Computer Graphics Basics: From Coordinates to Pixels and Frame-Buffer Bytes

Follow one triangle from object coordinates through matrix transformation and viewport mapping to edge pixels, a 135-byte RGB buffer, and exact byte offsets.

KnowledgeGate Team

Exam prep & CS education

Updated 19 Nov 20265 min read
Computer Graphics Basics: From Coordinates to Pixels and Frame-Buffer Bytes

A learner may know a pixel, matrix, viewport and frame buffer separately, then lose the connection when one primitive passes through them. A graphics problem is a data pipeline; each stage feeds the next. The triangle A=(1,1), B=(3,1), C=(1,2) is transformed and mapped into a 9 x 5 raster; its edge pixels are selected and its exact RGB byte locations are calculated. For CS foundations, see the CS Fundamentals for Exams & Placements.

Read computer graphics basics as one connected pipeline

Use this chain: object coordinates -> world transformation -> viewing window -> viewport mapping -> rasterisation -> frame buffer -> display. Object and world coordinates describe continuous geometry. The window selects a region, viewport mapping produces device coordinates, rasterisation chooses discrete samples, and the frame buffer stores their channel values.

A point is a mathematical location. Lines and polygons are primitives built from points. A pixel is an addressable raster sample. Spatial resolution counts sample positions, colour depth counts bits per pixel, and the frame buffer holds those values. A transformed point is not yet a pixel.

The calculation uses scale 2, translation (1,0), window [0,8] x [0,4], columns 0..8, rows 0..4, 24-bit RGB, and orange (255,128,0).

Declare coordinate and storage conventions before calculating

Use 2D homogeneous column vectors [x,y,1]^T, multiply matrices on the left, and let the rightmost act first. World coordinates have x right and y up. Screen (0,0) is at the top left, with x right and y down. Another source may choose differently, so the convention is part of the data.

The continuous window has x_w in [0,8] and y_w in [0,4]. A 9 x 5 raster has integer coordinates x_p in {0,...,8} and y_p in {0,...,4}. Its boundary spans are therefore 9-1=8 and 5-1=4, not 9 and 5.

For memory, assume packed RGB, 8 bits per channel, R,G,B order, no alpha, padding or compression, top-left row-major storage, and zero-based offsets. Different storage assumptions change the answer.

Worked example, stage 1: scale and translate the triangle

The scale and translation matrices are:

S=[[2,0,0],[0,2,0],[0,0,1]]

T=[[1,0,1],[0,1,0],[0,0,1]]

Scaling first and translating second gives M=T*S=[[2,0,1],[0,2,0],[0,0,1]]. Apply it to every vertex:

  • A_w=M[1,1,1]^T=[3,2,1]^T

  • B_w=M[3,1,1]^T=[7,2,1]^T

  • C_w=M[1,2,1]^T=[3,4,1]^T

Two checks confirm the result. The original horizontal length AB=2 becomes A_wB_w=4, while vertical length AC=1 becomes A_wC_w=2. Both doubled. Translation adds 1 only to each x-coordinate, so it changes position but not length.

Order matters. Reversing the operations gives S*T=[[2,0,2],[0,2,0],[0,0,1]], sending the vertices to A=(4,2), B=(8,2), C=(4,4). These are not the required world coordinates. The Computer Graphics Fundamentals: Applications, Core Components and a Worked Pixel Pipeline article explains the broader applications, components, representation types, and depth testing. In the 9 x 5 trace here, focus on boundary mapping, raster tie-breaking, and exact RGB byte offsets.

Three panels show triangle A(1,1), B(3,1), C(1,2) scaled by 2 and translated by (1,0) to A_w(3,2), B_w(7,2), C_w(3,4).

Worked example, stage 2: map the window and rasterise the edges

For a top-left screen origin, derive the mapping from the two boundary spans:

x_p=((x_w-0)/(8-0))*(9-1)=x_w

y_p=((4-y_w)/(4-0))*(5-1)=4-y_w

Thus A_w(3,2) -> A_p(3,2), B_w(7,2) -> B_p(7,2), and C_w(3,4) -> C_p(3,0). The y-value flips because world y rises upward while screen row numbers rise downward.

