lab()

Below are descriptions of the allowed values for both absolute and relative colors.

Absolute value syntax

lab(L a b[ / A])

The parameters are as follows:

L

A <number> between 0 and 100, a <percentage> between 0% and 100%, or the keyword none (equivalent to 0% in this case). This value specifies the color's lightness. Here the number 0 corresponds to 0% (black) and the number 100 corresponds to 100% (white).

a

A <number> between -125 and 125, a <percentage> between -100% and 100%, or the keyword none (equivalent to 0% in this case). This value specifies the color's distance along the a axis, which defines how green (moving towards -125) or red (moving towards +125) the color is. Note that these values are signed (allowing both positive and negative values) and theoretically unbounded, meaning that you can set values outside the ±125 (±100%) limits. In practice, values cannot exceed ±160.

b

A <number> between -125 and 125, a <percentage> between -100% and 100%, or the keyword none (equivalent to 0% in this case). This value specifies the color's distance along the b axis, which defines how blue (moving towards -125) or yellow ( moving towards +125) the color is. Note that these values are signed (allowing both positive and negative values) and theoretically unbounded, meaning that you can set values outside the ±125 (±100%) limits. In practice, values cannot exceed ±160.

A Optional

An <alpha-value> representing the alpha channel value of the color, where the number 0 corresponds to 0% (fully transparent) and 1 corresponds to 100% (fully opaque). Additionally, the keyword none can be used to explicitly specify no alpha channel. If the A channel value is not explicitly specified, it defaults to 100%. If included, the value is preceded by a slash (/).

Relative value syntax

lab(from <color> L a b[ / A])

The parameters are as follows:

from <color>

The keyword from is always included when defining a relative color, followed by a <color> value representing the origin color. This is the original color that the relative color is based on. The origin color can be any valid <color> syntax, including another relative color.

L

A <number> between 0 and 100, a <percentage> between 0% and 100%, or the keyword none (equivalent to 0% in this case) This value represents the lightness of the output color. Here the number 0 corresponds to 0% (black) and the number 100 corresponds to 100% (white).

a

A <number> between -125 and 125, a <percentage> between -100% and 100%, or the keyword none (equivalent to 0% in this case). This value represents the output color's distance along the a axis, which defines how green (moving towards -125) or red (moving towards +125) the color is. Note that these values are signed (allowing both positive and negative values) and theoretically unbounded, meaning that you can set values outside the ±125 (±100%) limits. In practice, values cannot exceed ±160.

b

A <number> between -125 and 125, a <percentage> between -100% and 100%, or the keyword none (equivalent to 0% in this case). This value represents the output color's distance along the b axis, which defines how blue (moving towards -125) or yellow (moving towards +125) the color is. Note that these values are signed (allowing both positive and negative values) and theoretically unbounded, meaning that you can set values outside the ±125 (±100%) limits. In practice, values cannot exceed ±160.

A Optional

An <alpha-value> representing the alpha channel value of the output color, where the number 0 corresponds to 0% (fully transparent) and 1 corresponds to 100% (fully opaque). Additionally, the keyword none can be used to explicitly specify no alpha channel. If the A channel value is not explicitly specified, it defaults to the alpha channel value of the origin color. If included, the value is preceded by a slash (/).

Note: Usually when percentage values have a numeric equivalent in CSS, 100% is equal to the number 1. This is not always the case for LAB's lightness and a and b axes, as mentioned above. With L, the range is from 0 to 100, with 100% equal to 100. The a and b values support both negative and positive values, with 100% being equal to 125 and -100% being equal to -125.

Defining relative color output channel components

When using relative color syntax inside a lab() function, the browser converts the origin color into an equivalent Lab color (if it is not already specified as such). The color is defined as three distinct color channel values — l (lightness), a (green/red axis), and b (blue/yellow axis) — plus an alpha channel value (alpha). These channel values are made available inside the function to be used when defining the output color channel values:

• The l channel value is resolved to a <number> between 0 and 100, inclusive.
• The a and b channels are each resolved to a <number> between -125 and 125, inclusive.
• The alpha channel is resolved to a <number> between 0 and 1, inclusive.

When defining a relative color, the different channels of the output color can be expressed in several different ways. Below, we'll study some examples to illustrate these.

In the first two examples below, we are using relative color syntax. However, the first one outputs the same color as the origin color and the second one outputs a color not based on the origin color at all. They don't really create relative colors! You'd be unlikely to ever use these in a real codebase, and would probably just use an absolute color value instead. We included these examples as a starting point for learning about relative lab() syntax.

Let's start with an origin color of hsl(0 100% 50%) (equivalent to red). The following function outputs the same color as the origin color — it uses the origin color's l, a, and b channel values (54.29, 80.8198, and 69.8997) as the output channel values:

lab(from hsl(0 100% 50%) l a b)

This function's output color is lab(54.29 80.8198 69.8997).

