Amber and Yellow color perception

2 Feb 2007 - 10:10pm
7 years ago
11 replies
2044 reads
DanP
2006

Hi All,

I wonder if anyone could point me to some research or data regarding
human ability to distinguish between the colors yellow and amber? I
have an alarm state prototype, and I'm trying to determine (with
actual research) whether this will pose a problem.

Thanks for any assistance, much appreciated!
-Dan

Comments

3 Feb 2007 - 3:54pm
cfmdesigns
2004

On Feb 2, 2007, at 7:10 PM, dnp607 wrote:

> I wonder if anyone could point me to some research or data regarding
> human ability to distinguish between the colors yellow and amber? I
> have an alarm state prototype, and I'm trying to determine (with
> actual research) whether this will pose a problem.

A great rule of thumb would seem to be "If you think it might cause a
problem, it will". This is one of those cases where I don't think
you should *need* research, just your internal alarm bells. But the
research you want would be on color blindness, color identification,
and low-light level reception, and maybe on LED quality and aging
(since I can imagine vagaries in manufacture having an effect here).

The place where this would apt to be less of a concern is when the
yellow can only occur in one part of the alarm hardware, and the
amber only in another. "If yellow appears at the left end of the
row, check your battery levels." "If amber appears at the right end,
the alarm is offline." Then you don't need to rely on fine color
gradation differences, just on gross ones. Actions like "yellow
always flashes on/off, amber is steady" would also be good, I would
think -- never rely just on color alone.

-- Jim

3 Feb 2007 - 4:53pm
.pauric
2006

We use 'yellow' as a label on the front of our boxes (signifies reduced
functionality) It is for all intents and purposes indistinguishable from
Amber. I believe Amber is produced by combining a Green and Yellow/Red LED,
but I'm not 100% sure which, its been a while since we played with
combinations.

I do remember one case where a vendor produced Amber instead of Yellow on
some alpha hardware. I think some loaded a component feeder with the wrong
part. It simply looked odd to users, subtly different but not
distinguishable. From my experience with LEDs I would strongly advise
against using yellow/amber to represnt different states.

But, I dont have and hard research.

3 Feb 2007 - 9:36pm
Jeff Howard
2004

This is one of the few cases where George Miller's 1956 study "The
Magic Number Seven, Plus or Minus Two" actually applies. Based on
Miller's research I'd say that yes, most people are physiologically
capable of distinguishing between yellow and amber. But it's still
going to pose a problem.

http://psychclassics.yorku.ca/Miller/

Miller was studying the human ability to distinguish values along
unidimensional and multi-dimensional scales. Things related to color,
sound and taste for example. Basically, he found that most people have
the ability to reliably distinguish (by memory) about seven variations
on any particular scale. Some people are better or worse at this, so
the range is generally between five and nine variations. For example,
most people can distinguish around seven musical tones on a scale and
identify them when played without making mistakes.

With color it's not as clear-cut since colors often vary along
multiple dimensions (hue, value and satuation) but given a set of
colors that vary only by hue ranging from yellow to amber, people in
a psychology experiment could, on average, distinguish about seven
different variations in yellow/amber. Given that you only need to
make them distinguish two variations, that's well within most
people's perceptual threshold.

But outside of psychology labs people aren't in the habit of making
those kind of discriminations unless they know it's important. Most
of the time it's not, so people gloss over the details. Think of
traffic lights. Older traffic signals use a yellowish green for
"go" while newer lights use blue-green LEDs. If you asked them,
most people could tell the difference but they've probably never
thought about it. There weren't any publc service announcement to
teach people about the new blue-green lights because we're used to
those subtle variations being meaningless.

If you want a reference for how people _really_ behave (as opposed to
what they're capable of) then I'd suggest the writings of Herbert
Simon on the concept of "chunking".

http://www.albany.edu/~dkw42/s6_chunked.html

In "How Big is a Chunk" (1974) Simon describes our tendency not to
attend to the overwhelming complexity of existence, but to mentally
"chunk" phenomenon into common patterns. I'd say that a relevant
example of "chunking" behavior is our ability to see the complexity
of the color spectrum as only seven (Miller) colors with the mnemonic
ROYGBIV. (And as far as I'm concerned you can throw out indigo--it
always seemed unnecessary.)

