A gentle touch...
Moderators: rjlittlefield, ChrisR, Chris S., Pau
-
- Posts: 318
- Joined: Thu Sep 22, 2011 9:55 am
- Location: Florida
First off, excellent image geddy. Very well posed, which is something that I desperately need practice on.
To offer a different perspective on the axis issue. When working with computer graphics, there is often times multiple coordinate systems one uses. A world coordinate system which is fixed, and a local coordinate system which is object based. Some times hierarchical and temporary coordinates are used, but I digress. It makes sense to use a local transformation system when one wants to quickly understand an objects orientation, such as the Z axis being the direction of focus. A world based coordinate system requires a great deal more information for one to determine a specific objects orientation.
A user based transformation system is obviously no good. I sit off to the side of my camera when stacking often times. If I were to describe this with a user based transformation, I would likely say my camera was moving in the X. While someone else sitting behind their camera may describe it traveling in the Z. This could lead people reading it to think I am making a panorama while the other is stacking, when we both infact are stacking.
An absolute world based transformation system exists for us with N E S W up and down, but would require understanding room layout and direction as well for each user.
A local transformation based upon the camera would give us its direction of focus(Z) right away. We would however have to include additional information about the camera's rotation being vertical or horizontal. We also would have to decide to use a true local transform where the Y axis will change depending on the rotation of the Z axis, or a psuedo local transformation where even if Z is rotated 90deg, "up" would still be referred to as Y.
I think using a psuedo-local transformation would be good. This would falling line with common compositing terms, and inline with Zerene's alignment information. Where Z is depth, X is left-right, and Y is up-down based on a properly oriented image. Then we can easily apply that information to images, and camera set ups at the same time.
To offer a different perspective on the axis issue. When working with computer graphics, there is often times multiple coordinate systems one uses. A world coordinate system which is fixed, and a local coordinate system which is object based. Some times hierarchical and temporary coordinates are used, but I digress. It makes sense to use a local transformation system when one wants to quickly understand an objects orientation, such as the Z axis being the direction of focus. A world based coordinate system requires a great deal more information for one to determine a specific objects orientation.
A user based transformation system is obviously no good. I sit off to the side of my camera when stacking often times. If I were to describe this with a user based transformation, I would likely say my camera was moving in the X. While someone else sitting behind their camera may describe it traveling in the Z. This could lead people reading it to think I am making a panorama while the other is stacking, when we both infact are stacking.
An absolute world based transformation system exists for us with N E S W up and down, but would require understanding room layout and direction as well for each user.
A local transformation based upon the camera would give us its direction of focus(Z) right away. We would however have to include additional information about the camera's rotation being vertical or horizontal. We also would have to decide to use a true local transform where the Y axis will change depending on the rotation of the Z axis, or a psuedo local transformation where even if Z is rotated 90deg, "up" would still be referred to as Y.
I think using a psuedo-local transformation would be good. This would falling line with common compositing terms, and inline with Zerene's alignment information. Where Z is depth, X is left-right, and Y is up-down based on a properly oriented image. Then we can easily apply that information to images, and camera set ups at the same time.
-
- Posts: 161
- Joined: Thu May 23, 2013 6:46 am
- Location: Richmond, Virginia, USA
Hurray up, Chris, I sense a gathering consensus...
So what do y'all think of "image-x, image-y, image-z" as unambiguous terms?
________________________________________
On a side note, the theater world uses a "local transformation" for stage directions, from the viewpoint of an actor facing the audience:
x = "stage left/stage right" (left/right)
y = "downstage/upstage" (forward/backward)
z = "in/out" (down/up)
So an actor at CENTERSTAGE might move DOWN LEFT before being being flown OUT suspended from a cable attached to a harness.
So what do y'all think of "image-x, image-y, image-z" as unambiguous terms?
________________________________________
On a side note, the theater world uses a "local transformation" for stage directions, from the viewpoint of an actor facing the audience:
x = "stage left/stage right" (left/right)
y = "downstage/upstage" (forward/backward)
z = "in/out" (down/up)
So an actor at CENTERSTAGE might move DOWN LEFT before being being flown OUT suspended from a cable attached to a harness.
