This is the summary of my bachelor thesis. It is about Musical Instrument recognition and shows:
* Features that can be extracted from sound.
* Machine learning basis.
* Decision algorithms.
* Statistical results.
Human Perception and Recognition of Musical Instruments: A ReviewEditor IJCATR
Musical Instrument is the soul of music. Musical Instrument and Player are the two fundamental component of Music. In
the past decade the growth of a new research field targeting the Musical Instrument Identification, Retrieval, Classification,
Recognition and management of large sets of music is known as Music Information Retrieval. An attempt to review the methods,
features and database is done.
2024 State of Marketing Report – by HubspotMarius Sescu
https://www.hubspot.com/state-of-marketing
· Scaling relationships and proving ROI
· Social media is the place for search, sales, and service
· Authentic influencer partnerships fuel brand growth
· The strongest connections happen via call, click, chat, and camera.
· Time saved with AI leads to more creative work
· Seeking: A single source of truth
· TLDR; Get on social, try AI, and align your systems.
· More human marketing, powered by robots
ChatGPT is a revolutionary addition to the world since its introduction in 2022. A big shift in the sector of information gathering and processing happened because of this chatbot. What is the story of ChatGPT? How is the bot responding to prompts and generating contents? Swipe through these slides prepared by Expeed Software, a web development company regarding the development and technical intricacies of ChatGPT!
Product Design Trends in 2024 | Teenage EngineeringsPixeldarts
The realm of product design is a constantly changing environment where technology and style intersect. Every year introduces fresh challenges and exciting trends that mold the future of this captivating art form. In this piece, we delve into the significant trends set to influence the look and functionality of product design in the year 2024.
Human Perception and Recognition of Musical Instruments: A ReviewEditor IJCATR
Musical Instrument is the soul of music. Musical Instrument and Player are the two fundamental component of Music. In
the past decade the growth of a new research field targeting the Musical Instrument Identification, Retrieval, Classification,
Recognition and management of large sets of music is known as Music Information Retrieval. An attempt to review the methods,
features and database is done.
2024 State of Marketing Report – by HubspotMarius Sescu
https://www.hubspot.com/state-of-marketing
· Scaling relationships and proving ROI
· Social media is the place for search, sales, and service
· Authentic influencer partnerships fuel brand growth
· The strongest connections happen via call, click, chat, and camera.
· Time saved with AI leads to more creative work
· Seeking: A single source of truth
· TLDR; Get on social, try AI, and align your systems.
· More human marketing, powered by robots
ChatGPT is a revolutionary addition to the world since its introduction in 2022. A big shift in the sector of information gathering and processing happened because of this chatbot. What is the story of ChatGPT? How is the bot responding to prompts and generating contents? Swipe through these slides prepared by Expeed Software, a web development company regarding the development and technical intricacies of ChatGPT!
Product Design Trends in 2024 | Teenage EngineeringsPixeldarts
The realm of product design is a constantly changing environment where technology and style intersect. Every year introduces fresh challenges and exciting trends that mold the future of this captivating art form. In this piece, we delve into the significant trends set to influence the look and functionality of product design in the year 2024.
How Race, Age and Gender Shape Attitudes Towards Mental HealthThinkNow
Mental health has been in the news quite a bit lately. Dozens of U.S. states are currently suing Meta for contributing to the youth mental health crisis by inserting addictive features into their products, while the U.S. Surgeon General is touring the nation to bring awareness to the growing epidemic of loneliness and isolation. The country has endured periods of low national morale, such as in the 1970s when high inflation and the energy crisis worsened public sentiment following the Vietnam War. The current mood, however, feels different. Gallup recently reported that national mental health is at an all-time low, with few bright spots to lift spirits.
To better understand how Americans are feeling and their attitudes towards mental health in general, ThinkNow conducted a nationally representative quantitative survey of 1,500 respondents and found some interesting differences among ethnic, age and gender groups.
Technology
For example, 52% agree that technology and social media have a negative impact on mental health, but when broken out by race, 61% of Whites felt technology had a negative effect, and only 48% of Hispanics thought it did.
