Hi Carusoswi
On 09 Aug 09 22:37 "Carusoswi" said:

Seems if I copy the background layer leaving the mode normal, I can then perform most any operation on that new level and give it a name suggestive of that operation. Then, make a copy of that new level, and perform some other operation on the new level, rename it to suggest that second operation, and so on. Is that how it works? Seems to give me a result that I can follow up and down the stack by turning on and off the visibility of the levels in sequence (or out of sequence, for that matter.

I feel like I'm on the right track. Would appreciate verification and/or additional advice.

Of course it depends on the kind of image on which you are working and you don't tell us that. Nor do you tell us what you are doing on the levels dialogue, only how you are protecting yourself from "mistakes".

I'm going to assume you are working with a photograph. If that's the case, then this page might suggest some of the reasons you should use Levels:
http://www.gregtutor.plus.com/digphoto/levels.html

Greg Chapman http://www.gregtutor.plus.com
Helping new users of KompoZer and The GIMP

< big snip >

The .xcf file archives my progressive work on the photo so that I can go back and refine my work/revise my edit decisions, etc. The TIFF gives me a full resolution "final" product. The RAW file archives the original image as it came out of the camera.

This business with the layers is new to me as of today. I'd never worked with layers in this manner. In fact, I had never worked with layers much at all until today.

So, my question: Does the above make sense with regard to layers, or is there a better way to work with layers within a photo than what I describe? Nothing seemed to work for me before because I couldn't get much to show in terms of adjustments in layers added beyond that initial duplication of the background layer.

The short answer is yes, this is how I would do things. The only difference is that when I am satisfied with a stage I would merge the layers otherwise the file becomes very large. Also, from time to time I would save my work and keep watching 'meetthegimp'

Norman

Hi Robert,

On 10 Aug 09 07:29 Norman Silverstone said:

The short answer is yes, this is how I would do things. The only difference is that when I am satisfied with a stage I would merge the layers otherwise the file becomes very large. Also, from time to time I would save my work and keep watching 'meetthegimp'

Norman stole my words! :-)

Greg Chapman http://www.gregtutor.plus.com
Helping new users of KompoZer and The GIMP

I've learned the hard way to SAVE OFTEN!!! Or you can lose your work when GIMP freezes and dies.

Greg Chapman wrote:

Hi Robert,

On 10 Aug 09 07:29 Norman Silverstone said:

The short answer is yes, this is how I would do things. The only difference is that when I am satisfied with a stage I would merge the layers otherwise the file becomes very large. Also, from time to time I would save my work and keep watching 'meetthegimp'

Norman stole my words! :-)

Greg Chapman http://www.gregtutor.plus.com
Helping new users of KompoZer and The GIMP

I've learned the hard way to SAVE OFTEN!!! Or you can lose your work when GIMP freezes and dies.

I don't suffer from Gimp freezing or dying but I have been known to press the wrong button and loose an hours careful cloning.

Norman

Carusoswi wrote:

Ok, but, when you merge, do you not lose the ability to go back and fix something that you might decide needs adjusting?

Yes, which is why you don't merge until. you're absolutely sure that everything you're merging is to your satisfaction. And if there are a couple of layers which you're not quite certain about yet, one can merge the layers below those, and still reduce the file size. I haven't explored this yet, and may be wrong, but it might just be that if there is one small area that needs a fix, that one can make a layer just big enough to manage the fix, rather than making the layer the full size of the image. One thing that might reduce the size of the file a bit, is that if there is only a small bit of something that needs fixing, to make

I would thing that layers in an xcf file would only represent references to adjustments and the underlying file

I think a better visualization of layers is to consider them like an overlay on a projector, and that the layer containing the change is independent of the layer to which the change relates, until the two are merged together.

I'm just a-wonderin' why the xcf files grow so large.

I suspect that becuase you have a number of layers all the same size as the image.

ns

Ok, but, when you merge, do you not lose the ability to go back and fix something that you might decide needs adjusting?