- rjlittlefield
- Site Admin
- Posts: 23621
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
OK... But among whom, and for what purpose?Bill Eldridge wrote:Hurray up, Chris, I sense a gathering consensus...
Mathematics is widely touted for being unambiguous, but I can't help noting that the function f(x) = 3x+5 either is or is not "linear", depending on which common definition of that term you're using.
So, the key to success (= getting the right result) is to
a) recognize that the same term may have different meanings in different contexts, and
b) make dang sure that you and whoever you're trying to converse with are either using the same definitions or making the appropriate translations.
Playing devil's advocate, I'll inquire whether that would be before or after Zerene Stacker applies its Options > Preferences > Preprocessing > Image Pre-rotation. Equally, what if the camera is rotated? What if the camera is physically rotated but is set to compensate for that using its internal orientation sensor, and "compensate" means to set a flag in the image header which it turns out that not all image reading libraries will honor? No, that last one is not hypothetical. I have seen images that some softwares display "right side up" and others don't.So what do y'all think of "image-x, image-y, image-z" as unambiguous terms?
As you can probably gather from my comments, I am not a great believer in the concept of "unambiguous" terms.
My personal soapbox is exactly the opposite: to repeatedly point out that even highly technical terms often are ambiguous, and that most of the time you're better off to just recognize that fact and deal with it.
--Rik
-
- Posts: 161
- Joined: Thu May 23, 2013 6:46 am
- Location: Richmond, Virginia, USA
Well, OK, Rik, "x" and "y" can get flipped by camera orientation which may not be honored by some software. This suggests the rather ungainly "intended-image-x, intended-image-y"!
But geddy, johan, TheLostVertex and I seem to prefer "z" to mean focus-stepping axis in all cases. A consensus of four... Perhaps "z (focus-stepping axis)" is a clear term.
And the purpose of standardizing terminology? Avoiding the chaos you to which you refer:
But geddy, johan, TheLostVertex and I seem to prefer "z" to mean focus-stepping axis in all cases. A consensus of four... Perhaps "z (focus-stepping axis)" is a clear term.
And the purpose of standardizing terminology? Avoiding the chaos you to which you refer:
In defense of geddy's labeling, it's very common for stacking systems to call the focus-stepping direction "z" regardless of how it happens to be oriented with respect to gravity.
This avoids the chaos that would otherwise occur when a StackShot happens to be horizontal at one moment, vertical at a second, and oblique at a third. The situation is similar
for the common term "depth map", which always means distance along the focus-stepping axis again independent of gravity.
-
- Posts: 318
- Joined: Thu Sep 22, 2011 9:55 am
- Location: Florida
A good thing to consider, though I did mention the caveat of "up-down based on a properly oriented image". Though what is proper for an image of crystals under a microscope for instance, may not be determinable.rjlittlefield wrote:Playing devil's advocate, I'll inquire whether that would be before or after Zerene Stacker applies its Options > Preferences > Preprocessing > Image Pre-rotation. Equally, what if the camera is rotated? What if the camera is physically rotated but is set to compensate for that using its internal orientation sensor, and "compensate" means to set a flag in the image header which it turns out that not all image reading libraries will honor? No, that last one is not hypothetical. I have seen images that some softwares display "right side up" and others don't.So what do y'all think of "image-x, image-y, image-z" as unambiguous terms?
I agree very much. Often times the main barrier to some one understanding is how we all interpret language slightly differently. Ultimately I feel like the most important point to be taken away is what you said,rjlittlefield wrote:As you can probably gather from my comments, I am not a great believer in the concept of "unambiguous" terms.
My personal soapbox is exactly the opposite: to repeatedly point out that even highly technical terms often are ambiguous, and that most of the time you're better off to just recognize that fact and deal with it.
