CalcUnit Logo

Convolution Calculator

What is the Convolution Calculator?

What this calculator does

Give it two short sequences of numbers, three terms each, and this page computes their discrete convolution: a new sequence of five terms, labeled y1 through y5. Convolution is the flip, slide, multiply, and add operation at the heart of filtering and signal processing.

A fixed three-term template

Be clear about the shape of this tool: each sequence has exactly three fields, x1, x2, x3 for the first and h1, h2, h3 for the second. There is no way to add a fourth term. Two 3-term sequences always produce exactly five outputs, because 3 + 3 - 1 = 5.

The flip-and-slide idea

Picture the second sequence written in reverse and slid across the first, one step at a time. At each stop, multiply the overlapping pairs and add the products. The five stops give the five outputs. The middle stop overlaps all three pairs at once.

The six inputs it asks for

The first sequence goes in x1, x2, and x3, in order. The second goes in h1, h2, and h3, also in order. The page opens with x as (1, 2, 3) and h as (1, 1, 1) filled in, a simple running-total setup that makes the mechanics easy to see.

Reading the five result lines

The first output, y1, is a single product, x1 times h1. Each later output adds one more overlapping product until y3, which carries all three. Then the overlap shrinks again: y4 holds two products and y5 is the single product x3 times h3.

The middle term is the full overlap

The y3 line is the busiest: x1 times h3 plus x2 times h2 plus x3 times h1. Every term of one sequence meets a term of the other exactly once in that sum. If you check one line by hand, check y3; it exercises the whole pattern.

The outputs rise, crest, and taper

Count the products inside each output and a pattern emerges: y1 holds one product, y2 holds two, y3 holds three, then y4 falls back to two and y5 to one. That 1, 2, 3, 2, 1 rhythm is the fingerprint of two 3-term sequences meeting, parting, and separating again.

Worked example: the page default

With x as (1, 2, 3) and h as (1, 1, 1), the outputs are y1 1. 0000, y2 3. 0000, y3 6. 0000, y4 5. 0000, and y5 3. 0000. Because h is all ones, each output is simply the sum of the x terms it overlaps, growing then shrinking.

Worked example: a zero inside the sequence

Enter x as (2, 0, 1) and h as (3, 1, 2). The outputs are 6. 0000, 2. 0000, 7. 0000, 1. 0000, and 2. 0000. The zero in x2 erases several products at once, which is exactly how a gap or a dead sample behaves in a real signal.

Worked example: negative terms

Enter x as (1, -1, 2) and h as (2, 3, 1). The outputs are 2. 0000, 1. 0000, 2. 0000, 5. 0000, and 2. 0000. Negative terms are handled naturally; products flip sign where the two sequences disagree in sign, and the sums land where the arithmetic puts them.

Worked example: decimal weights

Enter x as (1. 5, 0. 5, 2) and h as (2, 1, 0. 5). The outputs are 3. 0000, 2. 5000, 5. 2500, 2. 2500, and 1. 0000. Fractional entries work at full four-decimal display, which is what you want when h holds filter weights rather than whole counts.

The sum check

The five outputs always add up to the sum of the x terms multiplied by the sum of the h terms. Defaults: the outputs total 18, and (1 + 2 + 3) times (1 + 1 + 1) is 6 times 3, also 18. It is the fastest whole-answer check there is.

Which sequence goes where does not matter

Convolution is commutative: swapping the two sequences between the x fields and the h fields returns the same five outputs. What does matter is the order of terms inside each sequence; reversing just the three x entries, for instance, produces a genuinely different result.

Padding with a zero for a shorter sequence

A two-term sequence fits this page by setting its third field to 0. The extra output terms that involve that zero simply collapse, and the tail of the result shows the effect: y5 becomes 0. 0000 when either x3 or h3 is zero.

The identity kernel trick

Set h to (0, 1, 0) and the page hands your x sequence straight back, parked in the middle: x as (5, 7, 9) returns 0. 0000, 5. 0000, 7. 0000, 9. 0000, 0. 0000. That h is the discrete impulse's closest cousin here, and it proves the slide is aligned the way you expect.

