Photoshop question

[bb113]

If I measure the areas of 20 different selections (from the same layer), how does photoshop determine the order of the selections in the measurement log?

At first I thought it was top "row" (probably average pixel location or something) left to right , second row left to right, etc... and it usually does work this way, but then I noticed under certain circumstances the order is strange. I can give further details if anyone is interested... I just spent a couple hours trying to figure it out and only got more confused though.

[1714904982]

1714904982

[1714904982]

1714904982

[1714904982]

1714904982

[1714904982]

1714904982

[amencon]

Sure lets here the further details.  Not sure I fully understand what you are asking but sounds like an interesting photoshop question.

[Phinnaeus Gage]

If I measure the areas of 20 different selections (from the same layer), how does photoshop determine the order of the selections in the measurement log?

At first I thought it was top "row" (probably average pixel location or something) left to right , second row left to right, etc... and it usually does work this way, but then I noticed under certain circumstances the order is strange. I can give further details if anyone is interested... I just spent a couple hours trying to figure it out and only got more confused though.

After you find the answer, shop the image below. Incorporate a Bitcoin logo within that ring ejected from the sun. Once created, we'll have a caption contest.

~Bruno~

[bb113]

So here it is:



Each of those is a slice of a brain, #1 is towards the front, # 36 is most towards the back. What I want to do is compare the area of the left side with the area of the right side for each slice and see how this changes going from front to back. Let's ignore the right side for now.

If I use the quick selection tool to select each of the left sides (#6 to #36) and then "record measurements" (under analyze), I get this ordering:

Code:

"Document"	"Scale Units"	"Count"	"Area"
"280.psd"	"cm"	"31"	"14.302058"
"280.psd"	"cm"		"0.217312"
"280.psd"	"cm"		"0.244420"
"280.psd"	"cm"		"0.262937"
"280.psd"	"cm"		"0.270576"
"280.psd"	"cm"		"0.308789"
"280.psd"	"cm"		"0.337050"
"280.psd"	"cm"		"0.352047"
"280.psd"	"cm"		"0.358225"
"280.psd"	"cm"		"0.376706"
"280.psd"	"cm"		"0.401809"
"280.psd"	"cm"		"0.412669"
"280.psd"	"cm"		"0.422004"
"280.psd"	"cm"		"0.424898"
"280.psd"	"cm"		"0.461796"
"280.psd"	"cm"		"0.472783"
"280.psd"	"cm"		"0.462867"
"280.psd"	"cm"		"0.494974"
"280.psd"	"cm"		"0.502196"
"280.psd"	"cm"		"0.514262"
"280.psd"	"cm"		"0.536235"
"280.psd"	"cm"		"0.545036"
"280.psd"	"cm"		"0.561738"
"280.psd"	"cm"		"0.562155"
"280.psd"	"cm"		"0.573813"
"280.psd"	"cm"		"0.646003"
"280.psd"	"cm"		"0.663023"
"280.psd"	"cm"		"0.623948"
"280.psd"	"cm"		"0.533713"
"280.psd"	"cm"		"0.565140"
"280.psd"	"cm"		"0.608116"
"280.psd"	"cm"		"0.584818"

If I hit "record measurements" after selecting each left side (going from top row left to right, then second row left to right, then third row left to right)... It orders them this way:

Code:

"280.psd"	"cm"	"1"	"0.217312"
"280.psd"	"cm"	"1"	"0.244420"
"280.psd"	"cm"	"1"	"0.262937"
"280.psd"	"cm"	"1"	"0.270576"
"280.psd"	"cm"	"1"	"0.308789"
"280.psd"	"cm"	"1"	"0.337050"
"280.psd"	"cm"	"1"	"0.352047"
"280.psd"	"cm"	"1"	"0.358225"
"280.psd"	"cm"	"1"	"0.376706"
"280.psd"	"cm"	"1"	"0.401809"
"280.psd"	"cm"	"1"	"0.412669"
"280.psd"	"cm"	"1"	"0.422004"
"280.psd"	"cm"	"1"	"0.424898"
"280.psd"	"cm"	"1"	"0.461796"
"280.psd"	"cm"	"1"	"0.472783"
"280.psd"	"cm"	"1"	"0.462867"
"280.psd"	"cm"	"1"	"0.494974"
"280.psd"	"cm"	"1"	"0.502196"
"280.psd"	"cm"	"1"	"0.514262"
"280.psd"	"cm"	"1"	"0.536235"
"280.psd"	"cm"	"1"	"0.545036"
"280.psd"	"cm"	"1"	"0.533713"
"280.psd"	"cm"	"1"	"0.561738"
"280.psd"	"cm"	"1"	"0.562155"
"280.psd"	"cm"	"1"	"0.565140"
"280.psd"	"cm"	"1"	"0.573813"
"280.psd"	"cm"	"1"	"0.608116"
"280.psd"	"cm"	"1"	"0.584818"
"280.psd"	"cm"	"1"	"0.646003"
"280.psd"	"cm"	"1"	"0.663023"
"280.psd"	"cm"	"1"	"0.623948"

The ordering is the same until it gets to the last row (starting at #27). The last 10 measurements are ordered in some strange way. The last of the common ordering is # 26:
"280.psd"   "cm"   "1"   "0.545036"

It is much faster to select all of the left hemisphere then hit record than to hit record after selecting each. The second way of ordering is what I want to get by using the first method. I can't figure out why the last row is getting put out of order. I can send the file if interested.

[LoupGaroux]

Don't quote me on this, but from my own experiments with a different release of PShop, I believe it has to do with the order that you created the original selections. If you started with the US English convention of upper left corner of the upper left image, worked clockwise, left to right and sequentially in order... it should present in the same order. That would certainly be the case in AutoCAD, which I am significantly more experienced with.

Any calculation or operation polling the data in the file would start with the first data-sets in the record, no?

But then I wonder... did you select one at a time with the wand, or did you window selection by variance? Wand would hold together for a certain amount of data, but then would hit cache storage which would explain the odd ordering at the end; while windowed selection would read it as one set of datapoints as a whole.

[bb113]

I have tested this. It doesn't matter what order I select them in. It will order in the same way. This may have something to do with how they were added to the image, but if so it is in a non-obvious way. The way they were added is somewhat difficult to explain without images. I'll do it if someone wants to replicate the process though.

Edit: sorry I misread at first

But then I wonder... did you select one at a time with the wand, or did you window selection by variance? Wand would hold together for a certain amount of data, but then would hit cache storage which would explain the odd ordering at the end; while windowed selection would read it as one set of datapoints as a whole.

Also this may make sense, but why would it start ordering semi-randomly. I did move some sections around to organize them, maybe you are right.

[bb113]

Another thing, I didn't test it with this example, but with another I went through to find out which I could select while preserving the ordering. It turned out I could select the entire 2nd and 3rd rows and it would be fine. It would also be fine if I added the second to last of the first row, but any others from the first row would mess up the ordering (only of the last row) in different ways. Then I found out if I moved this second to last section up half a cm (according to the ruler) it would also mess up the ordering. If I moved any single other one down half a cm I would be able to select it while getting the correct the ordering. Eventually I moved the entire first row down half a centimeter and the ordering was correct.

[bb113]

But then I wonder... did you select one at a time with the wand, or did you window selection by variance? Wand would hold together for a certain amount of data, but then would hit cache storage which would explain the odd ordering at the end; while windowed selection would read it as one set of datapoints as a whole.

Also, the same thing happens if I select all of the left hemispheres and create a channel, then go into the channel and select all at once (with contiguous turned off).