HiFi Audio Glossary

This glossary explains common HiFi and digital-audio terms in practical language. It is written for ECHO users who want to understand what they are hearing, what a setting changes, and how to troubleshoot without turning audio into a numbers contest.

The most important rule is simple:

Terms should help you listen, adjust, and diagnose. They are not proof that a setting is automatically better.

If playback is unstable, return to System or WASAPI Shared, disable EQ, DSP, ReplayGain, speed changes, FIR, exclusive mode, ASIO, and DSD, then test a normal MP3 or FLAC first.

Quick Index

Topic	Start with
Listening impressions	Resolution, separation, imaging, soundstage, layering, dynamics, transient response
File formats	PCM, DSD, FLAC, WAV, sample rate, bit depth, bitrate, lossless, Hi-Res
Output path	WASAPI, ASIO, exclusive mode, bit-perfect, resampling, cables, Bluetooth codec, latency
Hardware	DAC, amp, impedance, sensitivity, power, noise floor, SNR, distortion, crosstalk
DSP	EQ, preamp, headroom, FIR, IIR, phase, compressor, limiter, crossfeed

Listening Terms Are Not Measurements

Words such as resolution, separation, imaging, and black background describe perception. They are useful, but they are not single-number measurements.

For fair comparison:

Use tracks you know well.
Change one setting at a time.
Match volume as closely as possible.
Separate “more exciting” from “more accurate.”
Remember that a change can suit one track and hurt another.

Resolution

Resolution describes how clearly a system reveals detail: breaths, string texture, reverb tails, drum brush noise, or quiet backing parts.

Good resolution is not the same as boosted treble. A device or EQ curve can fake detail by emphasizing upper mids and treble, but that often becomes sharp or tiring.

Healthy resolution usually means:

Small details are audible but not forced forward.
Treble is clear without becoming sharp.
Reverb tails and weak sounds appear naturally.
Complex passages remain readable.

Separation

Separation describes how easily you can distinguish different instruments or voices when they play together. Poor separation makes the mix sound smeared or congested.

It is affected by recording quality, distortion, crosstalk, excessive bass, room acoustics, and DSP choices.

Before blaming the player, try lowering excess bass, disabling stereo enhancement, checking clipping risk, and comparing with a clean recording.

Imaging

Imaging describes whether sounds have stable positions in the stereo field. A centered vocal should stay centered; instruments should feel placed rather than floating randomly.

Bad imaging can come from channel imbalance, reversed left/right channels, poor fit, room reflections, aggressive spatial DSP, or the recording itself.

In ECHO, use channel balance, mono checks, and left/right swap tools when the image feels off-center.

Soundstage

Soundstage is the perceived size and shape of the listening space: width, depth, height, and distance.

Bigger is not always better. Overdone virtual space can make vocals hollow or distant. A smaller but stable stage can be more natural than a huge artificial one.

Soundstage depends on the recording, headphones or speakers, room reflections, channel matching, crossfeed, HRTF, and EQ.

Layering

Layering describes front-to-back order and musical priority. Separation is about whether sounds are distinct; layering is about whether they sit in a believable order.

Layering suffers when bass or low mids mask other content, dynamics are over-compressed, volume is too high, or spatial processing is unnatural.

Dynamics

Dynamics describe changes from quiet to loud. Macro-dynamics are large swings such as drum hits or orchestral climaxes. Micro-dynamics are small expressive changes such as vocal phrasing or piano touch.

Over-compressed masters, clipping, weak amplification, heavy limiting, and aggressive loudness processing can all flatten dynamics.

Transient Response

Transient response describes how cleanly sound starts and stops: drum hits, plucked strings, piano attacks, electronic kicks.

Fast transients can sound lively and rhythmic. Too much upper-mid or treble emphasis can fake speed while becoming sharp.

Tone Words

Word	Meaning	Possible downside
Warm	Fuller lows and low mids, smoother highs	Muddy or slow
Cool	Leaner lows, clearer edges	Thin or hard
Bright	More upper mids or treble	Sharp or sibilant
Dark	Reduced treble, softer tone	Veiled or lacking air

These are preferences, not rankings.

