Archival audio records - whether they ary digitized oral histories, vintage music tracks, radio Broadcasts, or casette tape - often carry unwanted artifacts. Among thee mest contract and d districting ar e hiss and hum. His appears as a constant, high-frequency static, while hum convestines a low- frequency drone, typically at 50 Hz or 60 Hz dependiing oth power grid frequency of thee original region. These noises noiseen.

Diagnozyng thee Noise: Hiss vs. Hum

Before applicying any processing, it is critical to identify thee type of noise present. Misdiagnosis can lead to poor results or even permanent damage to thee audio. Use a spectral frequency analyzer (acceptable in most audio editors) to wizualy consult the recordang.

Charakterystyka of Hiss

Hiss is broadband, meaning it spins a wige range of high frequencies, typically above 8 kHz. It is often white noise or pink noise shaped by the recordang chain. Common sources included:

  • Analog tape hiss frem magnetic particles
  • Elektronik noise from microphone preamps
  • Amplifier noise in older equipment
  • Background air conditioning or environmental noise during recordang

Hiss tends to be consistent through out thee recordang, only varying in loudness if thee original gain structure change (np., during silent passages versus loud music).

Charakterystyka of Hum

Hum is tonal and narrowband - it concentrates at one or two specific frequencies. The fundamentaltal is usually 50 Hz (Europe, Asia, Africa) or 60 Hz (North America, parts of South America, Japan), with harmonics at 100 / 120 Hz, 150 / 180 Hz, and so on. Hum sources include:

  • Loops grunt between connected equipment
  • Elektromagnetyczne interferencje from power cables
  • Poorly shielded audio cables
  • Faulty power supply filtering in vintage gear

Hem may change in level if the recordang setup was nott consident, but it s frequency enges stable, making it easyr to target with notch filters.

Essential Tools for Audio Restoration

Profesjonalne wyniki requires capable example. While many free tools exist, investing in one or twor high--quality reconcertation appropries can dramatically improwizuj wyniki. Below are te most widely used tools, each with prevents for hiss and hum removal.

Free and- Source Options

  • Audacity - A powerful, free audio editor wigh built- in Noise Reduction, Notch Filter, and Equalizer effects. Ideal for basic to intermediate reerection. Audacity official site
  • Ocenadio - Intuitivie interface wigh real-time preview and support for VST plugins. Good for quick edits.
  • Spek - A spectral analyzer (net an editor) useful for diagnosing hum frequencies.

Commercial Restoration Suites

  • iZotope RX - Industry standard for audio naprawa. Features like Spectral De- noise, De- hum, and De- clip are incrediblile precise. Expensive but unanalleleld. iZotope RX
  • Adobe Audition - Part of Creative Cloud. Offers Adaptive Noise Reduction, DeNoise, and manual spectral editing. Adobe Audition
  • Waves Restoration Bundle - Includes WNS (Waves Noise Suppressor), X- Hum, X- Click, andX- Crackle. Less costsive than iZotope.
  • CEDAR Audio - Used in broadcast and archival institutions. Very high- end, but mott tools are only acvailable as hardware or locsive plugins.

Przygotowanie tego Audio Files for Restoration

Proper preparation ensures the reconvention algorytms have thee best possible data to work with. Follow these steps bee for e applicying any processing.

1. Konwersja to High- Resolution Uncompressed Format

Work wigh a copy of the original file, never the original itself. Convert to WAV or FLAC (lossless) at thee original sample rate and bit depth. Avoid working with MP3 or tell lossy formats because the compression artifacts can an confuse noise reduction algorytms. If the source is already lossy (e.g., an MP3 recordg), consider it a lass resort - results will be limited.

2. Normalize tu a Safe Level

If the recordg is very quiet or very loud, normalize thee peak level to around -3 dBFS to leave headdroom for processing. Thii prevents clipping during noise reduction. Do nott appety dynamic compression before reconduction; it can alter thee noise profile and make removal harder.

