Re: [hugin-ptx] when to sharpen?
Sharpening in Photoshop or GIMP. On Wed, Sep 12, 2012 at 1:42 AM, Bruno Postle br...@postle.net wrote: On Tue 11-Sep-2012 at 13:09 +0200, Rogier Wolff wrote: On Tue, Sep 11, 2012 at 10:38:55AM +0100, Bruno Postle wrote: Artificially 'sharpened' images are a special case, you don't find this kind of data in 'normal' photos, these don't really suffer any loss of focus in the standard remapping used by Hugin. IMHO, when I click optimal size it recommends a size where each source pixel maps to at least one remapped pixel. Bluntly said: If I take three portrait 2500x4000 images and align them next to each other, my optimal size will have a height of 4000 pixels (plus whatever is needed because they don't align perfectly). In this situation, I have the impression I can clearly see that the remapped images are softer, fuzzier than the originals. Are these photos sharpened? You can see this effect by drawing a one pixel line in an image, then remapping it in Hugin, it will become 'fuzzy', but remap it again and it won't get any fuzzier. Real, unsharpened, photos don't have hard edges like this, there is always a transition between two colours - This is nothing to do with the quality of the lens. -- Bruno -- You received this message because you are subscribed to the Google Groups Hugin and other free panoramic software group. A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ To post to this group, send email to hugin-ptx@googlegroups.com To unsubscribe from this group, send email to hugin-ptx+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/hugin-ptx -- *Emaad* www.flickr.com/emaad -- You received this message because you are subscribed to the Google Groups Hugin and other free panoramic software group. A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ To post to this group, send email to hugin-ptx@googlegroups.com To unsubscribe from this group, send email to hugin-ptx+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/hugin-ptx
Re: [hugin-ptx] when to sharpen?
On Tue, Sep 11, 2012 at 09:42:00PM +0100, Bruno Postle wrote: Are these photos sharpened? No, from camera. Not a sony, but a Nikon. You can see this effect by drawing a one pixel line in an image, then remapping it in Hugin, it will become 'fuzzy', but remap it again and it won't get any fuzzier. Real, unsharpened, photos don't have hard edges like this, there is always a transition between two colours - This is nothing to do with the quality of the lens. In theory, when I photograph a vertical divide between a white and a black plane, and the camera is just a few pixels non-horizontal, I should have pixels where the one on the left is entirely focussed on the white plane, and the one on the right is completley focussed on the black plane. In practise, there is lens fuzzyness, but you say that has nothing to do with it. In practise, to combat aliasing, there is a fuzzy-filter in front of the sensor. This makes the bayer pattern work in difficult situations. In this case, it will cause some black to leak to the left pixel and some white to the right pixel. So in this case, TWO pixels end up not being completely white or completely black. Roger. -- ** r.e.wo...@bitwizard.nl ** http://www.BitWizard.nl/ ** +31-15-2600998 ** **Delftechpark 26 2628 XH Delft, The Netherlands. KVK: 27239233** *-- BitWizard writes Linux device drivers for any device you may have! --* The plan was simple, like my brother-in-law Phil. But unlike Phil, this plan just might work. -- You received this message because you are subscribed to the Google Groups Hugin and other free panoramic software group. A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ To post to this group, send email to hugin-ptx@googlegroups.com To unsubscribe from this group, send email to hugin-ptx+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/hugin-ptx
[hugin-ptx] when to sharpen?
This probably an FAQ, but I couldn't find an answer: when do you sharpen spherical panoramas? Before compositing (i.e. sharpen the original images) or afterwards, and if so, using what type of projection (I need to output equirectangular in the end)? -- You received this message because you are subscribed to the Google Groups Hugin and other free panoramic software group. A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ To post to this group, send email to hugin-ptx@googlegroups.com To unsubscribe from this group, send email to hugin-ptx+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/hugin-ptx
Re: [hugin-ptx] when to sharpen?
