Characterizing and Counteracting Wave Steepening in Horn-Loaded Loudspeakers
(2025) BMEM01 20251Department of Biomedical Engineering
- Abstract
- This thesis aims to characterize and counteract the wave steepening effect in a horn loaded loudspeaker. The report consists of two main parts. The first part, the characterization of the horn speaker system, focuses on investigating and understanding the signal distortions introduced by the speaker system. In particular, non-linear distortion is examined to determine any parameter dependence of the distortions. This is realized by analyzing the Total Harmonic Distortion (THD) ratio, the frequency response of the system, as well as the frequency spectra of the outputs. The system is measured using two types of signals, a logarithmic frequency sweep with frequencies from 400 Hz to 12.5 kHz and a stepped frequency sweep with distinct... (More)
- This thesis aims to characterize and counteract the wave steepening effect in a horn loaded loudspeaker. The report consists of two main parts. The first part, the characterization of the horn speaker system, focuses on investigating and understanding the signal distortions introduced by the speaker system. In particular, non-linear distortion is examined to determine any parameter dependence of the distortions. This is realized by analyzing the Total Harmonic Distortion (THD) ratio, the frequency response of the system, as well as the frequency spectra of the outputs. The system is measured using two types of signals, a logarithmic frequency sweep with frequencies from 400 Hz to 12.5 kHz and a stepped frequency sweep with distinct frequencies following the R10 format. The speaker is tested with multiple power levels to find any power dependence. All tests had the same setup and were conducted in an anechoic chamber. The findings from this part will serve as a foundation for the second part.
The second part of the report aims to find a way to counteract the distortions characterized in the first part. It is investigated if a static linear Finite Impulse Response (FIR) filter can be used to counteract the distortions introduced by the system. This is realized by using the Normalized Least Mean Square (NLMS) algorithm to adapt a filter. Multiple filters are tested in order to generate a filter capable of counteracting the harmonic distortion. This is done by varying the learning rate and the filter length as well as adapting filters on different white noise signals. The filters are then evaluated by filtering the input signal before feeding it to the speaker. The filtered output is then compared to the unfiltered output in terms of THD and frequency response. The time domain signal and the corresponding frequency spectra are analyzed for specific frequencies.
The speaker system and the distortions it introduces are non-linear. Harmonic distortion is introduced and the THD reveals that it is both frequency and power dependent. The distortion originate both from the speaker driver and from the air distortions. The wave steepening effects caused by air distortion are especially prominent in the speaker pass-band, the frequency range where the speaker has the highest efficiency.
The FIR-filter with the best performance regarding wave steepening cancellation had a learning rate of 0.005 and 248 filter taps. It was adapted on 1-3 kHz bandpassed white noise. The band-passed range was chosen to target the area with the highest level of wave steepening. It was found that the filtered output had significantly lower harmonic distortions compared to the unfiltered output. The frequency response was flattened out in addition to this, a desirable effect in loudspeaker design. (Less) - Popular Abstract (Swedish)
- Linjär lösning på ett icke-linjärt problem - effektiv filtrering i hornhögtalare
Hornhögtalare är effektiva och skapar högt ljud men till priset av ljudkvaliteten. Genom linjär filtrering kan man få det bästa av två världar: minskade icke-linjära distorsioner och hög effektivitet.
I vägen mot en säkrare värld spelar hornhögtalaren en viktig roll. Hornhögtalaren är förknippad med höga ljudnivåer och används ofta för larmsignaler. Men i vissa situationer vill man skicka ut ett röstmeddelande via högtalarna för att varna eller ge instruktioner för till exempel utrymning av lokal. Hornhögtalare är som sagt väldigt bra på att producera höga ljudvolymer, däremot är de inte lika bra på att skapa rent ljud. Anledningen till detta är att... (More) - Linjär lösning på ett icke-linjärt problem - effektiv filtrering i hornhögtalare
Hornhögtalare är effektiva och skapar högt ljud men till priset av ljudkvaliteten. Genom linjär filtrering kan man få det bästa av två världar: minskade icke-linjära distorsioner och hög effektivitet.