Sibilance

Sibilance is sharpness around vocal s, sh, z, and similar sounds. It often lives somewhere around 5 kHz to 10 kHz, but the exact area depends on the voice and system.

It can come from the recording, headphone peaks, EQ, high volume, compression, or exciters. Fix it gently; do not remove all treble just to hide sibilance.

Noise Floor And Black Background

Noise floor is the background hiss, hum, or electrical noise under the signal. A “black background” means quiet passages feel clean and free of noise.

Noise can come from the DAC, amp gain, sensitive IEMs, USB power, ground loops, drivers, or the recording itself. If hiss remains while playback is paused, the hardware chain is likely involved.

PCM

PCM is the most common digital-audio representation. It stores audio as a sequence of samples.

Common examples:

44.1 kHz / 16-bit: CD standard.
48 kHz / 24-bit: common for video, system audio, and production.
96 kHz / 24-bit or 192 kHz / 24-bit: common Hi-Res PCM.

PCM is not inferior. Most modern production and playback processing uses PCM somewhere in the chain.

DSD

DSD is a different digital-audio format, commonly seen as DSD64, DSD128, or DSD256.

Important points:

DSD playback depends heavily on the DAC and driver.
Unsupported paths may convert it to PCM.
DoP packages DSD in a PCM-like transport frame; it is not ordinary PCM playback.
A DSD file is not guaranteed to come from a native DSD recording chain.
If DSD causes problems, return to PCM and verify basic playback.

FLAC, WAV, ALAC, APE

Format	Type	Notes
WAV	Usually uncompressed PCM container	Large files, simple structure
FLAC	Lossless compression	Common, smaller, good metadata support
ALAC	Apple lossless	Common in Apple ecosystems
APE	Lossless compression	Older libraries, sometimes less convenient

Lossless means the decoded PCM can match the source data. It does not guarantee a good master.

MP3, AAC, Opus

These are lossy codecs. They discard information to save space.

Lossy does not always mean bad. A high-quality AAC, Opus, or MP3 encode from a good master can sound better than a bad or fake lossless file.

Sample Rate

Sample rate is how many samples are stored per second. 44.1 kHz means 44100 samples per second.

Higher sample rates can be useful for production, processing, and filter design, but they do not automatically improve everyday listening. Very high Windows default formats can increase load, trigger driver issues, or resample everything without improving quality.

Bit Depth

Bit depth describes how much numerical precision each sample has. Common values are 16-bit, 24-bit, and 32-bit float.

It mainly affects dynamic range and processing headroom. It does not directly mean more bass or treble.

Bitrate

Bitrate is the amount of data used per second, usually shown as kbps or Mbps.

For lossy codecs, higher bitrate usually preserves more information. For lossless codecs, bitrate often reflects musical complexity and compression efficiency. Do not compare different codecs by number alone.

Hi-Res

Hi-Res usually means audio above CD specification, such as 24-bit / 96 kHz. It is a specification label, not a guarantee of better sound.

The master, source, playback chain, volume matching, and device support matter more than the badge.

Bit-Perfect

Bit-perfect means the player tries to send the audio samples unchanged: no EQ, no volume change, no resampling, and no DSP.

It is useful for verifying the output path. It is not always the best listening choice. If you need headphone correction, room correction, channel balance, ReplayGain, or EQ, you are intentionally changing the samples.

In ECHO, EQ, FIR, channel tools, headphone correction, resampling, and ReplayGain can remove bit-perfect candidate status. That is expected.

Resampling, Upsampling, Oversampling

Term	Meaning
Resampling	Convert one sample rate to another
Upsampling	Convert to a higher sample rate
Downsampling	Convert to a lower sample rate
Oversampling	Internal high-rate processing, often inside a DAC

Resampling quality depends on the algorithm. Windows shared mode often resamples multiple apps to one device format. WASAPI Exclusive can reduce this, but stability comes first.

Jitter

Jitter is timing error in digital-audio clocks. It is real, but often overused as a marketing explanation.

Modern competent DACs usually control jitter well through buffering, reclocking, and asynchronous USB. For most users, file quality, output stability, clipping, noise, and device matching are higher-priority problems.