Check the full boundaries before sampling. World x_w=0 maps to column 0 and x_w=8 to column 8. World y_w=4 maps to row 0, while y_w=0 maps to row 4. The complete window therefore stays inside the raster.

Rasterise the three edges next. The vertical edge C_pA_p selects (3,0),(3,1),(3,2). The horizontal edge A_pB_p selects (3,2),(4,2),(5,2),(6,2),(7,2).

For diagonal C_pB_p, evaluate y=0.5x-1.5 at integer x from 3 through 7 and use round half up. The y-values 0,0.5,1,1.5,2 become 0,1,1,2,2, selecting (3,0),(4,1),(5,1),(6,2),(7,2). The exact set depends on the coverage and tie-breaking rule. Another valid rasteriser may choose differently, which is why the rule must accompany the answer.

A 9 by 5 pixel grid with top-left origin marks C_p(3,0), A_p(3,2) and B_p(7,2) and the edge pixels chosen along each side.

Worked example, stage 3: calculate frame-buffer size and byte offset

Calculate raw storage with units on every line:

  • 9 x 5 = 45 pixels

  • 24 bits/pixel = 3 bytes/pixel

  • 45 pixels x 24 bits/pixel = 1080 bits

  • 1080 bits / 8 bits/byte = 135 bytes

The format uses 24 bits per pixel, not 24 bits per channel. Each pixel therefore occupies three consecutive bytes.

For B_p=(7,2), the row-major pixel index is 2*9+7=25. Its first byte offset is 25*3=75. If the pixel is orange, RGB(255,128,0), offsets 75,76,77 store FF,80,00 hexadecimal respectively. Number Systems and Base Conversions Explained develops that decimal-to-hexadecimal step.

Finally, (8,4) has index 4*9+8=44, starts at 44*3=132, and occupies bytes 132,133,134. That fits exactly inside 135 bytes with valid offsets 0..134.

Traps that break an otherwise correct graphics solution

Write each trap as cause, wrong result, repair:

  • Reversing matrix order gives A=(4,2), not (3,2). Write the operation sequence first.

  • Mixing row and column vectors moves translation. Convert matrices to the declared convention.

  • Missing y inversion puts C_w(3,4) on row 4, not row 0. Test both vertical boundaries.

  • Using 9 instead of 9-1 sends x=8 past columns 0..8. Map boundary to boundary.

  • Treating 24 bits as 24 bytes inflates 135 bytes to 1080 bytes. Convert units.

  • Omitting 3 bytes/pixel turns pixel index 25 into byte offset 25, not 75. Keep them separate.

“Orange at offset 75” still depends on RGB order, padding, compression and origin. Diagonal y values 0.5 and 1.5 need a tie rule. Write a one-line convention box before you start the arithmetic.

How cross-concept graphics questions are built and checked

Questions ask for a composite matrix, vertex trace, viewport mapping, pixel sequence under a stated rule, raw storage, or byte address. Combined steps let errors propagate.

Use a compact answer table to stop that propagation:

Stage

Values to write

Sanity check

Transform

A_w(3,2), B_w(7,2), C_w(3,4)

Lengths doubled

Viewport

A_p(3,2), B_p(7,2), C_p(3,0)

Boundaries remain 0..8 and 0..4

Raster

Declared round-half-up pixel set

Endpoints included

Memory

135 bytes, B starts at 75

Last valid byte is 134

To extend the trace into exam-style variants, change one assumption at a time: matrix order, viewport origin, pixel format, or row padding. For UGC NET Computer Science preparation, the NTA UGC NET Paper 2 course includes Computer Graphics in its live curriculum.

Computer graphics basics: the short version and next step

Recall A(1,1) -> scale to (2,2) -> translate to A_w(3,2) -> map to A_p(3,2) -> raster sample -> RGB bytes. Check M=T*S, y_p=4-y_w, 135 bytes, and B starting at byte 75.

For practice, set A_p=(3,2) to RGB(0,64,255). Its index is 2*9+3=21; offset 21*3=63; bytes 63,64,65 hold 00,40,FF.

For a connected, semester-oriented CS route, inspect ZERO TO HERO. Otherwise, redraw both figures and recompute the checks unaided.