ELEC1010 Chap.4 Filters and Signal Selection
Filters and Signal Selection
Multiply at each supplied frequency
A filter response describes how a system treats each frequency. In the course model, output spectral amplitude at a frequency equals input amplitude there multiplied by the response gain at the same frequency. The calculation is local: align an input stem with the response, read one gain, multiply, and repeat.
The smooth response curve is an operator and should never be copied as though it were the output signal.
Gain zero removes a supplied component, gain one preserves its amplitude, and a value between zero and one attenuates it. A value above one amplifies it. A nonzero response where the input has no component still produces zero, because zero input multiplied by gain remains zero.
This observation prevents invented output lines in regions where the filter could pass a signal but none was supplied.
Low-pass, high-pass and band-pass labels summarize which frequency regions are retained. A low-pass response favours frequencies below its cutoff, a high-pass response favours those above, and a band-pass response retains a bounded interval.
Boundary wording matters: through a stated frequency includes that point, whereas below it does not. If a transition gain is provided, use the stated value instead of forcing an ideal zero-or-one model.
Ordinary filtering changes amplitude and possibly phase but does not move a component to a new frequency in this treatment. Horizontal translation belongs to modulation.
After calculating an output, compare its set of frequencies with the input set; every surviving line should occupy an original location. A shifted or newly created line signals that the response curve or another operation was misread.
Units provide an independent audit. If the input is a voltage spectrum, the output amplitudes remain voltages, while gain is dimensionless.
Copying gain values as output amplitudes loses the input scale and often loses the unit. Use a four-column table for frequency, input, gain and output, then cross out rows whose product is zero only after the multiplication is recorded.
To revise, solve examples with partial gains, passband edges and unequal input strengths. State both survivors and rejected components, tying each result to its local product.
Then compare two lines where the larger input has the smaller gain. Such cases defeat intuition based on either picture alone and reinforce the central rule that only input multiplied by response determines the output spectrum.
What this chapter covers
- 01
Filter response
- 02
Pointwise spectral gain
- 03
Low-pass selection
- 04
High-pass selection
- 05
Band-pass selection
- 06
Output spectrum audits
Worked example · free
Apply a response line by line
- 2Align each input with its local gain.
- 4Multiply the four pairs.
- 2Keep nonzero output lines only.
Key terms
- Filter Response
- The gain assigned by a filter as a function of frequency.
- Filter Gain
- The multiplier applied to an input component at one frequency.
- Low-pass Filter
- A filter retaining low frequencies and suppressing higher ones.
- High-pass Filter
- A filter retaining high frequencies and suppressing lower ones.
- Band-pass Filter
- A filter retaining one bounded frequency interval.
Filters and Signal Selection FAQ
Does a filter move spectral lines?
Not in the taught pointwise model. Ordinary filtering changes component amplitude through gain but leaves each surviving line at its original frequency. Translation belongs to modulation.
Can gain create an output without an input line?
No. Output is input multiplied by gain. A nonzero response where the input amplitude is zero still produces zero and does not create a new component.
Why are the gain values not the output amplitudes?
Gain is a dimensionless multiplier. The signal output retains its amplitude unit and equals the original amplitude times gain, so copying the response values discards the input strengths.
How should a passband boundary be handled?
Use the inclusion wording or response value supplied by the question. Through a boundary includes it, below excludes it, and a plotted transition requires the local gain rather than an assumed ideal value.
Exam move
Redraw every filtering problem as a table with frequency, input amplitude, gain and output amplitude. Check that the output-frequency set is a subset of the input set. Include partial gains and edge frequencies so the method handles attenuation and boundary wording rather than only ideal zero-or-one responses. Audit the output in two passes. The numerical pass checks each multiplication and preserves units.
The structural pass checks that every nonzero output frequency already existed at the input and that no zero-input location became a new component. These passes catch different mistakes: correct locations can still carry copied gain heights, while correct products can be drawn at shifted frequencies. At a boundary, quote the wording or graph value that set the gain rather than silently choosing inclusion.
Finally describe the filter family from its response, not from the accidental contents of one supplied input.
Working through Filters and Signal Selection in ELEC1010? Sia is AskSia’s AI Electrical Engineering tutor — ask any ELEC1010 Filters and Signal Selection question and get a clear, step-by-step explanation grounded in how ELEC1010 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.