Getting back to your problem... If you absolutely had to use yellow
and amber, then make the distinction as pronounced as you can by
varying the saturation or value in addition to the hue (to the extent
that's possible with LEDs). Jim's suggestions about different
behaviors for the colors makes the distinction even more pronounced.
But it's safer to use some other color.

3 Feb 2007 - 10:41pm
DanP
2006

Jeff, Jim, Pauric,

Thank you for taking the time to respond - this was valuable info.

I think Jeff's conclusion summed it up. Vary the saturation and hue
as much as possible, but best not to do it in the first place. In the
way of research, I just read about computing predicted
discriminability from "CIE" coordinates. It's based on something
called "MacAdam Ellipses". I don't fully understand it yet, but from
what I read, there are formulas that will calculate how many units
(dichromats) should be in between each color in order to best
separate them visually (in a cone space). If forced to use similar
colors, this will at least give me the maximum discrimination given
the constraint..

Thanks again everyone.
-Dan

On Feb 3, 2007, at 6:36 PM, Jeff Howard wrote:

> This is one of the few cases where George Miller's 1956 study "The
> Magic Number Seven, Plus or Minus Two" actually applies. Based on
> Miller's research I'd say that yes, most people are physiologically
> capable of distinguishing between yellow and amber. But it's still
> going to pose a problem.

3 Feb 2007 - 11:18pm
.pauric
2006

"Jim's suggestions about different behaviors for the colors makes the
distinction even more pronounced."

Flashing an LED actually makes it less pronounced. Less time for the eye to
perceive subtle difference in colours. That is, flashing an LED will make
it harder to determine whether is it yellow or amber compared with having it
on all the time.

If you do use flashing, our researched found LEDs needed to be on for 30ms
to be registered properly. We didnt look in to how long it needed to be on
too tell the difference between and amber or yellow. As a general rule we
keep the number of states displayed via LED to a minimum, they're a little
bit 'binary' if you understand what I mean.

I think asking the user to build a matrix of green/amber/yellow/blue/red
over solid/flashing and have them translate states or meaning through that
matrix, could be phrased as an engineering based interaction solution.

On 2/3/07, Jeff Howard <id at howardesign.com> wrote:
>
> This is one of the few cases where George Miller's 1956 study "The
> Magic Number Seven, Plus or Minus Two" actually applies. Based on
> Miller's research I'd say that yes, most people are physiologically
> capable of distinguishing between yellow and amber. But it's still
> going to pose a problem.
>
> http://psychclassics.yorku.ca/Miller/
>
> Miller was studying the human ability to distinguish values along
> unidimensional and multi-dimensional scales. Things related to color,
> sound and taste for example. Basically, he found that most people have
> the ability to reliably distinguish (by memory) about seven variations
> on any particular scale. Some people are better or worse at this, so
> the range is generally between five and nine variations. For example,
> most people can distinguish around seven musical tones on a scale and
> identify them when played without making mistakes.
>
> With color it's not as clear-cut since colors often vary along
> multiple dimensions (hue, value and satuation) but given a set of
> colors that vary only by hue ranging from yellow to amber, people in
> a psychology experiment could, on average, distinguish about seven
> different variations in yellow/amber. Given that you only need to
> make them distinguish two variations, that's well within most
> people's perceptual threshold.
>
> But outside of psychology labs people aren't in the habit of making
> those kind of discriminations unless they know it's important. Most
> of the time it's not, so people gloss over the details. Think of
> traffic lights. Older traffic signals use a yellowish green for
> "go" while newer lights use blue-green LEDs. If you asked them,
> most people could tell the difference but they've probably never
> thought about it. There weren't any publc service announcement to
> teach people about the new blue-green lights because we're used to
> those subtle variations being meaningless.
>
> If you want a reference for how people _really_ behave (as opposed to
> what they're capable of) then I'd suggest the writings of Herbert
> Simon on the concept of "chunking".
>
> http://www.albany.edu/~dkw42/s6_chunked.html<http://www.albany.edu/%7Edkw42/s6_chunked.html>
>
> In "How Big is a Chunk" (1974) Simon describes our tendency not to
> attend to the overwhelming complexity of existence, but to mentally
> "chunk" phenomenon into common patterns. I'd say that a relevant
> example of "chunking" behavior is our ability to see the complexity
> of the color spectrum as only seven (Miller) colors with the mnemonic
> ROYGBIV. (And as far as I'm concerned you can throw out indigo--it
> always seemed unnecessary.)
>
> Getting back to your problem... If you absolutely had to use yellow
> and amber, then make the distinction as pronounced as you can by
> varying the saturation or value in addition to the hue (to the extent
> that's possible with LEDs). Jim's suggestions about different
> behaviors for the colors makes the distinction even more pronounced.
> But it's safer to use some other color.
> ________________________________________________________________
> Welcome to the Interaction Design Association (IxDA)!
> To post to this list ....... discuss at ixda.org
> List Guidelines ............ http://listguide.ixda.org/
> List Help .................. http://listhelp.ixda.org/
> (Un)Subscription Options ... http://subscription-options.ixda.org/
> Announcements List ......... http://subscribe-announce.ixda.org/
> Questions .................. lists at ixda.org
> Home ....................... http://ixda.org/
> Resource Library ........... http://resources.ixda.org
>