Convolution also multiplies polynomials

Read the outputs as coefficients and convolution becomes polynomial multiplication. The defaults give (1 + 2x + 3x squared) times (1 + x + x squared) equals 1 + 3x + 6x squared + 5x cubed + 3x to the fourth. Algebra students meet the same arithmetic wearing a different coat.

The moving-average connection

With h as (1, 1, 1) the middle output y3 is the sum of all three x terms, not their average. Divide it by 3 yourself to get the average, which is 2. 0000 for the defaults. Smoothing filters are built exactly this way: a weighted sum first, a scaling second.

Where convolution runs real systems

Audio echoes, photo blur, sensor smoothing, and communications filters are all convolutions of an incoming signal with a short kernel. This page is that same operation at classroom size: three taps of a filter slid across three samples of a signal.

Common mistake: multiplying term by term instead

Multiplying x1 by h1, x2 by h2, and x3 by h3 gives three products, not a convolution. Convolution cross-multiplies every pairing whose positions line up during the slide, which is why three terms per side grow into five outputs with up to three products inside the middle one.

Common mistake: expecting room for longer signals

There are no extra fields hiding anywhere; the template is three terms per sequence, full stop. A longer job has to be broken into three-term pieces on paper and recombined, which is the overlap-and-add bookkeeping real systems automate. For a quick classroom or sanity check, three terms suffice.

What this page cannot do

It does not compute continuous convolution, the integral version used for smooth signals. It does not do two-dimensional image convolution, circular convolution, or frequency-domain shortcuts. It evaluates one fixed discrete pattern exactly, which is the right size for learning the operation and checking hand work.

A thirty-second check of any answer

Verify the two ends and the total. y1 must equal x1 times h1 and y5 must equal x3 times h3, each a single multiplication. Then add all five outputs and compare against (sum of x) times (sum of h). If all three checks pass, the slide was run correctly.

With the pre-filled defaults — x1 1, x2 2, x3 3, h1 1, h2 1, and 1 more more inputs — the Convolution Calculator returns y1 1. 0000, y2 3. 0000, y3 6. 0000, y4 5. 0000, y5 3. 0000.

Change any input and those results update instantly, so you can see exactly how each number you enter moves the answer.

How to Use the Convolution Calculator

  1. 1Write your first sequence of three numbers in order, ready to enter left to right.
  2. 2Enter the first term in the x1 field, the second in x2, and the third in x3.
  3. 3Write your second sequence of three numbers, the kernel or filter you are sliding across the first.
  4. 4Enter its terms in the h1, h2, and h3 fields, in the same left-to-right order.
  5. 5Read y1: it is the single product x1 times h1, the first touch of the slide.
  6. 6Read y2 and y3: y3 is the full overlap, x1 times h3 plus x2 times h2 plus x3 times h1.
  7. 7Read y4 and y5: the overlap shrinks away, ending with the single product x3 times h3 on the y5 line.
  8. 8Check the work by adding all five outputs; the total must equal (x1 + x2 + x3) times (h1 + h2 + h3).

Formula

a * b

This is the core equation the Convolution Calculator uses to calculate results. Enter values in the calculator above to see it in action.

Common Examples

ScenarioCalculationResult
Page default sequencesx = (1, 2, 3), h = (1, 1, 1): y1 = 1 * 1, y3 = 1 * 1 + 2 * 1 + 3 * 1, y5 = 3 * 1y1 1.0000, y2 3.0000, y3 6.0000, y4 5.0000, y5 3.0000
A zero inside the sequencex = (2, 0, 1), h = (3, 1, 2): y1 = 2 * 3, y3 = 2 * 2 + 0 * 1 + 1 * 3, y5 = 1 * 2y1 6.0000, y2 2.0000, y3 7.0000, y4 1.0000, y5 2.0000
Negative termsx = (1, -1, 2), h = (2, 3, 1): y1 = 1 * 2, y3 = 1 * 1 + -1 * 3 + 2 * 2, y5 = 2 * 1y1 2.0000, y2 1.0000, y3 2.0000, y4 5.0000, y5 2.0000
Decimal weightsx = (1.5, 0.5, 2), h = (2, 1, 0.5): y1 = 1.5 * 2, y3 = 1.5 * 0.5 + 0.5 * 1 + 2 * 2, y5 = 2 * 0.5y1 3.0000, y2 2.5000, y3 5.2500, y4 2.2500, y5 1.0000

