or Connect
Styleforum › Forums › Men's Style › Classic Menswear › whnay.'s good taste thread
New Posts  All Forums:Forum Nav:

whnay.'s good taste thread - Page 425

post #6361 of 12603
Quote:
Originally Posted by hendrix View Post

can we hit the reset button on the last page before things get out of hand?

It gets rather frustrating when threads continually devolve into name calling and sarcasm.

this
post #6362 of 12603
Quote:
Originally Posted by mktitsworth View Post

Actually, I just went through and computed the difference in brightness and contrast between the elements of the pocket sqare and the suit, and there's a pretty significan't difference. The suit averages out to somewhere around

a0 = {134, 138, 183};

in RGB and the main colors of the PS are

b0 = {79, 19, 73};
c0 = {160, 114, 117};
d0 = {217, 195, 208};
e0 = {60, 70, 97};

so computing the brightnesses yields

N[Mean[a0]]=151.667
N[Mean[b0]]=57.
N[Mean[c0]]=130.333
N[Mean[d0]]=206.667
N[Mean[e0]]=75.6667

with extremal differences in brightness of 55 and -94. Not insignificant distances in color space. Contrast is similar, with a mean contrast from the suit of 120.509 and extemum of 171 and 74. All in all, you're just wrong.

Except, your analysis depends completely on the spots you select to measure the values for. The eyes are better at telling the story here:

contraststudy_zps0647851a.jpg

In fact, for any outfit, studying it in grayscale is a useful exercise. Helps you understand brightness and contrast more directly.
post #6363 of 12603
Quote:
Originally Posted by mafoofan View Post

Except, your analysis depends completely on the spots you select to measure the values for. The eyes are better at telling the story here:

contraststudy_zps0647851a.jpg

The problem with this is that you've projected out all of the variance. If you look at the color values, the difference from the grey line of two of them isn't terribly great, however the vector differences between the colors obviously are. You can grey scale it all you want, but you lose information, and the information you've lost makes you wrong. We could also look at this in HSV, which would also show the problem with your analysis.
post #6364 of 12603
Also, the area of the square is roughly divided up evenly amongst the four colors, so the averages for each should not be terribly far off. I could compute an average color vector for each of the sections, but that's really unnecessary to get the point across.
post #6365 of 12603
Quote:
Originally Posted by mafoofan View Post

Except, your analysis depends completely on the spots you select to measure the values for. The eyes are better at telling the story here:

contraststudy_zps0647851a.jpg

In fact, for any outfit, studying it in grayscale is a useful exercise. Helps you understand brightness and contrast more directly.

 

My eyes tell me that I'm looking at four different levels of darkness. And while I didn't like the color version, the B&W is something I really like. In the B&W, it seems to me that the square sits nicely between the tie and the coat and helps transition between the two. I don't think it plays that role at all in the color version.

post #6366 of 12603
Quote:
Originally Posted by Patrick R View Post

My eyes tell me that I'm looking at four different levels of darkness. And while I didn't like the color version, the B&W is something I really like. In the B&W, it seems to me that the square sits nicely between the tie and the coat and helps transition between the two. I don't think it plays that role at all in the color version.

Apparently the results of eyes are contradictory. That would seem to make them less effective.
Quote:
Originally Posted by mafoofan View Post

SquareFAIL. Colors are too random, also the muted saturation and level of brightness bring it too close to your jacket. If I were to convert this pic to grayscale, you'd see the square and jacket blend together. This leaves a big gap between the jacket/square and the tie. With the square, you really want to hit a spot somewhere between. Also, regardless of what stuffing technique you use, the result shouldn't turn out like this, where the square looks like a clenched mess of rumples floating conspicuously in the center of your pocket.

I just converted to grayscale in Photoshop and I was right. Go ahead and do the same. You'll see everything turn to a similar shade of light grey--except for your tie, which will be near black.
post #6367 of 12603
Quote:
Originally Posted by mktitsworth View Post

Apparently the results of eyes are contradictory. That would seem to make them less effective.

I wouldn't say "less effective," but I would say that it helps explain differences in opinions on the fit.

 

When I squint or otherwise blur my vision, the four darkness levels in the B&W picture become even more apparent. I don't know how wise it would be for me to claim that the more I move towards blindness the more it supports my perspective!

post #6368 of 12603
What a clusterfuck. lol8[1].gif
post #6369 of 12603
We need Foo on Pink (I love both looks).
post #6370 of 12603

The inherent limitations of 2D pictorial media having shown themselves in the (chillingly bizarre) post mortem above, and with a nod to the general aura and savoir-faire that a good fit exudes in person (often transcending mean little incongruities like "wrong" colors or "pattern density"), I move that all future entries be submitted in the form of live video. Of course, those that can afford a camera crew will be blamed for having an unfair advantage...

 

Oh, yeah, that Fresco stuff upthread looked good.
 

post #6371 of 12603

Just to be a jerk (after being impressed with the technicality of mktits): the actual curves are going to differ depending on the camera model and the processing software used.

post #6372 of 12603
Quote:
Originally Posted by sprout2 View Post

The inherent limitations of 2D pictorial media having shown themselves in the (chillingly bizarre) post mortem above, and with a nod to the general aura and savoir-faire that a good fit exudes in person (often transcending mean little incongruities like "wrong" colors or "pattern density"), I move that all future entries be submitted in the form of live video. Of course, those that can afford a camera crew will be blamed for having an unfair advantage...

Oh, yeah, that Fresco stuff upthread looked good.

lol8[1].gif
post #6373 of 12603
Quote:
Originally Posted by Claghorn View Post

Just to be a jerk (after being impressed with the tecnicality of mtits), the actual curves are going to differ depending on the camera model and the processing software used.

Indubidibly. However, we're talking about the image as is and I don't think GIMP and photoshop will do a terribly different job at color grabbing. Edit: Specifically, thinking about it (obviously) because the color coordinates of each pixel should be directly computable from the compressed file.

This is totally correct though.
Edited by mktitsworth - 2/14/13 at 10:09pm
post #6374 of 12603
Quote:
Originally Posted by mafoofan View Post

The eyes are better at telling the story here.

Sometimes, however, the eyes lie. Hips, on the other-hand, do not lie.

smile.gif
post #6375 of 12603
Doh.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Classic Menswear
Styleforum › Forums › Men's Style › Classic Menswear › whnay.'s good taste thread