DAC

A DAC converts digital audio into an analog signal. Look for stable drivers, the formats you actually need, suitable output level, low noise, low distortion, and good device matching.

Beautiful specs do not help much if the driver is unstable.

Do Not Buy A DAC By Chip Alone

Many buyers start with the DAC chip name: ESS, AKM, Cirrus Logic, ROHM, TI/Burr-Brown, and so on. The chip matters, but it is only one part of the complete device.

The same DAC chip can sound and behave differently in different products because the final result also depends on clocking, power supply, I/V conversion, filtering, analog output stage, headphone amp stage, USB receiver, drivers, firmware, PCB layout, grounding, shielding, and gain structure. A chip datasheet describes potential under controlled conditions. The finished unit still has to implement it well.

When choosing a DAC, look at:

Area	Why it matters
Driver stability	Windows compatibility, WASAPI / ASIO behavior, sleep/wake reliability
Outputs	RCA, XLR, coaxial, optical, USB, Bluetooth, or headphone output
Output level	Too low may not drive the next device well; too high may overload an input
Noise floor	Sensitive IEMs can reveal hiss even when the chip looks excellent
Output impedance	Can affect low-impedance or multi-driver IEM frequency response
Analog stage	Affects real distortion, noise, channel matching, and drive stability
Power and grounding	Can influence USB noise, ground loops, hum, and interference
Volume control	Digital volume, analog volume, remote control, or preamp mode
Features	DSD, MQA, Bluetooth, display, remote, and firmware only matter if you use them
Support	Driver and firmware support matter more than a famous chip when things break

In plain language: do not buy a DAC only because the chip looks impressive. Let your ears and your actual setup receive the final product. The right DAC for you is stable, quiet, compatible with your system, matched to your amp or powered speakers, and comfortable with music you know.

For auditioning:

Use tracks you know, not only spectacular demo tracks.
Match volume carefully; louder is often mistaken for better detail.
Listen for hiss during quiet passages and pauses.
Check complex passages for harshness, congestion, and clipping-like stress.
Test USB disconnects, pops during track changes, and sleep/wake behavior.
Make sure the connectors fit your actual system without awkward adapters.
If an AB comparison is tiny, do not let chip marketing pressure you into upgrading.

A great chip with poor implementation can disappoint. A modest chip in a well-designed unit can be very enjoyable.

Amplifier And Power

An amplifier drives headphones or speakers. “Power” is about enough voltage, current, and control at the desired volume without excessive distortion.

Too little power can mean low volume, weak bass, compression, or distortion. Too much gain can cause hiss and poor volume control with sensitive IEMs.

Impedance And Sensitivity

Impedance and sensitivity together affect how easy a headphone is to drive.

High impedance is not always hard to drive. Low impedance is not always easy. Sensitive IEMs can reveal noise. Some low-impedance multi-driver IEMs are sensitive to output impedance.

Output Impedance

Output impedance is the impedance of the device output. Lower output impedance is usually safer for most headphones and IEMs because it reduces frequency-response interaction.

Higher output impedance can change tonal balance, especially with multi-driver IEMs.

SNR, Dynamic Range, THD+N, Crosstalk

Term	Meaning
SNR	Signal-to-noise ratio
Dynamic Range	Distance between usable loudest signal and noise floor
THD+N	Total harmonic distortion plus noise
Crosstalk	How much left and right channels leak into each other

Measurements are useful, but real experience also depends on gain, output level, driver stability, and headphone matching.

Balanced Output

Balanced outputs such as 2.5 mm, 4.4 mm, or XLR can provide more power or better separation in some designs. They are not automatically better.

Use only proper balanced cables. Do not force single-ended wiring into a balanced output with unsafe adapters.

Cables

Cables matter, but mostly for practical electrical reasons: resistance, capacitance, inductance, shielding, characteristic impedance, connector contact, mechanical reliability, length, and safety. A cable should deliver the correct signal reliably with low loss and low interference. It should not be treated as a magic tone enhancer.