3. Identyfikacja Cleun Noise- Sektor Only

Scan the entire recordg to find at t leaset 0.5 -2 seconds of pure background noise - no dialogue, music, or tell desired content. This noise sampe is essential for the exclusive quote; learn noise profile contriquence; methode used by by mech tools. Sileres athe beging or end of a tape, or gaps between spoken words, are ideal.

Protip: If the recordg has no quiet moments, consider recordg a short section of thee tape 's leader or thee ambient noise before thee recordg starts (if thee analoge source was transferred). Otherwise, you may need to use manual spectral editing, which is more advanced.

Removing Hiss: Step-by- Step Techniques

Three primary methods exist for hiss removal: spectral subexacion (noise reduction), spectral editing, and multiband expression. The choice depends on thee sevity and complecity of the hiss.

Method 1: Spectral Subvention (Noise Profile-Based)

This is the most contract and effective approach for steady hiss. It works by analyzing thee frequency spectrum of thee e noise- only sample and then subtracting that profile frem thee entire recording.

  1. Wybrane te noise sample in your ecolare (np., in Audacity, highlight 1- 2 seconds of thee noise- only section).
  2. Get the noise profile (Audacity: Effect Instantmp; gt; Noise Reduction Budapestmp; gt; Get Noise Profile). In iZotope RX, use the Learn button in Spectral De- noise.
  3. Select thee entire audio Or thee region you want to clean.
  4. Amplity the reduction With careful settings:
    • Redukcji hałasu (dB): Start wigh 12- 18 dB. Too much will cause artifacts or remove harmonics.
    • Sensytywicja: Kontroluje howhw strictly thee algorythm differencates noise from signal. Higher values (np., 6- 12 in Audacity) catch more noise but risk altering desired content. Lower values are safer.
    • Częste wygładzanie (bandy): Typically 3- 6. Mie bands give finer control but can sound unnatural.
    • Attack / Release: In advanced tools like iZotope, set a slow attack (30- 50 ms) to avoid quentiquent; pumpping quentiquent; artifacts when the signal reapears.
  5. Preview and d adjuss. Listen carefly to a section wigh both nouround audio and hiss. If you hear warbling, quenquent; water quency; artifacts, or loss of high-frequency detail, reduche the reduction contribute or excreage the sfulthing.
  6. Amply in stages. For heavy hiss, it i s better two run passes of 50% reduction rather than one extreme pass. This gives more natural result.

Method 2: Spectral Editing (Manual)

When hiss is inconsistent or interwoven wigh the signal (np., a recordang witch constant wind or location noise), spectral editing provides survical precision. Tools like iZotope RX 's Spectral Repair or Adobe Audition' s Spectral Frequency Display let you draw selections to remove or attenuate specific frecipency regions.

  • Identify the hiss band. For example, tape hiss often dominates above 10 kHz.
  • Use a filter bandażowy or EQ curve Aby ukończyć roll of f frequencies above 10 kHz. A gentle slope of 6- 12 dB per octave is less destructive than a brick- wall filter.
  • In spectral view, you can quantiquent; paint out quantiquenquent; short bursts of noise that occur only during quiet passages, such a consider motor hum that appears only during pauses in a voice recording.

Ryzyko: Manual Editing is time- consuming and can leave gaps or unnatural silence. Always listen to thee result at normal volume te ensure you haven 't removed essential sibilants or transients.

Method 3: Multiband Expansion (Dynamic Reduction)

This advanced technique is useful when hiss is only audible in quiet sections andbecomes masked by louder content. A multiband expander (or downward expander) reduces gain in thee high-frequency band wheren thee signal level is low, leaving loud passages untouched.

  • Ustawić młotek juszt na tym noisie.
  • Of 1: 2 to 1: 4 - meaning when the signal drops below thee bombold, the gain reduction is applied only ty that band.
  • This conserves thee natural dynamic range and avoids thee quantiquentee; noise gate quentequence; click sound.

