1. Recording, mixing and mastering
A microphone captures the vibrations in the air, much as your ears do, and turns them into an electrical signal. Where you place it changes how much of the instrument and room it picks up. Mixing brings the separate recordings together and sets how loud, bright and left or right each part is. Mastering works with the finished mix to prepare the release and can change its overall tone and loud-and-quiet contrast.
Dynamic range is about the difference between quieter and louder levels, but always check which two levels are being compared. In a song, it can describe musical contrast. In electronics, a specification often compares the strongest clean signal with background noise. Clipping happens when a device runs out of room and cuts off the signal's peaks.
Follow recording, mixing and mastering in the Sound Journey.
2. Digital audio and the source
Sample rate tells you how many times per second the digital system records the wave's level for each channel. At 48 kHz, that's 48,000 readings per second, enough in principle for frequencies below 24 kHz when properly filtered. Bit depth tells you how finely each reading can represent that level. More bits mean finer steps and potentially less rounding noise, not automatically a better performance or master.
Lossless compression reconstructs the original encoded audio data. Lossy encoding discards or approximates information to reduce size. Audible differences depend on the encoder, settings, material and listener. Compare versions derived from the same master if you want to isolate the codec.
A DAC (digital-to-analogue converter) turns stored digital sound back into an electrical audio signal for an amplifier. If you're worried the input is overloaded, compare how strong the source signal can be with how much the amplifier input can handle cleanly. An input's sensitivity number alone is not that maximum limit.
3. Amplification and headroom
An amplifier provides the electrical push and current that speakers or headphones need. Their impedance is the changing electrical opposition they present to it; the printed ohm number is only a guide. Headroom is the extra clean output available when a sudden drum hit rises above the average music level.
Worked example. If an amplifier supplies 20 V RMS into an ideal 8 Ω resistor, it delivers 50 W (20² ÷ 8). The same voltage into an ideal 4 Ω resistor requires 100 W and twice the RMS current, 5 A instead of 2.5 A. This calculation assumes the amplifier can sustain 20 V into either load. A real loudspeaker's impedance and phase vary with frequency.
Read a wattage rating alongside its load, distortion threshold, frequency range and number of driven channels. For a distortion figure such as THD+N, check test level, frequency and measurement bandwidth before comparing equipment.
4. Speakers, headphones and the room
Frequency response is a graph of how strongly a speaker or headphone plays low, middle and high notes in a particular test. Sensitivity tells you how much sound it produces from a stated electrical input. When comparing two models, check that the input, measurement method and, for speakers, distance match. Headphones may be rated per volt or per milliwatt, which are not the same reference.
Dispersion describes how a speaker's output changes with angle and frequency. A seat off to the side receives a different angle of direct sound. The walls also receive sound from off-axis directions and reflect some of it back. For headphones, fit and seal can strongly affect the sound at the ear, particularly bass response.
When someone describes imaging, they usually mean how precisely voices or instruments appear to occupy positions. Soundstage describes the apparent extent of that scene. The recording, speaker setup, reflections and listener all contribute.
5. Check what a claim can tell you
- Read the conditions. Check the units, reference level, load, angle, distance and bandwidth as applicable.
- Identify the evidence. A specification states a result or limit under defined conditions. A measurement records what happened in one setup. A model shows what follows from its assumptions. A listening impression describes what someone experienced.
- Match the comparison to the question. If you are testing a codec, use the same master and match playback levels. A blind ABX trial asks whether you can identify a hidden sample as A or B under those conditions. Preference is a separate question.
Sources and further reading
- KLIPPEL Near Field Scanner documentation explains angular loudspeaker measurements and derived summaries.
- Spinorama measurement guide explains what loudspeaker response plots can tell you.
- The linked Decoder lessons provide worked examples and topic-specific sources for digital audio, amplifiers and transducers.