While technology has helped us keep in touch with friends and family in faraway places, it appears to have degraded our ability to connect in person. Staying connected online is a double-edged sword since the same news feed that brings us pictures of the grandkids and fluffy kittens also feeds us news about the wars in Israel and Ukraine, the dysfunction in Washington, the latest mass shooting and the climate crisis.
Hispanics may have a built-in defense against the isolation technology breeds, owing to their large, multigenerational households, strong social support systems, and tendency to use social media to stay connected with relatives abroad.
Age and Gender
When asked how individuals rate their mental health, men rate it higher than women by 11 percentage points, and Baby Boomers rank it highest at 83%, saying it’s good or excellent vs. 57% of Gen Z saying the same.
Gen Z spends the most amount of time on social media, so the notion that social media negatively affects mental health appears to be correlated. Unfortunately, Gen Z is also the generation that’s least comfortable discussing mental health concerns with healthcare professionals. Only 40% of them state they’re comfortable discussing their issues with a professional compared to 60% of Millennials and 65% of Boomers.
Race Affects Attitudes
As seen in previous research conducted by ThinkNow, Asian Americans lag other groups when it comes to awareness of mental health issues. Twenty-four percent of Asian Americans believe that having a mental health issue is a sign of weakness compared to the 16% average for all groups. Asians are also considerably less likely to be aware of mental health services in their communities (42% vs. 55%) and most likely to seek out information on social media (51% vs. 35%).
AI Trends in Creative Operations 2024 by Artwork Flow.pdfmarketingartwork
This article is all about what AI trends will emerge in the field of creative operations in 2024. All the marketers and brand builders should be aware of these trends for their further use and save themselves some time!
A report by thenetworkone and Kurio.
The contributing experts and agencies are (in an alphabetical order): Sylwia Rytel, Social Media Supervisor, 180heartbeats + JUNG v MATT (PL), Sharlene Jenner, Vice President - Director of Engagement Strategy, Abelson Taylor (USA), Alex Casanovas, Digital Director, Atrevia (ES), Dora Beilin, Senior Social Strategist, Barrett Hoffher (USA), Min Seo, Campaign Director, Brand New Agency (KR), Deshé M. Gully, Associate Strategist, Day One Agency (USA), Francesca Trevisan, Strategist, Different (IT), Trevor Crossman, CX and Digital Transformation Director; Olivia Hussey, Strategic Planner; Simi Srinarula, Social Media Manager, The Hallway (AUS), James Hebbert, Managing Director, Hylink (CN / UK), Mundy Álvarez, Planning Director; Pedro Rojas, Social Media Manager; Pancho González, CCO, Inbrax (CH), Oana Oprea, Head of Digital Planning, Jam Session Agency (RO), Amy Bottrill, Social Account Director, Launch (UK), Gaby Arriaga, Founder, Leonardo1452 (MX), Shantesh S Row, Creative Director, Liwa (UAE), Rajesh Mehta, Chief Strategy Officer; Dhruv Gaur, Digital Planning Lead; Leonie Mergulhao, Account Supervisor - Social Media & PR, Medulla (IN), Aurelija Plioplytė, Head of Digital & Social, Not Perfect (LI), Daiana Khaidargaliyeva, Account Manager, Osaka Labs (UK / USA), Stefanie Söhnchen, Vice President Digital, PIABO Communications (DE), Elisabeth Winiartati, Managing Consultant, Head of Global Integrated Communications; Lydia Aprina, Account Manager, Integrated Marketing and Communications; Nita Prabowo, Account Manager, Integrated Marketing and Communications; Okhi, Web Developer, PNTR Group (ID), Kei Obusan, Insights Director; Daffi Ranandi, Insights Manager, Radarr (SG), Gautam Reghunath, Co-founder & CEO, Talented (IN), Donagh Humphreys, Head of Social and Digital Innovation, THINKHOUSE (IRE), Sarah Yim, Strategy Director, Zulu Alpha Kilo (CA).
Trends In Paid Search: Navigating The Digital Landscape In 2024Search Engine Journal
The search marketing landscape is evolving rapidly with new technologies, and professionals, like you, rely on innovative paid search strategies to meet changing demands.