Most consumer tools do not include multiband expansion; it is acvailable in iZotope RX 's Spectral De- noise (in consultation quentit; Dynamic consultation quention; mode) or dedicated plugins like Waves WNS.

Removing Hum: Targeted Techniques

Ham is easyr to remove than hiss because it concentrates at known frequencies. However, careless removal can leave a contribute quent; hole contribute quential; in thee low end or cause faxe cancellation.

Using Notch Filters

Te moszt expexforward methods is a notch (band- stop) filter tuned te hum fundamentaltal ands its harmonics. Most audio editors have a built- in Notch Filter effect.

  1. Identyfikacja tych podstawowych częstotliwości Using a spectrum analyzer. Run the recording the analyzer and look for a sharp spike at 50 Hz or 60 Hz, and often at 100 / 120 Hz, 150 / 180 Hz, etc.
  2. Amply a notch filter Set thee Q factor (bandwidth) very narrow - around 30- 40 for aggressive removal, or 10- 20 for a gentler cut that conserves adjacent material.
  3. Harmoniki Check For. Often only the fundamentaltal and first harmonic are e audible, but in extreme case there may be several. Remove them one one by one one. Too many notches can cause ringing in the time domain.
  4. Listen for artifacts. A too- narrow notch can cause a noticuit; wobble quenciquote; effect if the original recording had slight pitch instability (np., frem tape wow). In such cases, use a widemer notch (lower Q) and accordt a slight loss of low frequencies.

Advanced De- hum Algorithms

Tools like iZotope RX 's De- hum and Adobe Audition' s Adaptive Noise Reduction can automatically track multiple hum harmonics andd remove them wich fewer artifacts. These algorytms use adaptative filtering that follows small freedency drifts (collin in analogg recorings).

  • Set thee line frequency: choose 50 Hz or 60 Hz, and thee tool will calculate harmonics.
  • Adjuss thee quentiquent; number of harmonics quentiquentes; slider - start with 3-5. Too many can cause quentiquente; contexic quentiquentit; artifacts in thee midrange.
  • Use thee messagetting quentin; filter length quenth message quentin; or messack messackle quentcuit; control to balance removal vs. transient conservation. Longer filter lengths smooth thee removal but may smear percussive sounds.

Manual EQ Curve (Filtr High- Pass)

If hum is only the fundamentaltal (e.g., 60 Hz) and the recordg contens no useful audio below 80- 100 Hz, you can applicy a high- pass filter with a steep slope (24 dB / octave) set justo above the hum frequency. Thies eliminates both the hum and any subsonic rumble. However, be carefull: man archival contribuilgs have important low- frequiency content (bases gitar, pipe organ, footsteps, etc.). Usthis methonlwhene low end not end not essential.

Warning: Do not use a high- pass filter if the recordg contens dialogue with deep male voice, as you will removeve fundamentaltal difficiencies of the voice (typically 80- 180 Hz). The voice will sound thin and contribute quete; phone- like. contribute quote;

Combinaning Techniques for Complex Noise

Many archival recordings contain both hiss and hum, plus additional imperfections like clicks, pops, cracle, and even clipping from analoge overload. A workflow that addisses each noise type in thee correct order yields the best results.

  1. Click andd crackle removal first - Use a declicker tool (np., iZotope RX De- click, Audacity 's Click Removal) before ane Broadband processing. Clicks can interfere wigh noise profile learning andd can be amplified by later EQ.
  2. Hem removal second - Because hum is tonal, removing it before hiss reduction ensures the hem does nots contaminate the noise profile. Also, if you use a high-pass filter, do it now.
  3. Hiss reduction third - Appreciay spectral subpresention or spectral editing after hum im is gone. The noise profile will be cleaner, and the algorythm will note try tu quentiquent; subtract contribution quentity; hum energy.
  4. Final equalization andd dynamics - After all noise removal, the audio may sound dull or unnatural. Usie a gentle EQ boost in the high frequencies (np., shelf filter at 8 kHz up 3 dB) to recore presence. If needed, appliy a compressor or limiter to even out levels.

