Mask (computing)
From Wikipedia, the free encyclopedia
In computer science, a mask is data that is used for bitwise operations.
Using a mask, multiple bits in a byte, nibble, word (etc.) can be set either on, off or inverted from on to off (or vice verse) in a single bitwise operation.
Contents |
[edit] Common bitmask functions
[edit] Masking bits to 1
To turn certain bits on, the bitwise OR operation can be used. Recall that Y or 1 = 1 and Y or 0 = Y. Therefore, to make sure a bit is on, OR can be used with a 1. To leave a bit alone, OR is used with a 0.
e.g.
10011101 10010101
OR 00001000 00001000
= 10011101 10011101
[edit] Masking bits to 0
As seen above, there is no way to change a bit from on to off using the OR operation. Instead, bitwise AND is used. When a value is ANDed with a 1, the result is simply the original value, as in: Y AND 1 = Y. However, AND with a value with 0, is guaranteed to get a 0 back so it is possible to turn a bit off by ANDing it with 0: Y AND 0 = 0. To leave the other bits alone, ANDing them with a 1 can be done.
The most common mask used, also known as a bitmask, extracts the status of certain bits in a binary string or number (a bit field or bit array). For example, to extract the status of the fifth bit from the binary string 10011101 it is possible to use a bitmask such as 00001000 and use the bitwise AND operator. Recalling that 1 AND 1 = 1, with 0 otherwise, finds the status of the fifth bit, since
10011101 10010101
AND 00001000 00001000
= 00001000 00000000
[edit] Querying the status of a bit
It is possible to use bitmasks to easily check the state of individual bits regardless of the other bits. To do this, turning off all the other bits using the bitwise AND is done as discussed above and the value is compared with 0. If it is, then the bit was off, but if the value is any other value, then the bit was on. What makes this convenient is that it is not necessary to figure out what the value actually is, just that it is not 0.
[edit] Toggling bit values
So far the article has covered how to turn bits on and turn bits off, but not both at once. Sometimes it is not really cared what the value is, it must be made the opposite of what it currently is. This can be achieved using the XOR (exclusive or) operation. XOR returns 1 if and only if an odd number of bits are 1. Therefore, if two corresponding bits are 1, the result will be a 0, but if only one of them is 1, the result will be 1. Therefore inversion of the values of bits is done by XORing them with a 1. If the original bit was 1, it returns 1 XOR 1 = 0. If the original bit was 0 it will get 0 XOR 1 = 1. Also note that XOR masking is bit-safe, meaning it will not affect unmasked bits because Y XOR 0 = Y, just like an OR.
[edit] Uses of bitmasks
[edit] Arguments to functions
In programming languages such as C, bit masks are a useful way to pass a set of named boolean arguments to a function. For example, in the graphics API OpenGL, there is a command, glClear() which clears the screen or other buffers. It can clear up to four buffers (the color, depth, accumulation, and stencil buffers), so the API authors could have had it take four arguments. But then a call to it would look like
glClear(1,1,0,0); // This is not how glClear actually works and would make for unreadable code.
which is not very descriptive. Instead there are four defined bit fields, GL_COLOR_BUFFER_BIT, GL_DEPTH_BUFFER_BIT, GL_ACCUM_BUFFER_BIT, and GL_STENCIL_BUFFER_BIT and glClear() is declared as
void glClear(GLbitfield mask);
Then a call to the function looks like this
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
Internally, a function taking a bitfield like this can use logical and to extract the individual bits. For example, an implementation of glClear() might look like:
void glClear(GLbitfield mask) { if (mask & GL_COLOR_BUFFER_BIT) { // Clear color buffer. } if (mask & GL_DEPTH_BUFFER_BIT) { // Clear depth buffer. } if (mask & GL_ACCUM_BUFFER_BIT) { // Clear accumulation buffer. } if (mask & GL_STENCIL_BUFFER_BIT) { // Clear stencil buffer. } }
While elegant, in the simplest implementation this solution is not type-safe. A GLbitfield is simply defined to be an unsigned int, so the compiler would allow a meaningless call to glClear(42) or even glClear(GL_POINTS). In C++ an alternative would be to create a class to encapsulate the set of arguments that glClear can accept, at no cost of performance, and could be cleanly encapsulated in a library (see the external links for an example).
[edit] Inverse Masks
Masks are used with IP addresses in IP ACLs (Access Control Lists) to specify what should be permitted and denied. Masks in order to configure IP addresses on interfaces start with 255 and have the large values on the left side, for example, IP address 209.165.202.129 with a 255.255.255.224 mask. Masks for IP ACLs are the reverse, for example, mask 0.0.0.255. This is sometimes called an inverse mask or a wildcard mask. When the value of the mask is broken down into binary (0s and 1s), the results determine which address bits are to be considered in processing the traffic. A 0 indicates that the address bits must be considered (exact match); a 1 in the mask is a "don't care". This table further explains the concept. Mask Example
network address (traffic that is to be processed) 10.1.1.0
mask 0.0.0.255
network address (binary) 00001010.00000001.00000001.00000000
mask (binary) 00000000.00000000.00000000.11111111
Based on the binary mask, it can be seen that the first three sets (octets) must match the given binary network address exactly (00001010.00000001.00000001). The last set of numbers are "don't cares" (.11111111). Therefore, all traffic that begins with 10.1.1. matches since the last octet is "don't care". Therefore, with this mask, network addresses 10.1.1.1 through 10.1.1.255 (10.1.1.x) are processed.
Subtract the normal mask from 255.255.255.255 in order to determine the ACL inverse mask. In this example, the inverse mask is determined for network address 172.16.1.0 with a normal mask of 255.255.255.0.
255.255.255.255 - 255.255.255.0 (normal mask) = 0.0.0.255 (inverse mask)
ACL equivalents
The source/source-wildcard of 0.0.0.0/255.255.255.255 means "any".
The source/wildcard of 10.1.1.2/0.0.0.0 is the same as "host 10.1.1.2"
[edit] Image masks
In computer graphics, when a given image is intended to be placed over a background, the transparent areas can be specified through a binary mask. This way, for each intended image there are actually two bitmaps: the actual image, in which the non used areas are given a pixel value with all bits set to 0's, and an additional mask, in which the correspondent image areas are given a pixel value of all bits set to 0's and the surrounding areas a value of all bits set to 1's. In the sample at right, black pixels have the all-zero bits and white pixels have the all-one bits.
At run time, to put the image on screen over the background, the program first masks the screen pixel's bits with the image mask at the desired coordinates using the bitwise AND operation. This preserves the background pixels of the transparent areas while reset with zeros the bits of the pixels which will be obscured by the overlapped image.
Then, the program renders the image pixels' bits by blending them with the background pixels' bits using the bitwise OR operation. This way, the image pixels are appropiately placed while keeping the background surrounding pixels preserved. The result is a perfect compound of the image over the background.
This technique is used for painting pointing devices' cursors; in typical 2-D videogames for characters, bullets and so on (the sprites); GUI icons; video titling and other image mixing applications.
Although related (due to being used for the same purposes), transparent colors and alpha channels are techniques which do not involve the image pixels' mixage by binary masking.
[edit] Hash tables
To create a hashing function for a hash table often a function is used that has a large domain. To create an index from the output of the function, a modulo can be taken to reduce the size of the domain to match the size of the array, however it is often faster on many processors to restrict the size of the hash table to powers of two sizes and use a bit mask instead.
