On Sun, Aug 15, 2004 at 08:18:14AM -0700, William Skaggs wrote:

Does Photoshop do it a lot faster with the same parameters?

Very much so. I created a panorama earlier this evening and ran it through the workflow action I'm porting to GIMP and the haze removal step took only a few seconds, and this on a 275MB image.

On Sun, Aug 15, 2004 at 08:57:46PM +0200, Steve Crane wrote:

On Sun, Aug 15, 2004 at 08:18:14AM -0700, William Skaggs wrote:

Does Photoshop do it a lot faster with the same parameters?

Very much so. I created a panorama earlier this evening and ran it through the workflow action I'm porting to GIMP and the haze removal step took only a few seconds, and this on a 275MB image.

what did you use for haze removal?

when i use value levels to remove haze, it works very very fast on my gimp.

carol

On Sun, Aug 15, 2004 at 01:05:29PM -0700, Carol Spears wrote:

what did you use for haze removal?

USM with radius 60, amount 0.30 and threshold 1.

On Sun, Aug 15, 2004 at 10:33:33PM +0200, Steve Crane wrote:

On Sun, Aug 15, 2004 at 01:05:29PM -0700, Carol Spears wrote:

what did you use for haze removal?

USM with radius 60, amount 0.30 and threshold 1.

without having the chance (or a photo handy) to check this with, are you certain that unsharp mask is better than using levels?

carol

On Sun, Aug 15, 2004 at 03:00:01PM -0700, carol wrote:

On Sun, Aug 15, 2004 at 10:33:33PM +0200, Steve Crane wrote:

On Sun, Aug 15, 2004 at 01:05:29PM -0700, Carol Spears wrote:

what did you use for haze removal?

USM with radius 60, amount 0.30 and threshold 1.

without having the chance (or a photo handy) to check this with, are you certain that unsharp mask is better than using levels?

the method i always used is to move the pointers on the dialog to where the colors start on the values histogram (black and white sides) then a tiny (was supposed to be 10% of the greater movement of the ends) adjustment to the middle triangle, to adjust the contrast some.

i have never seen unsharp mask used for haze reduction.

carol

On Sun, Aug 15, 2004 at 03:00:01PM -0700, Carol Spears wrote:

USM with radius 60, amount 0.30 and threshold 1.

without having the chance (or a photo handy) to check this with, are you certain that unsharp mask is better than using levels?

As I explained in my original post I am in the progress of converting a Photoshop action that performs a workflow to process photos from specific cameras. This is just one of the steps it uses. As the step is quick in Photoshop and very slow in GIMP, I merely wondered if this was to be expected or the symptom of a problem. One of the other replies suggested that the large radius could be expected to make the USM filter slow as the algorithm had not been revised for some years and was perhaps not working as well as it could.

Hi,

Steve Crane writes:

On Sun, Aug 15, 2004 at 03:00:01PM -0700, Carol Spears wrote:

USM with radius 60, amount 0.30 and threshold 1.

without having the chance (or a photo handy) to check this with, are you certain that unsharp mask is better than using levels?

As I explained in my original post I am in the progress of converting a Photoshop action that performs a workflow to process photos from specific cameras. This is just one of the steps it uses. As the step is quick in Photoshop and very slow in GIMP, I merely wondered if this was to be expected or the symptom of a problem. One of the other replies suggested that the large radius could be expected to make the USM filter slow as the algorithm had not been revised for some years and was perhaps not working as well as it could.

Steve, please don't be so ignorant. Carol has a point here. If unsharp mask is slow, it makes sense to look for alternatives. There's no point in sticking to your workflow if it turns out that the same result can be better achieved differently. So, are you certain that unsharp mask is better than using levels?

Sven

sven@gimp.org (2004-08-17 at 1052.52 +0200):

Steve, please don't be so ignorant. Carol has a point here. If unsharp mask is slow, it makes sense to look for alternatives. There's no point in sticking to your workflow if it turns out that the same result can be better achieved differently. So, are you certain that unsharp mask is better than using levels?

I tried levels 20 1.0 235 and USM 25 .3 1 in the first image of http://www.lonestardigital.com/photoshop_quicktips.htm and while the results look similar, the histogram shows periodic holes (predictable from levels), so problably that is the reason the action Steve is porting used USM. Maybe he should give a look at USM code and try to emulate with blur and other ops.

GSR

On Tue, Aug 17, 2004 at 10:52:52AM +0200, Sven Neumann wrote:

Steve, please don't be so ignorant.