--
Job type: In house
Field: Embedded & physical interfaces. Web/cli

4 Feb 2007 - 2:20am
Jeff Howard
2004

One other thing I wanted to mention is that it's easier to apply a
pattern if we have a name for it. I'd be surprised if anyone who
didn't design LEDs for a living used the word "amber" to describe
anything other than fossilized tree resin or waves of grain. Without
the right words, we're left trying to describe rather than identify
(like yellowish-green in my example with traffic lights). Describing
a thing's properties is different than recognizing it by those
properties as a pattern. If you don't have the word "amber" in
your lexicon, it's harder to see.

To pauric's point, the overall distinction between a blinking LED
and one that's not blinking is greater, even if the color
distinction is masked. Whether or how that difference carries meaning
is another matter. The first time my younger brother encountered a
blinking red traffic signal, he had no idea what it was telling him
to do, but he recognized that it was different from a regular stop
light.

4 Feb 2007 - 3:50am
Chris McLay
2005

On 04/02/2007, at 4:20 PM, Jeff Howard wrote:

> To pauric's point, the overall distinction between a blinking LED
> and one that's not blinking is greater, even if the color
> distinction is masked. Whether or how that difference carries meaning
> is another matter. The first time my younger brother encountered a
> blinking red traffic signal, he had no idea what it was telling him
> to do, but he recognized that it was different from a regular stop
> light.

What does a blinking red traffic light mean? We have blinking orange
(amber) lights here which happen when there is a fault with the
signals. Is blinking red the same thing?

Chris

--
Chris McLay.
Interaction & Visual Designer

chris at eeoh.com.au http://eeoh.com.au/chris/

4 Feb 2007 - 4:32am
Jeff Howard
2004

Chris McLay wrote:
> What does a blinking red traffic light mean?

In the US, a blinking red light functions like a stop sign. Come to a
complete stop and then proceed if it's clear. The opposing traffic
often has a blinking yellow, or the intersection can become a
four-way stop if both directions are blinking red. In the particular
intersection I was referring to, it changed from a regular stop light
to a blinking stop light after midnight.

There's a blinking yellow light near where I live. It's at a
horrible intersection and I noticed today that a sign had been
installed next to the blinking light that essentially explained how
it should be interpreted. "Proceed with caution on flashing
yellow."

4 Feb 2007 - 8:27am
.pauric
2006

Sorry Jeff, you are right it is easy to tell the difference between a
flashing and solid LED. However I think what was being proposed was to have
a solid yellow and flashing amber LED.

The blinking state is generally a transient or dependent state of the
'solid' state
In your example, solid red means stop. Flashing red means stop and proceed.
We use flashing green power LED for booting and solid for booted.

For all intent and purposes flashing amber will look like flashing yellow.
Viewers will assume the flashing amber LED is trying to tell them something
based upon the solid yellow LED state.