Yes, which is why you don't merge until. you're absolutely sure that everything you're merging is to your satisfaction. And if there are a couple of layers which you're not quite certain about yet, one can merge the layers below those, and still reduce the file size. I haven't explored this yet, and may be wrong, but it might just be that if there is one small area that needs a fix, that one can make a layer just big enough to manage the fix, rather than making the layer the full size of the image. One thing that might reduce the size of the file a bit, is that if there is only a small bit of something that needs fixing, to make

I would thing that layers in an xcf file would only represent references to adjustments and the underlying file

I think a better visualization of layers is to consider them like an overlay on a projector, and that the layer containing the change is independent of the layer to which the change relates, until the two are merged together.

I'm just a-wonderin' why the xcf files grow so large.

I suspect that becuase you have a number of layers all the same size as the image.

Also the size of the file will grow even larger when you start to use layer masks.

Norman

I really appreciate the replies. In the work flow example I gave earlier in

this thread, am I correct that there is no other practical way to accomplish those steps on a photo other than to perform them destructively on a single layer (set levels, hit ok, correct color, hit ok, etc. so that you can't go back without undoing in sequence or just scrapping your work and starting over from scratch)? And flattening or merging layers is really sort of a 'kicked down the road' way of 'hitting OK' albeit you are committing to a batch of changes rather than accepting them one at a time - better, I guess, in that you get to see the net effect of all the layers in the stack before committing to their effect.

So, if I'm correct, then, I finally think I've grasped how to use layers in editing a photo.

I am no expert but I think you are nearly there. One thing to bear in mind is that, with a tool like levels, every time you use it and press OK you will lose some detail. So it will pay not to accept your changes until you are really satisfied with the result. If you need to go back then it will probably be better to redo that layer from fresh. In other cases with a layer stack you can, of course, turn layers on and off with the little icon on the left and so work on any layer you wish.

Norman

It seems to me that gimp works a bit differently from photoshop. In photoshop (actually, I'm still using 'elements-2', can you believe? Although I have got Richard Lynch's excellent "hidden power" installed which releases a lot more of the underlying photoshop 7 functionality - and the book was great for learning about image manipulation in general - but I digress...) In photoshop you can add an adjustment layer, which operates on the layer underneath it. And you can go on adding new adjustment layers, and then going back to earlier ones to "tweak" the parameters - so I might have a base image with a brightness/contrast adjustment layer above it and a hue/saturation layer above that. Both adjustment layers operate on the base layer, I can go back to the middle (in my eg, the brightness/contrast), tweak the parameters and view the result - with the effects of the upper hue/sat layer still applied.  
Correct me if I'm wrong, please, but it seems that in gimp I have to make a copy of the base layer and apply any adjustments to the copy; and repeat this for any new adjustment. This seems to be much less flexible, as subsequent changes to the middle layer would be obscured by the upper layer?  
The photoshop method appears to be far more flexible. I was thinking that doing things this way might also have a beneficial effect on the file size, though judging by the size of photoshop format files, I doubt this is actually the case. But I tend to save as layered tiff with zip compression applied to the layers, which makes them much smaller and preserves much of the layer information (though things like selections won't be saved. But I can live with that).  
These, I guess, are my main reasons for hanging on to photoshop rather than migrating to gimp. So if I'm wrong, I'd love to know...  
/Gary

Ok, perhaps I'm on the right track, now.  Someone tell me if I'm moving in the right direction.  Seems if I copy the background layer leaving the mode normal, I can then perform most any operation on that new level and give it a name suggestive of that operation.  Then, make a copy of that new level, and perform some other operation on the new level, rename it to suggest that second operation, and so on.  Is that how it works?  Seems to give me a result that I can follow up and down the stack by turning on and off the visibility of the levels in sequence (or out of sequence, for that matter.

I feel like I'm on the right track.  Would appreciate verification and/or additional advice.

On Wed, Aug 12, 2009 at 9:36 PM, Gary Collins wrote:

It seems to me that gimp works a bit differently from photoshop. In photoshop (actually, I'm still using 'elements-2', can you believe? Although I have got Richard Lynch's excellent "hidden power" installed which releases a lot more of the underlying photoshop 7 functionality - and the book was great for learning about image manipulation in general - but I digress...) In photoshop you can add an adjustment layer, which operates on the layer underneath it. And you can go on adding new adjustment layers, and then going back to earlier ones to "tweak" the parameters - so I might have a base image with a brightness/contrast adjustment layer above it and a hue/saturation layer above that. Both adjustment layers operate on the base layer, I can go back to the middle (in my eg, the brightness/contrast), tweak the parameters and view the result - with the effects of the upper hue/sat layer still applied.