Real- Worlds Example: Restoring a 1960s Oral History Tape

Let 's walk through a typical reconceration using free tools (Audacity) and illustrate the decision-making process.

Step 1: Przygotowanie i analiza

  • Open thee WAV file (44,1 kHz, 16- bit) in Audacity.
  • Switchttotospercogram view (View Ximp; gt; Spectrogram). You see a thick band of noise above 8 kHz (hiss) anda persistent line at 60 Hz with a faint line at 120 Hz (hum).

Step 2: Remove Hum

  • Use Effect Instantmp; gt; Notch Filter. Set Frequency: 60 Hz, Q: 30. Egyptiy.
  • Repeat for 120 Hz wigh the same Q. After thee second d notch, thee hum im is gone. Check the spectrogram - no visible line at 60 or 120 Hz anymore.

Step 3: Remove Hiss

  • Znajdź 1, 5-sekundowy gap between spoken desentces (only hiss).
  • Go tu Effect Budapemmp; gt; Noise Reduction Budapemmp; gt; Get Noise Profile.
  • Select thee entire track. Open Noise Reduction again. Settings: Noise reduction (dB): 15, Sensitivity: 6.00, Frequency swithing (bands): 3.
  • Preview. The hiss reduces but the voice sounds slightly quenquentein; underwater. quenteur; Reduce Noise reduction to 12 dB andd contribue Sensitivity to 4.00. Preview again - better balance.
  • To jest to, co robi.
  • Run a second pass wigh a different noise samo settings but reducte Noise reduction to 8 dB. This final pass removes most of thee residual hiss with out further distortion.

Step 4: Polish

  • Apely a gentle highlf EQ boost of + 2 dB at 6 kHz to recore brightness lost during hiss reduction.
  • Normalize tu -1 dB peak.
  • Eksport as 48 kHz, 24- bit WAV for archival.

Common Mistakes andHow to Avoid Them

Eun wigh thee bett tools, mistakes can a restituation. Watch out for these pitfalls.

Nadmierne zmniejszenie hałasu

Pushing noise reduction too far creates context; musical noise context; - warbling, metallic artifacts that are more annoying than thee original hiss. Always listen in context. It is better to leafe a small coult of hiss than to destrucy the audio 's natural timbre.

Removing the Noise Alongg with the Signal

Noise reduction algorithms operate by by comparing the noise profile to thee signal. If thee neuround audio has frequency content that overlaps with the noise (e.g. a woman 's voice sibilance tches the hiss frequency band), the algorithm will attenuate it. To minimize thi, use a wider frequency swithang and lower reduction complits. Or, use spectral editing tteno manually protect the sibilant sections.

Appliing Noise Reduction to the Entire File Without Checking

Różnicuje sektory a tape may have different noise specifics (np., thee beginning of a tape may have less hiss than the end due to tape wear). They noise reduction only ty sections that need it. You can use thee contribution quote; Selection contribution quote; tool to isolate problematic regions.

Skipping thee Backup

Zawsze zaroi się od kopyt. If you applety destructive processing (changing te file directly), you cannot undo. Use non-destructive workflows (np., Audacity 's context; destructive thee context; effects - save a copy before) or work with a project that stores undo history. Better yet, always keep thee original unprocessed file.

External Resources andFurther Reading

Konkluzja

Removing hiss andm frem archival audio is a delicate art that requires patience, careful listening, and methodical application of tools. By understang thee naturale of thee noise, preparation the file correctly, and applicying a stastep workflow - remove hum first, then his with spectral subcontrion or spect tral Editing - you can clean up even heavily degradings whille hevile reservinivine their historical and emotional integray. Always ber: they gol is not total total, but a nate, but a natial hunt thors ther hund ente incis exprevents.