This seems a bit rude and uncalled for.

Carol has a point here. If unsharp mask is slow, it makes sense to look for alternatives.

Of course it does.

There's no
point in sticking to your workflow if it turns out that the same result can be better achieved differently.

I don't intend to, nor did I say I planned to.

So, are you certain that
unsharp mask is better than using levels?

No.

The point I was trying to make in my reply to Carol, and maybe I didn't put it across clearly, was the following. In general GIMP (in Linux) performs roughly the same as Photoshop (in Windows) on my machine. So when I found this one filter that is so much slower with the same settings I became curious, wondering if it is to be expected, due to the way GIMP handles USM. That is all. I was not saying I will blindly use the function because the workflow uses it. I was not saying that I won't look for alternatives. I was not saying that GIMP sucks or has a bug, merely asking for an opinion from those more knowledgeable than myself in the workings of the USM filter.

I do appreciate Carol's pointing out an alternative. I sometimes feel though, that on mailing lists in general, a lot of friction could be avoided if we only answer what is asked, not what we think is being asked by trying to read between the lines.

In response to Alan Horkan, it is the plug-in version that I'm using.

Cheers.

On Wed, Aug 18, 2004 at 12:14:15AM +0200, Steve Crane wrote:

On Tue, Aug 17, 2004 at 10:52:52AM +0200, Sven Neumann wrote:

Steve, please don't be so ignorant.

This seems a bit rude and uncalled for.

well, gimp developers are not known for being polite in all situations. pointing out when he is rude is sort of a compliment in some situations. maybe not this one, however.

Carol has a point here. If unsharp mask is slow, it makes sense to look for alternatives.

Of course it does.

it is interesting that the autosharp plug-in is running slower than the photoshop equivelent of the same thing. there is a good chance that one or the other of us is doing the wrong thing or that the gimp code is old, simply ported along and not rewritten.

that would probably best sent in a more terse format to the gimp-developer list. even then, someone needs to be interested in it. mentioning your "workflow" is not the best way to approach people who are trying to think about writing software or people who are volunteering to write software.

There's no
point in sticking to your workflow if it turns out that the same result can be better achieved differently.

I don't intend to, nor did I say I planned to.

we were wondering if value levels (a tool and not a plug-in) worked better than what you are using.

we do not assume that you are using photoshop properly or not. i simply would have accomplished what you described using the levels tool on the image values.

the situation you described sounded like this tool would work better. this list is generally the sort where images are shown that demonstrate the problem. can you show us a photo? original, unsharp maskified and also with the levels tweaking i suggested? there is a very good chance that i am incorrect here.

So, are you certain that
unsharp mask is better than using levels?

No.

The point I was trying to make in my reply to Carol, and maybe I didn't put it across clearly, was the following. In general GIMP (in Linux) performs roughly the same as Photoshop (in Windows) on my machine. So when I found this one filter that is so much slower with the same settings I became curious, wondering if it is to be expected, due to the way GIMP handles USM. That is all. I was not saying I will blindly use the function because the workflow uses it. I was not saying that I won't look for alternatives. I was not saying that GIMP sucks or has a bug, merely asking for an opinion from those more knowledgeable than myself in the workings of the USM filter.

most of the gimp developers use linux. photoshop does not run on linux. half the time we do not know what they are doing! usually when they do have something new, it is just a trick of the modes or something not so clever as you might think.

yes, photoshop and gimp share many of the same stuff. that is because it is all computers. graphic images on computers. i always thought gimp was more like paint shop pro, personally. are you certain it resembles photoshop that much?

yes, please do not blindly use any gimp functions. i am also assuming that you did not use unsharp mask blindly. if you could show us some examples it would make the chat about your work and your workflow more productive.

it is way too late in the game to say gimp sucks or that photoshop sucks. no one even thought this.

I do appreciate Carol's pointing out an alternative. I sometimes feel though, that on mailing lists in general, a lot of friction could be avoided if we only answer what is asked, not what we think is being asked by trying to read between the lines.

ah, you are worried about your workflow. i am also. and about your work as well. please do not read between the lines of what i ask as well. a simple "did you try this" might actually improve your work at the same time it improves your work flow. this list has not ever been about peoples work flow, it has been about the best way to handle images with the gimp.

In response to Alan Horkan, it is the plug-in version that I'm using.

examples and actual version numbers are helpful to everyone.

cheers back, carol