Correct me if I'm wrong, please, but it seems that in gimp I have to make a copy of the base layer and apply any adjustments to the copy; and repeat this for any new adjustment. This seems to be much less flexible, as subsequent changes to the middle layer would be obscured by the upper layer?

This is certainly true; all of your points are true. This is being worked on.
However, the specific idea of 'effect layers' is regarded as severely broken (basically cause it makes nonsense of the whole layers concept: all layers have content, but oh! effect layers don't. all layers have blending mode, but oops! effect layers don't. it's user-unfriendly in this marked inconsistency.)
The implementation I believe we are currently aiming for is instead oriented around the idea of being able to attach any number of effects to a given layer group (btw, martin nordholts is doing some great work on layer trees presently and in the last few months.. they are shaping up well.)

The photoshop method appears to be far more flexible. I was thinking that doing things this way might also have a beneficial effect on the file size, though judging by the size of photoshop format files, I doubt this is actually the case.

Photoshop format generally saves a lot of cached data -- for instance, there is a thumbnail for each layer, and a composited version of the image rendered at full size.

But I tend to save as layered tiff with zip compression applied to the layers, which makes them much smaller and preserves much of the layer information (though things like selections won't be saved. But I can live with that).

I hope you don't change GIMP too much. I'm just getting used to the way it works now.

I've never used PS so I don't care how PS does things.

David Gowers wrote:

This is certainly true; all of your points are true. This is being worked on.
However, the specific idea of 'effect layers' is regarded as severely broken (basically cause it makes nonsense of the whole layers concept: all layers have content, but oh! effect layers don't. all layers have blending mode, but oops! effect layers don't. it's user-unfriendly in this marked inconsistency.)
The implementation I believe we are currently aiming for is instead oriented around the idea of being able to attach any number of effects to a given layer group (btw, martin nordholts is doing some great work on layer trees presently and in the last few months.. they are shaping up well.)

On 08/12/2009 04:21 PM, David Gowers wrote: btw, martin nordholts is doing some great work on layer trees presently and in the last few months.. they are shaping up well.

You are confusing me with another M.N.

It is Michael Natterer that has been doing great work recently on preparing for layer groups

/ Martin

On Thu, Aug 13, 2009 at 5:19 AM, Martin Nordholts wrote:

On 08/12/2009 04:21 PM, David Gowers wrote: btw, martin nordholts is doing some great work on layer trees presently and in the last few months.. they are shaping up well.

You are confusing me with another M.N.

It is Michael Natterer that has been doing great work recently on preparing for layer groups

Oops! Sorry, Martin and Michael :)

David

Thanks for the info. It seems exciting things are on the horizon. I eagerly await with much anticipation.....  
/Gary

Correct me if I'm wrong, please, but it seems that in gimp I have to make a copy of the base layer and apply any adjustments to the copy; and repeat this for any new adjustment. This seems to be much less flexible, as subsequent changes to the middle layer would be obscured by the upper layer?

This is certainly true; all of your points are true. This is being worked on.
However, the specific idea of 'effect layers' is regarded as severely broken (basically cause it makes nonsense of the whole layers concept: all layers have content, but oh! effect layers don't. all layers have blending mode, but oops! effect layers don't. it's user-unfriendly in this marked inconsistency.)
The implementation I believe we are currently aiming for is instead oriented around the idea of being able to attach any number of effects to a given layer group (btw, martin nordholts is doing some great work on layer trees presently and in the last few months.. they are shaping up well.)

The photoshop method appears to be far more flexible. I was thinking that doing things this way might also have a beneficial effect on the file size, though judging by the size of photoshop format files, I doubt this is actually the case.

Photoshop format generally saves a lot of cached data -- for instance, there is a thumbnail for each layer, and a composited version of the image rendered at full size.

But I tend to save as layered tiff with zip compression applied to the layers, which makes them much smaller and preserves much of the layer information (though things like selections won't be saved. But I can live with that).