The useful rule is:

Cables are best at fixing faults, noise, bad contact, wrong length, and wrong specifications. They are weakest as a promise of instant resolution, soundstage, or tonal miracles.

Cable	Common connectors	What matters	Common problems
Headphone cable	`3.5 mm`, `6.35 mm`, `2.5 mm`, `4.4 mm`, XLR	Pinout, contact, flexibility, microphonics, balanced compatibility	Reversed channels, unsafe adapters, loose plugs
Analog interconnect	RCA, XLR, TRS	Shielding, capacitance, grounding, balanced/unbalanced matching	Hum, buzz, RF pickup, high-frequency rolloff over long runs
Speaker cable	Banana, spade, bare wire	Gauge, resistance, contact area, polarity	Loose bass, reversed polarity, shorts, oxidized terminals
Coaxial digital	RCA, BNC	`75 ohm` characteristic impedance, shielding, termination	No lock, clicks, dropouts, reflections
AES/EBU	XLR	`110 ohm` balanced digital cable, twisted pair, shield	Short analog XLR may work, but long/pro links need the right cable
USB	USB-A/B/C	Standards compliance, data rate, shielding, power capability, firm fit	DAC disconnects, pops, power noise, charge-only cable
Optical	TOSLINK, Mini-TOSLINK	Alignment, bend radius, length, transmitter/receiver strength	No lock, format limits, but useful electrical isolation
Ethernet	RJ45	Rated cable, connector quality, length, shielding only when needed	Network dropouts and buffering, not analog tone shaping
Power cable	IEC and regional mains plugs	Safety certification, ground, current rating, connector contact	Safety risk, bad ground, loose contact

Why cables matter:

Resistance causes voltage drop and loss. Speaker cables are the clearest example because speakers are low-impedance, high-current loads.
Capacitance can interact with source impedance and slightly roll off treble, especially in long unbalanced analog cables or high-output-impedance gear.
Inductance can matter in speaker cables or unusual cable geometries, though it is rarely the first concern in short line-level cables.
Shielding reduces interference in low-level analog cables. Speaker cables normally do not need shielding for audio quality because signal level and current are high.
Balanced lines reject common-mode noise when the source, cable, and receiver are all truly balanced.
Digital cables need the right physical specification. S/PDIF coax is about 75 ohm, AES/EBU is about 110 ohm, and USB-C cables also involve data rate, power capability, and compliance.
Connectors matter. Many audible “cable upgrades” are really old plugs, oxidation, loose sockets, broken shields, or bad solder joints being fixed.
Length magnifies everything. A short desktop RCA run is not the same problem as a long stage or studio run.

How to read cable specifications:

Spec	Where it matters	Meaning
AWG / gauge	Speaker cables, power cables, some USB cables	Lower AWG usually means thicker wire; longer speaker runs need enough copper
Resistance	Speaker, headphone, and power cables	Lower resistance reduces voltage drop, especially in low-impedance or high-current use
Capacitance	RCA, phono, long analog runs	High capacitance can interact with high source impedance and reduce treble
Inductance	Speaker cables and unusual cable geometries	Excessive values can affect high frequencies or amplifier stability
Shield coverage	RCA, microphone, USB, coaxial digital	Better shielding helps low-level signals; speaker cables usually do not need shielding for sound quality
Characteristic impedance	S/PDIF, AES/EBU, USB, HDMI, Ethernet	Critical for digital/high-speed links; S/PDIF uses `75 ohm`, AES/EBU uses `110 ohm`
Connector quality	All cables	The point is stable contact and oxidation resistance, not magic plating
Bend radius	Optical, thick, and portable cables	Over-bending can damage fibers, conductors, or connector strain relief

Changing cables is meaningful when:

Speaker cable is too thin or too long.
Plugs are loose, oxidized, or intermittent.
RCA hum improves when routing or shielding changes.
A USB DAC disconnects, pops, or is recognized inconsistently.
A balanced headphone cable has the wrong pinout or unsafe adapter.
TOSLINK is bent, loose, or fails to lock.
Cable stiffness, weight, or microphonics affects headphone comfort.
Long studio or stage runs need balanced wiring and reliable shielding.