The next function uses absolute values for the output color's channel values, outputting a completely different color not based on the origin color:

lab(from hsl(0 100% 50%) 29.692% 44.89% -29.034%)

In the above case, the output color is lab(29.692 56.1125 -36.2925).

The following function creates a relative color based on the origin color:

lab(from hsl(0 100% 50%) l -100 b)

This example:

• Converts the hsl() origin color to an equivalent lab() color — lab(54.29 80.8198 69.8997).
• Sets the l and b channel values for the output color to those of the origin lab() equivalent's L and b channel values — those values are 54.29 and 69.8997, respectively.
• Sets the output color's a channel value to a new value not based on the origin color: -100.

The final output color is lab(54.29 -100 69.8997).

In the examples we've seen so far in this section, the alpha channels have not been explicitly specified for either the origin or output colors. When the output color alpha channel is not specified, it defaults to the same value as the origin color alpha channel. When the origin color alpha channel is not specified (and it is not a relative color), it defaults to 1. Therefore, the origin and output alpha channel values are 1 for the above examples.

Let's look at some examples that specify origin and output alpha channel values. The first one specifies the output alpha channel value as being the same as the origin alpha channel value, whereas the second one specifies a different output alpha channel value, unrelated to the origin alpha channel value.

lab(from hsl(0 100% 50% / 0.8) l a b / alpha)
/* Computed output color: lab(54.29 80.8198 69.8997 / 0.8) */

lab(from hsl(0 100% 50% / 0.8) l a b / 0.5)
/* Computed output color: lab(54.29 80.8198 69.8997 / 0.5) */

In the following example, the hsl() origin color is again converted to the lab() equivalent — lab(54.29 80.8198 69.8997). calc() calculations are applied to the L, a, b, and A values, resulting in an output color of lab(74.29 60.8198 29.8997 / 0.9):

lab(from hsl(0 100% 50%) calc(l + 20) calc(a - 20) calc(b - 40) / calc(alpha - 0.1))

<lab()> = 
  lab( [ <percentage> | <number> | none ] [ <percentage> | <number> | none ] [ <percentage> | <number> | none ] [ / [ <alpha-value> | none ] ]? )  

<alpha-value> = 
  <number>      |
  <percentage>  

The following example shows the effect of varying the lightness value of the lab() function.

HTML

<div data-color="red-dark"></div>
<div data-color="red"></div>
<div data-color="red-light"></div>

<div data-color="green-dark"></div>
<div data-color="green"></div>
<div data-color="green-light"></div>

<div data-color="blue-dark"></div>
<div data-color="blue"></div>
<div data-color="blue-light"></div>

CSS

body {
  display: flex;
  flex-wrap: wrap;
  gap: 10px;
}
div {
  height: 50px;
  flex: 0 0 28%;
  border: 1px solid black;
}

[data-color="red-dark"] {
  background-color: lab(5 125 71);
}
[data-color="red"] {
  background-color: lab(40 125 71);
}
[data-color="red-light"] {
  background-color: lab(95 125 71);
}

[data-color="green-dark"] {
  background-color: lab(10% -120 125);
}
[data-color="green"] {
  background-color: lab(50% -120 125);
}
[data-color="green-light"] {
  background-color: lab(90% -120 125);
}

[data-color="blue-dark"] {
  background-color: lab(10 -120 -120);
}
[data-color="blue"] {
  background-color: lab(50 -120 -120);
}
[data-color="blue-light"] {
  background-color: lab(90 -120 -120);
}

Result

This example demonstrates the effects of setting the a and b values of the lab() function to the ends and midpoints of the a-axis, which goes from green (-125) to red (125) and the b-axis, which goes from yellow (-125) to blue (125).

HTML

<div data-color="redyellow"></div>
<div data-color="redzero"></div>
<div data-color="redblue"></div>

<div data-color="zeroyellow"></div>
<div data-color="zerozero"></div>
<div data-color="zeroblue"></div>

<div data-color="greenyellow"></div>
<div data-color="greenzero"></div>
<div data-color="greenblue"></div>

CSS

Using the CSS background-color property, we vary the a and b values of the lab() color function along the a-axis and b-axis, showing the effects of maximum, midpoint, and minimum values in each case.

body {
  display: grid;
  grid-template-columns: 1fr 1fr 1fr;
  gap: 10px;
}
div {
  height: 50px;
  flex: 0 0 28%;
  border: 1px solid black;
}

/* a-axis max, variable b-axis */
[data-color="redyellow"] {
  background-color: lab(50 125 125);
}
[data-color="redzero"] {
  background-color: lab(50 125 0);
}
[data-color="redblue"] {
  background-color: lab(50 125 -125);
}

/* a-axis center, variable b-axis */
[data-color="zeroyellow"] {
  background-color: lab(50 0 125);
}
[data-color="zerozero"] {
  background-color: lab(50 0 0);
}
[data-color="zeroblue"] {
  background-color: lab(50 0 -125);
}

/* a-axis min, variable b-axis */
[data-color="greenyellow"] {
  background-color: lab(50 -125 125);
}
[data-color="greenzero"] {
  background-color: lab(50 -125 0);
}
[data-color="greenblue"] {
  background-color: lab(50 -125 -125);
}

Result

The left column is at the yellow end (-125) of the b-axis and the right column is at the blue end (125). The top row displays colors at the red end of the a-axis (-125) and the bottom row is at the green end (125). The middle column and row are at the midpoints (0) of each axis, with the middle cell being grey; it contains no red, green, yellow, or blue, with a 0 value for both axes.

