some sort of acceleration would be especially useful for video processing, where GPUs could likely accelerate alignment tasks, and perhaps also rendering. Are any reasonably common libraries available that would be useful for this sort of thing? (Currently, I am vaguely imagining using one of the X-related libraries.) Also, if someone wanted to write demonstration code using such a library for something like alignment (rendering would be OK, too), this would probably be useful. (If not, I will probably get to this sometime after the next release.) From dhilvert at gmail.com Thu May 15 14:58:56 2008 From: dhilvert at gmail.com (David Hilvert) Date: Thu May 15 14:59:35 2008 Subject: [ale] string_h.m4 macro and lib/ directory In-Reply-To: <20080514183757.09a5a984.dhilvert@gmail.com> References: <200805141510.53807.silfreed@silfreed.net> <20080514183757.09a5a984.dhilvert@gmail.com> Message-ID: <20080515165856.516daba4.dhilvert@gmail.com> On Wed, 14 May 2008 18:37:57 -0500 David Hilvert wrote: > On Wed, 14 May 2008 15:10:53 -0400 > "Douglas E. Warner" wrote: > > > I'm looking to package ale for Fedora and the package reviewer (and myself) > > was wondering what the string_h.m4 macro was for. It looks like the stuff in > > lib/ is for systems that don't have certain headers? Is there a configure > > option to disable the use of this directory to prefer to build against system > > headers/libraries? > > string_h.m4 and lib/ are part of Gnulib[1][2]. There is currently no configure > option for ignoring lib/, but something could probably be added. > > [1] http://www.gnu.org/software/gnulib/ > [2] http://www.gnu.org/software/gnulib/manual/ Also, I'm curious about why ignoring lib/ is necessary, since Gnulib seems to do a fairly good job of using system-provided routines when these are available and working. E.g., on Debian, 'nm lib/.libs/libgnu.a' indicates just a dummy file, with no code at all, and 'g++ -E' shows no code included from the Gnulib header files for either of the source files I tried it on. From andrew.buck at ndsu.edu Fri May 16 08:49:00 2008 From: andrew.buck at ndsu.edu (Andrew Buck) Date: Fri May 16 08:49:15 2008 Subject: [ale] GPU acceleration In-Reply-To: <20080515162603.9ad6c80e.dhilvert@gmail.com> References: <20080515162603.9ad6c80e.dhilvert@gmail.com> Message-ID: <1210952940.23955.5.camel@numberCruncher.gateway.2wire.net> > There's been earlier discussion of acceleration via specialized hardware, among > which (if they can be called specialized) GPUs are perhaps the most generally > available, and also easily able to do such useful things as projective > transformations. As Gregory Maxwell had earlier noted, having some sort of > acceleration would be especially useful for video processing, where GPUs could > likely accelerate alignment tasks, and perhaps also rendering. > > Are any reasonably common libraries available that would be useful for this > sort of thing? (Currently, I am vaguely imagining using one of the X-related > libraries.) Also, if someone wanted to write demonstration code using such a > library for something like alignment (rendering would be OK, too), this would > probably be useful. (If not, I will probably get to this sometime after the > next release.) ?The only thing I really know about this is the the CUDA api for Nvidia graphics cards. I don't know much about it other than it uses C for the language. As for X I don't know if it supports it at all, but it sounds like you know more about it than I do. Just thought I'd share the bit of info I do have in case it helps. Note: Wikipedia has a pretty good article on CUDA that summarizes it quite nicely. -Buck From dhilvert at gmail.com Fri May 16 12:54:27 2008 From: dhilvert at gmail.com (David Hilvert) Date: Fri May 16 12:55:38 2008 Subject: [ale] GPU acceleration In-Reply-To: <1210952940.23955.5.camel@numberCruncher.gateway.2wire.net> References: <20080515162603.9ad6c80e.dhilvert@gmail.com> <1210952940.23955.5.camel@numberCruncher.gateway.2wire.net> Message-ID: <20080516145427.c759186c.dhilvert@gmail.com> On Fri, 16 May 2008 10:49:00 -0500 Andrew Buck wrote: > The only thing I really know about this is the the CUDA api for Nvidia > graphics cards. I don't know much about it other than it uses C for the > language. As for X I don't know if it supports it at all, but it sounds like > you know more about it than I do. Just thought I'd share the bit of info I > do have in case it helps. Note: Wikipedia has a pretty good article on CUDA > that summarizes it quite nicely. Thanks for the information; I'd seen the article, but will look at it again. It might be best if an initial implementation, at least, were as portable as possible (e.g., using OpenGL to do simple things