Changing cables is usually not the first move when:

The system is already quiet, stable, short-run, and correctly wired.
The reason is only that someone described a cable as more resolving.
Fit, placement, room acoustics, EQ, or amplifier matching are still unresolved.
Volume was not matched in the comparison.
The cable budget would crowd out bigger improvements elsewhere.

“Cable burn-in” is often overstated. Unlike a driver suspension, a cable does not have an obvious mechanical break-in mechanism. Perceived changes can come from restored contact, cleaned oxidation after replugging, changed headphone fit, unmatched volume, or listener adaptation. In daily use, secure contact, correct routing, and avoiding strain matter more.

Practical buying advice:

Use the correct pinout, especially for 2.5 mm, 4.4 mm, XLR, and proprietary headphone connectors.
Keep cables only as long as needed.
Prefer clear specifications: gauge, shielding, impedance, USB data/power rating, S/PDIF 75 ohm, AES/EBU 110 ohm.
Choose reliable plugs and strain relief.
For power cables, prioritize safety certification and grounding. Never defeat protective earth to chase sound.

For ECHO troubleshooting:

Pops, dropouts, or DAC disconnects: try a short compliant USB cable directly into the computer.
Hum or buzz: keep RCA away from power cords, use a shared power strip when appropriate, or try optical isolation.
Off-center vocals: use mono, left/right swap, and cable swapping to separate recording, headphone, cable, and device faults.
Loose speaker bass: check polarity, binding posts, strand shorts, and cable gauge.
Remote library buffering: check Ethernet, router, NAS, Wi-Fi, and server throughput before blaming decoding quality.
Compare cables at matched volume, changing only one thing at a time.

References:

Bluetooth Codecs

Bluetooth headphones use codecs such as SBC, AAC, aptX, LDAC, or LHDC. Bluetooth is convenient, but it adds encoding, latency, wireless stability concerns, and operating-system behavior.

Bluetooth playback is usually not bit-perfect. Higher codec modes can be less stable in poor radio conditions.

DSP

DSP means Digital Signal Processing. It includes EQ, headphone correction, room correction, FIR, channel balance, crossfeed, resampling, compression, and limiting.

DSP is not anti-HiFi. It is controlled digital processing. But it changes the samples, can add latency, and can cause clipping or phase problems if configured poorly.

Use DSP with purpose, small changes, and output safety checks.

EQ

EQ changes the level of frequency ranges.

Type	Use
Graphic EQ	Fixed bands, easy to use
Parametric EQ	Adjustable frequency, gain, and Q
Shelf filter	Raise or lower lows or highs broadly
High-pass	Remove content below a frequency
Low-pass	Remove content above a frequency
Notch	Cut a narrow problem frequency

For parametric EQ:

Frequency: center frequency.
Gain: boost or cut amount.
Q: width of the affected range.

Prefer small moves. If you boost several bands, reserve headroom.

Preamp And Headroom

Preamp is overall gain before or around processing. Headroom is reserved level space to prevent digital clipping.

If you boost bass by +4 dB and treble by +3 dB, the signal may exceed 0 dBFS. Lower preamp or enable headroom.

Practical defaults:

Mild EQ: around -3 dB can be enough.
Strong bass boost or multiple positive bands: consider -6 dB.
FIR, correction, and channel tools together: watch output safety closely.

Clipping And 0 dBFS

Digital audio has a maximum level called 0 dBFS. If a signal exceeds it, it clips. Clipping sounds hard, harsh, distorted, or broken.

Common causes include EQ boosts, high preamp, ReplayGain, FIR peak gain, stacked DSP, and already-clipped masters.

If ECHO shows output safety warnings, lower headroom or preamp first.

FIR, IIR, Convolution

Term	Meaning
FIR	Filter type useful for precise frequency and phase control
IIR	Efficient filter type common in normal EQ
Convolution	Processing audio with an impulse response
IR	Impulse response, a captured or designed response of a system

FIR and convolution are powerful, but wrong IR files, sample rates, channels, gain, or latency settings can cause problems. Use reliable IR files and reserve headroom.