So, the most we get out of a single LED are 3 states; off, flashing & on.
We used to have varying states of flashing but again back to the matrix of
possible states and difficulty in interpreting them.

Although we dont use amber & yellow together, I would fully expect to see
users assume that the meaning of those two LED is related. With the Amber
being associated with a fault as they generally look kind of redish.

4 Feb 2007 - 2:19pm
cfmdesigns
2004

On Feb 3, 2007, at 8:18 PM, pauric wrote:

> "Jim's suggestions about different behaviors for the colors makes the
> distinction even more pronounced."
>
> Flashing an LED actually makes it less pronounced. Less time for
> the eye to
> perceive subtle difference in colours. That is, flashing an LED
> will make
> it harder to determine whether is it yellow or amber compared with
> having it
> on all the time.

Yes, I would agree with that. I was thinking about it in terms of
other ways of differentiating it, though: the user doesn't have to
tell the difference between yellow and amber if all he has to do is
tell the difference between yelllow-ish flashing and yellow-ish solid.

-- Jim

5 Feb 2007 - 9:51am
.pauric
2006

We separate colour and specific meaning. That is, people normally associate
green with all is good. We use green in different contexts for successful
boot-up, max speed and good connection. We use yellow for partial fault,
limited connection. We use red for critical faults.

These associations are separate from any labels that can be applied to the
LED. When I need to display a large amount of information through LEDs I
employ a Mode button. The trick is that through pressing the mode button
the user can use common associations across different contexts, thus
reducing the load on the user. Here's a mechanical drawing of a section of
one of our overlays. The LEDs convey speed, type of connection and power
over ethernet, across 52 ports.

<http://web.mac.com/pauric_ocallaghan/overlay_cad.jpg>
and in practice
<http://web.mac.com/pauric_ocallaghan/2ustackLED.png>

Apologies for the small images, dont have anything larger to hand right now
but I hope this demonstrates you don't necessarily need to rely on specific
colour = specific meaning.

Alternatively, simply use LEDs to illuminate the label, whatever that may be
<
http://images.google.com/imgres?imgurl=http://www.comtecs.co.uk/images/on.jpg&imgrefurl=http://www.comtecs.co.uk/technology.xhtml&h=171&w=171&sz=38&hl=en&sig2=RCcbwdg219dZOAGwbXfMwg&start=8&tbnid=gXOl3IK3G_u-4M:&tbnh=100&tbnw=100&ei=2j_HRZCJMZCsiwGT0ciZDg&prev=/images%3Fq%3Dpower%2Bbutton%26svnum%3D10%26hl%3Den%26client%3Dfirefox-a%26rls%3Dorg.mozilla:en-US:official%26sa%3DG
>

On 2/4/07, Jim Drew <cfmdesigns at earthlink.net> wrote:
>
>
> On Feb 3, 2007, at 8:18 PM, pauric wrote:
>
> > "Jim's suggestions about different behaviors for the colors makes the
> > distinction even more pronounced."
> >
> > Flashing an LED actually makes it less pronounced. Less time for
> > the eye to
> > perceive subtle difference in colours. That is, flashing an LED
> > will make
> > it harder to determine whether is it yellow or amber compared with
> > having it
> > on all the time.
>
> Yes, I would agree with that. I was thinking about it in terms of
> other ways of differentiating it, though: the user doesn't have to
> tell the difference between yellow and amber if all he has to do is
> tell the difference between yelllow-ish flashing and yellow-ish solid.
>
> -- Jim
>
> ________________________________________________________________
> Welcome to the Interaction Design Association (IxDA)!
> To post to this list ....... discuss at ixda.org
> List Guidelines ............ http://listguide.ixda.org/
> List Help .................. http://listhelp.ixda.org/
> (Un)Subscription Options ... http://subscription-options.ixda.org/
> Announcements List ......... http://subscribe-announce.ixda.org/
> Questions .................. lists at ixda.org
> Home ....................... http://ixda.org/
> Resource Library ........... http://resources.ixda.org
>

--
Job type: In house
Field: Embedded & physical interfaces. Web/cli

Syndicate content Get the feed