This example includes linear gradients to demonstrate the progression of values of the lab() function along the a-axis (from red to green) and along the b-axis (from yellow to blue). In each gradient image, one axis remains static while the other axis progresses from the low end to the high end of the axis values.

<div data-color="redtogreen-yellow">
  <span>red</span><span>`b`= -125 (yellow)</span><span>green</span>
</div>
<div data-color="redtogreen-zero">
  <span>red</span><span>no yellow or blue</span><span>green</span>
</div>
<div data-color="redtogreen-blue">
  <span>red</span><span>`b`= 125 (blue)</span><span>green</span>
</div>

<div data-color="yellowtoblue-red">
  <span>yellow</span><span>`a` = -125 (red)</span><span>blue</span>
</div>
<div data-color="yellowtoblue-zero">
  <span>yellow</span><span>no red or green</span><span>blue</span>
</div>
<div data-color="yellowtoblue-green">
  <span>yellow</span><span>`a` = 125 (green)</span><span>blue</span>
</div>

CSS

div {
  height: 50px;
  padding: 5px;
  margin: 5px;
  border: 1px solid black;
  display: flex;
  justify-content: space-between;
  align-items: center;
}
span {
  background-color: #ffffffcc;
  padding: 3px;
}

/* a-axis gradients */
[data-color="redtogreen-yellow"] {
  background-image: linear-gradient(to right, lab(50 125 125), lab(50 -125 125));
}
[data-color="redtogreen-zero"] {
  background-image: linear-gradient(to right, lab(50 125 0), lab(50 -125 0));
}
[data-color="redtogreen-blue"] {
  background-image: linear-gradient(to right, lab(50 125 -125), lab(50 -125 -125));
}

/* b-axis gradients */
[data-color="yellowtoblue-red"] {
  background-image: linear-gradient(to right, lab(50 125 125), lab(50 125 -125));
}
[data-color="yellowtoblue-zero"] {
  background-image: linear-gradient(to right, lab(50 0 125), lab(50 0 -125));
}
[data-color="yellowtoblue-green"] {
  background-image: linear-gradient(to right, lab(50 -125 125),lab(50 -125 -125));
}

Result

The following example shows the effect of varying the A (alpha) value of the lab() functional notation. The red and red-alpha elements overlap the #background-div element to demonstrate the effect of opacity. Giving A a value of 0.4 makes the color 40% opaque.

HTML

<div id="background-div">
  <div data-color="red"></div>
  <div data-color="red-alpha"></div>
</div>

CSS

div {
  width: 50px;
  height: 50px;
  padding: 5px;
  margin: 5px;
  display: inline-block;
  border: 1px solid black;
}
#background-div {
  background-color: lightblue;
  width: 150px;
  height: 40px;
}

[data-color="red"] {
  background-color: lab(80 125 125);
}
[data-color="red-alpha"] {
  background-color: lab(80 125 125 / 0.4);
}

Result

This example styles three <div> elements with different background colors. The middle one is given the unmodified --base-color, while the left and right ones are given lightened and darkened variants of that --base-color.

These variants are defined using relative colors — the --base-color custom property is passed into a lab() function, and the output colors have their lightness channel modified to achieve the desired effect via a calc() function. The lightened color has 15% added to the lightness channel, and the darkened color has 15% subtracted from the lightness channel.

<div id="container">
  <div class="item" id="one"></div>
  <div class="item" id="two"></div>
  <div class="item" id="three"></div>
</div>

CSS

#container {
  display: flex;
  width: 100vw;
  height: 100vh;
  box-sizing: border-box;
}

.item {
  flex: 1;
  margin: 20px;
}

:root {
  --base-color: orange;
  /* equivalent of lab(75 24 79) */
}

#one {
  background-color: lab(from var(--base-color) calc(l + 15) a b);
}

#two {
  background-color: var(--base-color);
}

#three {
  background-color: lab(from var(--base-color) calc(l - 15) a b);
}

Result

The output is as follows:

• <color> data type
• <color-function> data type
• Using relative colors
• CSS colors module
• LCH colors in CSS: what, why, and how?
• Safari Technology Preview 122 release notes: includes lab() and lch() colors