Phase

Phase describes timing relationships in waveforms. Phase issues can thin bass, shift imaging, or make space feel strange.

Common terms:

Minimum phase: common in normal EQ, low latency.
Linear phase: preserves phase relationships across frequency, but can add latency and pre-ringing.
Polarity: positive/negative inversion, not the same as all phase behavior.
Pre-ringing: ringing before a transient, sometimes caused by linear-phase filters.

Compressor, Limiter, Loudness

A compressor reduces dynamic range. A limiter is a stronger ceiling. They are common in production, but casual playback use can flatten music.

Important parameters:

Threshold: when processing starts.
Ratio: compression strength.
Attack: how fast it reacts.
Release: how fast it stops.
Makeup Gain: volume added after compression.

Louder is not automatically better.

ReplayGain

ReplayGain makes different tracks or albums play at more consistent loudness. It does not remaster the music.

Track gain matches individual songs. Album gain preserves relative levels inside an album. If you want bit-perfect playback, disable it; for shuffle listening, it can be very useful.

Crossfeed

Crossfeed mixes a little of each channel into the other to make headphone listening more speaker-like. It can reduce hard left/right separation and in-head localization, but too much can narrow or blur the image.

Use it as a taste tool, not a mandatory HiFi switch.

HRTF And Virtual Space

HRTF describes how your head, ears, and body shape sounds arriving from different directions. Virtual surround and spatial audio often use HRTF.

It is personal. A preset that works for someone else may sound hollow or strange to you.

Channel Balance, Mono, Left/Right Swap

Use these as diagnosis tools:

Channel balance fixes small left/right level differences.
Mono checks center image and phase behavior.
Left/right swap helps find reversed wiring or device issues.
Channel delay should be used carefully.

Latency And Buffer

Buffering prepares audio ahead of time. More buffer is usually more stable but higher latency. Less buffer responds faster but can pop or drop out.

For music playback, latency is usually less important than stability. For gaming, video, or production, latency matters more.

WASAPI Shared, WASAPI Exclusive, ASIO

Mode	Best for	Notes
System / Windows Output	Stability and compatibility	Uses the system path
WASAPI Shared	Daily music, video, games	Multiple apps can play together
WASAPI Exclusive	Direct device control when stable	Other apps may be silent
ASIO	Professional interfaces and low-latency production	Not automatically better for listening

Do not force ASIO just because it sounds professional.

EQ Frequency Cheat Sheet

Range	Affects	Too much can sound
`20-60 Hz`	Sub-bass, rumble, atmosphere	Boomy, heavy, unstable
`60-120 Hz`	Kick and bass weight	Bloated
`120-250 Hz`	Warmth and thickness	Muddy
`250-500 Hz`	Body	Boxy
`500 Hz-1 kHz`	Midrange core	Nasal or crowded
`1-3 kHz`	Vocal presence	Shouty or hard
`3-6 kHz`	Clarity and attack	Sharp or tiring
`6-10 kHz`	Sibilance and brightness	Sibilant
`10-16 kHz`	Air	Thin or artificial

Use this as a starting point, not a formula.

Practical ECHO Checklist

When the sound feels wrong:

Return output mode to System or WASAPI Shared.
Test a known-good MP3 or FLAC.
Disable EQ, FIR, channel tools, headphone correction, ReplayGain, and speed changes.
Set Windows default format back to 24-bit / 48 kHz or 16-bit / 44.1 kHz.
Check ECHO volume, system volume, and device volume.
Check channel balance, mono, and left/right swap.
Re-enable EQ, DSP, exclusive mode, ASIO, or DSD one at a time.

If a module immediately makes things worse, turn it off before stacking more settings.

Summary

HiFi is not about enabling every advanced term. Resolution, separation, imaging, and soundstage help describe what you hear. Sample rate, bit depth, bit-perfect playback, WASAPI, and ASIO help explain the output path. EQ, headroom, FIR, and crossfeed help process sound safely.

Start stable, adjust gently, and troubleshoot before chasing bigger numbers.

For hands-on usage, continue with DSP Beginner Guide, DSP Simple Guide, and Audio Setup Advice.