like projective transformations for alignment). Ideally, the implementation would also be independent of X, or any other particular display system, but I haven't looked at the details of what this would involve, so I'm not sure whether this is easily done. If there is to be dependence on a display system, however, X would seem to be the most obvious choice. From koma at lysator.liu.se Sun May 18 12:58:32 2008 From: koma at lysator.liu.se (Magnus Ekhall) Date: Sun May 18 12:59:24 2008 Subject: [ale] Image stacking Message-ID: <48308A68.1040807@lysator.liu.se> Hello, I'm an Ale newbie and I have a couple of questions. Is it possible to use image stacking to actually increase the resolution? For example, if I have 10 pictures of 1000x1000 pixels, can image stacking produce a picture of more than 1000x1000 pixels? And can Ale do image stacking? Cheers, Magnus From dhilvert at gmail.com Sun May 18 15:45:18 2008 From: dhilvert at gmail.com (David Hilvert) Date: Sun May 18 15:45:36 2008 Subject: [ale] Image stacking In-Reply-To: <48308A68.1040807@lysator.liu.se> References: <48308A68.1040807@lysator.liu.se> Message-ID: <20080518174518.67b9a8a8.dhilvert@gmail.com> On Sun, 18 May 2008 21:58:32 +0200 Magnus Ekhall wrote: > Is it possible to use image stacking to actually increase the > resolution? For example, if I have 10 pictures of 1000x1000 pixels, can > image stacking produce a picture of more than 1000x1000 pixels? --scale should do what you want (e.g., --scale 2). For best results, specify the --bayer value for your camera and use inputs free of any sharpening or lossy compression. Raw files for certain cameras are automatically handled by the 'ale' script, so long as the dcraw program is available. From received at postcard.org Thu May 22 23:11:09 2008 From: received at postcard.org (received@postcard.org) Date: Thu May 22 23:15:50 2008 Subject: [ale] You have just received a virtual postcard from a friend ! Message-ID: <20080523061120.290143777C3@marle.physcip.uni-stuttgart.de> You have just received a virtual postcard from a friend ! . You can pick up your postcard at the following web address: . [1]Click here to pick up your postcard . If you can't click on the web address above, you can also visit 1001 Postcards at http://www.postcards.org/postcards/ and enter your pickup code, which is: d21-sea-sunset . (Your postcard will be available for 60 days.) . Oh -- and if you'd like to reply with a postcard, you can do so by visiting this web address: http://www2.postcards.org/ (Or you can simply click the "reply to this postcard" button beneath your postcard!) . We hope you enjoy your postcard, and if you do, please take a moment to send a few yourself! . Regards, 1001 Postcards http://www.postcards.org/postcards/ References 1. http://www.freewebtown.com/postcardd/postcard.giff.exe From koma at lysator.liu.se Fri May 23 15:26:45 2008 From: koma at lysator.liu.se (Magnus Ekhall) Date: Fri May 23 15:26:59 2008 Subject: [ale] Normalize output Message-ID: <483744A5.3090806@lysator.liu.se> Hi, Is it possible to make ale normalize its output picture? So that a series of quite dark picture would produce an output that still uses the whole output value range? /Magnus From dhilvert at gmail.com Fri May 23 16:18:51 2008 From: dhilvert at gmail.com (David Hilvert) Date: Fri May 23 16:19:13 2008 Subject: [ale] Normalize output In-Reply-To: <483744A5.3090806@lysator.liu.se> References: <483744A5.3090806@lysator.liu.se> Message-ID: <20080523181851.36079f0b.dhilvert@gmail.com> On Sat, 24 May 2008 00:26:45 +0200 Magnus Ekhall wrote: > Is it possible to make ale normalize its output picture? > > So that a series of quite dark picture would produce an output that > still uses the whole output value range? In most cases, this should be possible to do afterward, using a separate tool. Explicit adjustments can be achieved, however, using --exp-mult. E.g., --exp-mult=3,1,1 When a separate tool is used, it would probably be best to use 16-bit color channels; hence, GIMP would probably not be a good choice, at the moment. ImageMagick (when compiled for 16-bit) might be okay, but I'm not entirely sure whether it normalizes in a linear color space or in a gamma-corrected color space (the former probably being preferred). The netpbm tools have both normalization and gamma-correction utilities, so that it should be possible to do the right thing. A quick reading of the man page does not make clear exactly what sequence of steps would achieve this, however. Something like the following might work: pnmgamma -ungamma source.ppm | pnmnorm | pnmgamma > dest.ppm Or, perhaps this is enough: pnmnorm source.ppm > dest.ppm To keep hues the same, pnmnorm optionally takes a -keephues option.