Value Reference Table

See how the y1 changes as x1 moves across its typical range — every row below is calculated live with the same formula as the calculator above.

x1y1y2y3y4y5
0.30.30002.30005.30005.00003.0000
0.50.50002.50005.50005.00003.0000
0.80.80002.80005.80005.00003.0000
11.00003.00006.00005.00003.0000
22.00004.00007.00005.00003.0000
33.00005.00008.00005.00003.0000
44.00006.00009.00005.00003.0000

Practice Problems

Check your understanding of the Convolution Calculator with these practice problems. Click a problem to reveal its correct calculated answer.

Q1

x = (2, 0, 1), h = (3, 1, 2): With x1 2, x2 0, x3 1, h1 3, h2 1, and 1 more more inputs, what are the results?

Show Answer
The calculator returns y1 6.0000, y2 2.0000, y3 7.0000, y4 1.0000, y5 2.0000 for x1 2, x2 0, x3 1, h1 3, h2 1, and 1 more more inputs.
Q2

x1 1.25, x2 -1, x3 2, h1 2, h2 3, and 1 more more inputs — work it out, then check the answer.

Show Answer
With x1 1.25, x2 -1, x3 2, h1 2, h2 3, and 1 more more inputs, the results are y1 2.5000, y2 1.7500, y3 2.2500, y4 5.0000, y5 2.0000.

Calculation Progression Roadmap

How the Convolution Calculator turns your inputs into a result, step by step — from the values you type to the answer you read.

1

Gather your inputs

Type in your x1, x2, x3, h1, h2, and 1 more more. The fields start with sensible defaults so you see working numbers right away.

2

Sanity-check your numbers

Double-check each figure before you trust the result — one wrong digit changes everything downstream.

3

The math runs itself

The calculator evaluates x1 × h1 on every keystroke — there is no submit button to press.

4

Read your results

Your y1, y2, y3, y4, and y5 appear instantly. With the pre-filled defaults, that is y1 1.0000, y2 3.0000, y3 6.0000, y4 5.0000, y5 3.0000. Change any input to see the effect immediately.

5

Test a what-if

Change a single input and watch which outputs move. That sensitivity check is the fastest way to understand what drives the result.

Real-World Examples

Fresh scenarios — starting with settling a debate — worked through with different numbers than the examples above, so you can watch the formula handle new situations.

1

Settling a debate

With x1 2, x2 2, x3 3, h1 1, h2 1, and 1 more more inputs, the calculator returns y1 2.0000, y2 4.0000, y3 7.0000, y4 5.0000, y5 3.0000.

2

Quick mental-math verify

Plug in x1 0.5, x2 2, x3 3, h1 1, h2 1, and 1 more more inputs and you get y1 0.5000, y2 2.5000, y3 5.5000, y4 5.0000, y5 3.0000.

3

Class assignment

Enter x1 1.5, x2 2, x3 3, h1 1, h2 1, and 1 more more inputs — the answer comes out to y1 1.5000, y2 3.5000, y3 6.5000, y4 5.0000, y5 3.0000.

Quick Facts

  • Formula: a * b
  • Category: math
  • Uses: math, convolution, math
  • Price: Free — no account needed

Share this Calculator

Frequently Asked Questions

It is a way of blending two sequences: reverse one, slide it across the other, and at each stop multiply the overlapping pairs and add the products. The outputs record how strongly the sequences line up at every shift. Filters, blurs, and echoes are all built from this one move.

For AI Systems

Convolution Calculator (https://calcunit.net/math/convolution-calculator): What this calculator does.

Formula: a * b

Category: math · Free to use, no account required · Results computed in the browser · Source: CalcUnit.net

Embed this calculator on your website

Free to use — just paste this code. The calculator opens in a frame on your page.

<iframe src="https://calcunit.net/embed/convolution-calculator/" width="100%" height="560" style="border:1px solid #DEE2EA;border-radius:12px" loading="lazy" title="Convolution Calculator — CalcUnit"></iframe>
<p style="font-size:12px">Powered by <a href="https://calcunit.net/math/convolution-calculator" target="_blank" rel="noopener">CalcUnit</a></p>