On 11 Sep 2012 07:54, TvE tvoneic...@gmail.com wrote: This probably an FAQ, but I couldn't find an answer: when do you sharpen spherical panoramas? Before compositing (i.e. sharpen the original images) or afterwards, and if so, using what type of projection (I need to output equirectangular in the end)? Sharpening doesn't survive remapping very well, so you should apply it to your final image at the intended display resolution. So for a cubic panorama it is the individual cubefaces that you sharpen etc... -- Bruno -- You received this message because you are subscribed to the Google Groups Hugin and other free panoramic software group. A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ To post to this group, send email to hugin-ptx@googlegroups.com To unsubscribe from this group, send email to hugin-ptx+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/hugin-ptx
Re: [hugin-ptx] when to sharpen?
On Tue, Sep 11, 2012 at 04:09:50AM -0400, Bruno Postle wrote: Sharpening doesn't survive remapping very well, so you should apply it to Right! It looks to me that the remapping takes the average of two surrounding pixels, on average. i.e. with an image that COULD map 1 to 1, every source pixel will still be smeared out over two destination pixels minimum. (or the other way around: every destination pixel is the average of two source pixels). If my math intuition is good, the -0.5 2 -0.5 convolution is the inverse of 0,1,1,0. Such a convolution might be possible to build into the remapping operation to keep the images as sharp as the original. The core of the problem is that you want to prevent aliasing effects. It is quite possible that 50 source pixels map to 49 destination pixels. This means that at one point (say at 0 and 49 destination pixels) they line up perfectly. While at another (that would be around 25 pixels) each destination pixel is halfway inbetween two source pixels. If you do it the obvious way, you'd just copy pixels 0, 1, 48, 49 to get a sharp image, but near pixels 24 and 25 you'd have to take the average of 24, 25 to get destination pixel 24, and then average 25,26 to get destination pixel 25. This would probably lead to visible artefacts. So there is some smart stuff in there to take a similar average near the point where just copying would be more obvious. I've described things as if they are one-dimensional. Things get a bit more complicated in two dimensions. Roger. -- ** r.e.wo...@bitwizard.nl ** http://www.BitWizard.nl/ ** +31-15-2600998 ** **Delftechpark 26 2628 XH Delft, The Netherlands. KVK: 27239233** *-- BitWizard writes Linux device drivers for any device you may have! --* The plan was simple, like my brother-in-law Phil. But unlike Phil, this plan just might work. -- You received this message because you are subscribed to the Google Groups Hugin and other free panoramic software group. A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ To post to this group, send email to hugin-ptx@googlegroups.com To unsubscribe from this group, send email to hugin-ptx+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/hugin-ptx
Re: [hugin-ptx] when to sharpen?
On 11 September 2012 10:03, Rogier Wolff rew-googlegro...@bitwizard.nl wrote: On Tue, Sep 11, 2012 at 04:09:50AM -0400, Bruno Postle wrote: Sharpening doesn't survive remapping very well, so you should apply it to If my math intuition is good, the -0.5 2 -0.5 convolution is the inverse of 0,1,1,0. Such a convolution might be possible to build into the remapping operation to keep the images as sharp as the original. Artificially 'sharpened' images are a special case, you don't find this kind of data in 'normal' photos, these don't really suffer any loss of focus in the standard remapping used by Hugin. -- Bruno -- You received this message because you are subscribed to the Google Groups Hugin and other free panoramic software group. A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ To post to this group, send email to hugin-ptx@googlegroups.com To unsubscribe from this group, send email to hugin-ptx+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/hugin-ptx
Re: [hugin-ptx] when to sharpen?