I vägen mot en säkrare värld spelar hornhögtalaren en viktig roll. Hornhögtalaren är förknippad med höga ljudnivåer och används ofta för larmsignaler. Men i vissa situationer vill man skicka ut ett röstmeddelande via högtalarna för att varna eller ge instruktioner för till exempel utrymning av lokal. Hornhögtalare är som sagt väldigt bra på att producera höga ljudvolymer, däremot är de inte lika bra på att skapa rent ljud. Anledningen till detta är att insignalen blir vågbrantad i hornet vilket betyder att det skapas övertoner. Om man skickar in en ren sinussignal på 500 Hz kommer utsignalen från högtalaren innehålla grundtonen 500 Hz men också dess övertoner 1000 Hz, 1500 Hz, 2000 Hz och så vidare. Detta gör att vågformen förändras från en sinus till en framåtlutad sågtandsliknande våg. Denna typ av distorsion kallas för icke-linjär distorsion. En vågbrantad signal kommer inte längre att låta som den ska. Detta kan göra att ett röstmeddelande blir otydligt och instruktioner blir svåra att förstå. Fenomenet är tryckberoende, ju högre tryck desto starkare blir övertonerna.
För att motverka vågbrantningen och därmed förbättra ljudkvaliteten har vi skapat ett digitalt filter, ett så kallat FIR-filter, som dämpar övertonerna. Framför allt är vi intresserade av frekvensområdet där högtalaren är speciellt effektiv och därmed skapar väldigt höga ljudvolymer. Eftersom vågbrantning är tryckberoende är de icke-linjära distorsionerna som värst i detta område. Genom att träna filtret på frekvenser inom detta område lyckades vi skapa ett filter som kunde bibehålla hög ljudvolym men samtidigt dämpa övertonerna. Genom att implementera ett linjärt filter i hornhögtalaren kan röstmeddelanden i till exempel offentliga larmsystem bli tydligare, utan att man behöver kompromissa med ljudstyrkan. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9196969
- author
- Nilhammar, Lina LU and Larsson, Vilma LU
- supervisor
- organization
- alternative title
- Analys och Filtrering av Vågbrantning i Hornhögtalare
- course
- BMEM01 20251
- year
- 2025
- type
- H2 - Master's Degree (Two Years)
- subject
- language
- English
- additional info
- 2025-08
- id
- 9196969
- date added to LUP
- 2025-06-16 08:20:12
- date last changed
- 2025-06-16 08:20:12
@misc{9196969, abstract = {{This thesis aims to characterize and counteract the wave steepening effect in a horn loaded loudspeaker. The report consists of two main parts. The first part, the characterization of the horn speaker system, focuses on investigating and understanding the signal distortions introduced by the speaker system. In particular, non-linear distortion is examined to determine any parameter dependence of the distortions. This is realized by analyzing the Total Harmonic Distortion (THD) ratio, the frequency response of the system, as well as the frequency spectra of the outputs. The system is measured using two types of signals, a logarithmic frequency sweep with frequencies from 400 Hz to 12.5 kHz and a stepped frequency sweep with distinct frequencies following the R10 format. The speaker is tested with multiple power levels to find any power dependence. All tests had the same setup and were conducted in an anechoic chamber. The findings from this part will serve as a foundation for the second part. The second part of the report aims to find a way to counteract the distortions characterized in the first part. It is investigated if a static linear Finite Impulse Response (FIR) filter can be used to counteract the distortions introduced by the system. This is realized by using the Normalized Least Mean Square (NLMS) algorithm to adapt a filter. Multiple filters are tested in order to generate a filter capable of counteracting the harmonic distortion. This is done by varying the learning rate and the filter length as well as adapting filters on different white noise signals. The filters are then evaluated by filtering the input signal before feeding it to the speaker. The filtered output is then compared to the unfiltered output in terms of THD and frequency response. The time domain signal and the corresponding frequency spectra are analyzed for specific frequencies. The speaker system and the distortions it introduces are non-linear. Harmonic distortion is introduced and the THD reveals that it is both frequency and power dependent. The distortion originate both from the speaker driver and from the air distortions. The wave steepening effects caused by air distortion are especially prominent in the speaker pass-band, the frequency range where the speaker has the highest efficiency. The FIR-filter with the best performance regarding wave steepening cancellation had a learning rate of 0.005 and 248 filter taps. It was adapted on 1-3 kHz bandpassed white noise. The band-passed range was chosen to target the area with the highest level of wave steepening. It was found that the filtered output had significantly lower harmonic distortions compared to the unfiltered output. The frequency response was flattened out in addition to this, a desirable effect in loudspeaker design.}}, author = {{Nilhammar, Lina and Larsson, Vilma}}, language = {{eng}}, note = {{Student Paper}}, title = {{Characterizing and Counteracting Wave Steepening in Horn-Loaded Loudspeakers}}, year = {{2025}}, }