As even if many of us in this thread agree to a standard concept, not everyone else may be aware of that or even accept it. And perhaps other circumstances we havent thought of may arise making things a little more muddy.My original post was really just a long winded way of my saying, I do not think that any method will be perfect, but I think settling on an image based descriptor would be simple for most cases.rjlittlefield wrote: b) make dang sure that you and whoever you're trying to converse with are either using the same definitions or making the appropriate translations.
-
- Posts: 161
- Joined: Thu May 23, 2013 6:46 am
- Location: Richmond, Virginia, USA
Rik and geddy, what do you think about pulling this extended discussion of axial nomenclature into a new thread in the "Macro and Micro Technique and Technical Discussions" forum,
starting with Chris S.'s challenge? I'd be interested in reading the thoughts of others who might wish to weigh in on this topic but might not find it under the heading "A gentle touch..."
starting with Chris S.'s challenge? I'd be interested in reading the thoughts of others who might wish to weigh in on this topic but might not find it under the heading "A gentle touch..."
- rjlittlefield
- Site Admin
- Posts: 23621
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
Right, and that's the key driver for me.TheLostVertex wrote:... even if many of us in this thread agree to a standard concept, not everyone else may be aware of that or even accept it.
Within the focus-stacking community, we might comfortably agree that the focus-stepping axis is called Z and the two axes orthogonal to that are X and Y, with the "obvious" assignments relative to the image. If more precision is needed, probably a labeled diagram is called for.
Outside the focus-stacking community, we should expect that the term "focus-stepping axis" will mean nothing at all, and X/Y/Z will probably be interpreted however the listener is used to using those. This provides great opportunity for confusion.
As far as I know, Chris S. is almost unique among us focus-stackers in that he specifies optical equipment that other people who are mechanical specialists are supposed to build. I'm guessing that this is the reason for his sensitivity to nomenclature. But I'm often wrong when I guess, so I look forward to either being corrected or having him fill in the details for himself.
--Rik
- rjlittlefield
- Site Admin
- Posts: 23621
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
I think that makes enormous sense.what do you think about pulling this extended discussion of axial nomenclature into a new thread in the "Macro and Micro Technique and Technical Discussions" forum
There's a nice mechanism in the forum software for "splitting" a thread in two. But there is no matching mechanism for re-joining a split thread, so before I do the split I'd like to know whether anybody objects or has a better idea.
Y'all have 8 hours to speak up about that, starting now...
--Rik
-
- Posts: 161
- Joined: Thu May 23, 2013 6:46 am
- Location: Richmond, Virginia, USA
Well, of course. Outside the music theory community, "pitch class 0" means nothing, and outside the theater community "in" means a variety of things.Outside the focus-stacking community, we should expect that the term "focus-stepping axis" will mean nothing at all
But inside a theater, when a flyman yells, "Pipe coming in!", you'd better look up and prepare to move aside or duck because a 70-foot long piece of steel
hung with a couple of dozen lighting instruments may be about to bonk you on the head! Just one example of the utility of standardized terms within a field...
If you would, please let me start that thread, as I have a graphic that I'd like to be at the top, not buried in a later post, as a common reference point for the discussion. Am working on the post now, but need to stop for dinner. Will have it up tonight.rjlittlefield wrote:Y'all have 8 hours to speak up about that, starting now...
OK?
--Chris
- rjlittlefield
- Site Admin
- Posts: 23621
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
I can't do what I think you're imagining. There's no mechanism for moving old posts to the end of a new post. Actually no mechanism for moving posts to a different thread, other than to split an existing thread to start a new one.Chris S. wrote:If you would, please let me start that thread, as I have a graphic that I'd like to be at the top, not buried in a later post, as a common reference point for the discussion. Am working on the post now, but need to stop for dinner. Will have it up tonight.rjlittlefield wrote:Y'all have 8 hours to speak up about that, starting now...
OK?
--Chris
So, you can start a new thread to put your graphic at the top, but we'll have to cross-reference between that thread and whatever we do with this one.
--Rik
Well, here it is: The six axes of translation and rotation.
Sorry that it is a long post--I didn't have time to write a short one.
--Chris
Sorry that it is a long post--I didn't have time to write a short one.
--Chris