On Tue, Sep 11, 2012 at 10:38:55AM +0100, Bruno Postle wrote: On 11 September 2012 10:03, Rogier Wolff rew-googlegro...@bitwizard.nl wrote: On Tue, Sep 11, 2012 at 04:09:50AM -0400, Bruno Postle wrote: Sharpening doesn't survive remapping very well, so you should apply it to If my math intuition is good, the -0.5 2 -0.5 convolution is the inverse of 0,1,1,0. Such a convolution might be possible to build into the remapping operation to keep the images as sharp as the original. Artificially 'sharpened' images are a special case, you don't find this kind of data in 'normal' photos, these don't really suffer any loss of focus in the standard remapping used by Hugin. IMHO, when I click optimal size it recommends a size where each source pixel maps to at least one remapped pixel. Bluntly said: If I take three portrait 2500x4000 images and align them next to each other, my optimal size will have a height of 4000 pixels (plus whatever is needed because they don't align perfectly). In this situation, I have the impression I can clearly see that the remapped images are softer, fuzzier than the originals. Roger. -- ** r.e.wo...@bitwizard.nl ** http://www.BitWizard.nl/ ** +31-15-2600998 ** **Delftechpark 26 2628 XH Delft, The Netherlands. KVK: 27239233** *-- BitWizard writes Linux device drivers for any device you may have! --* The plan was simple, like my brother-in-law Phil. But unlike Phil, this plan just might work. -- You received this message because you are subscribed to the Google Groups Hugin and other free panoramic software group. A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ To post to this group, send email to hugin-ptx@googlegroups.com To unsubscribe from this group, send email to hugin-ptx+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/hugin-ptx
Re: [hugin-ptx] when to sharpen?
With any image processing it is always recommended that sharpening is the last thing you do. Any adjustments after that affect the perception of sharpness that the eye sees. Sent from my BlackBerry® wireless handheld -Original Message- From: Felix Hagemann felix.hagem...@gmail.com Sender: hugin-ptx@googlegroups.com Date: Tue, 11 Sep 2012 14:49:42 To: hugin-ptx@googlegroups.com Reply-To: hugin-ptx@googlegroups.com Subject: Re: [hugin-ptx] when to sharpen? On 11 September 2012 08:54, TvE wrote: This probably an FAQ, but I couldn't find an answer: when do you sharpen spherical panoramas? Before compositing (i.e. sharpen the original images) or afterwards, and if so, using what type of projection (I need to output equirectangular in the end)? I've been experimenting with the projection to do the sharpening on quite some time ago. I was mainly trying to sort out if there is a practical difference between: (i) Sharpening the final equirectangular. In theory this should be a bad idea due to the messed up neighbourhoods near the poles. (ii) Create six rectilinear 90x90 images, sharpen those and reassemble to an equirectangular. While the difference images showed some very minor differences I was unable to distinguish the images created by those two methods visually. I've been sharpening equirects ever since... Felix -- You received this message because you are subscribed to the Google Groups Hugin and other free panoramic software group. A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ To post to this group, send email to hugin-ptx@googlegroups.com To unsubscribe from this group, send email to hugin-ptx+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/hugin-ptx -- You received this message because you are subscribed to the Google Groups Hugin and other free panoramic software group. A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ To post to this group, send email to hugin-ptx@googlegroups.com To unsubscribe from this group, send email to hugin-ptx+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/hugin-ptx
Re: [hugin-ptx] when to sharpen?
On Tue 11-Sep-2012 at 13:09 +0200, Rogier Wolff wrote: On Tue, Sep 11, 2012 at 10:38:55AM +0100, Bruno Postle wrote: Artificially 'sharpened' images are a special case, you don't find this kind of data in 'normal' photos, these don't really suffer any loss of focus in the standard remapping used by Hugin. IMHO, when I click optimal size it recommends a size where each source pixel maps to at least one remapped pixel. Bluntly said: If I take three portrait 2500x4000 images and align them next to each other, my optimal size will have a height of 4000 pixels (plus whatever is needed because they don't align perfectly). In this situation, I have the impression I can clearly see that the remapped images are softer, fuzzier than the originals. Are these photos sharpened? You can see this effect by drawing a one pixel line in an image, then remapping it in Hugin, it will become 'fuzzy', but remap it again and it won't get any fuzzier. Real, unsharpened, photos don't have hard edges like this, there is always a transition between two colours - This is nothing to do with the quality of the lens. -- Bruno -- You received this message because you are subscribed to the Google Groups Hugin and other free panoramic software group. A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ To post to this group, send email to hugin-ptx@googlegroups.com To unsubscribe from this group, send email to hugin-ptx+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/hugin-ptx