It’s important that you’re ready to implement new strategies in 2024.
Check this out and learn the top trends in paid search advertising that are expected to gain traction, so you can drive higher ROI more efficiently in 2024.
You’ll learn:
- The latest trends in AI and automation, and what this means for an evolving paid search ecosystem.
- New developments in privacy and data regulation.
- Emerging ad formats that are expected to make an impact next year.
Watch Sreekant Lanka from iQuanti and Irina Klein from OneMain Financial as they dive into the future of paid search and explore the trends, strategies, and technologies that will shape the search marketing landscape.
If you’re looking to assess your paid search strategy and design an industry-aligned plan for 2024, then this webinar is for you.
5 Public speaking tips from TED - Visualized summarySpeakerHub
From their humble beginnings in 1984, TED has grown into the world’s most powerful amplifier for speakers and thought-leaders to share their ideas. They have over 2,400 filmed talks (not including the 30,000+ TEDx videos) freely available online, and have hosted over 17,500 events around the world.
With over one billion views in a year, it’s no wonder that so many speakers are looking to TED for ideas on how to share their message more effectively.
The article “5 Public-Speaking Tips TED Gives Its Speakers”, by Carmine Gallo for Forbes, gives speakers five practical ways to connect with their audience, and effectively share their ideas on stage.
Whether you are gearing up to get on a TED stage yourself, or just want to master the skills that so many of their speakers possess, these tips and quotes from Chris Anderson, the TED Talks Curator, will encourage you to make the most impactful impression on your audience.
See the full article and more summaries like this on SpeakerHub here: https://speakerhub.com/blog/5-presentation-tips-ted-gives-its-speakers
See the original article on Forbes here:
http://www.forbes.com/forbes/welcome/?toURL=http://www.forbes.com/sites/carminegallo/2016/05/06/5-public-speaking-tips-ted-gives-its-speakers/&refURL=&referrer=#5c07a8221d9b
ChatGPT and the Future of Work - Clark Boyd Clark Boyd
Everyone is in agreement that ChatGPT (and other generative AI tools) will shape the future of work. Yet there is little consensus on exactly how, when, and to what extent this technology will change our world.
Businesses that extract maximum value from ChatGPT will use it as a collaborative tool for everything from brainstorming to technical maintenance.
For individuals, now is the time to pinpoint the skills the future professional will need to thrive in the AI age.
Check out this presentation to understand what ChatGPT is, how it will shape the future of work, and how you can prepare to take advantage.
A brief introduction to DataScience with explaining of the concepts, algorithms, machine learning, supervised and unsupervised learning, clustering, statistics, data preprocessing, real-world applications etc.
It's part of a Data Science Corner Campaign where I will be discussing the fundamentals of DataScience, AIML, Statistics etc.
Time Management & Productivity - Best PracticesVit Horky
Here's my presentation on by proven best practices how to manage your work time effectively and how to improve your productivity. It includes practical tips and how to use tools such as Slack, Google Apps, Hubspot, Google Calendar, Gmail and others.
The six step guide to practical project managementMindGenius
The six step guide to practical project management
If you think managing projects is too difficult, think again.
We’ve stripped back project management processes to the
basics – to make it quicker and easier, without sacrificing
the vital ingredients for success.
“If you’re looking for some real-world guidance, then The Six Step Guide to Practical Project Management will help.”
Dr Andrew Makar, Tactical Project Management
Unlocking the Power of ChatGPT and AI in Testing - A Real-World Look, present...Applitools
During this webinar, Anand Bagmar demonstrates how AI tools such as ChatGPT can be applied to various stages of the software development life cycle (SDLC) using an eCommerce application case study. Find the on-demand recording and more info at https://applitools.info/b59
Key takeaways:
• Learn how to use ChatGPT to add AI power to your testing and test automation
• Understand the limitations of the technology and where human expertise is crucial
• Gain insight into different AI-based tools
• Adopt AI-based tools to stay relevant and optimize work for developers and testers
* ChatGPT and OpenAI belong to OpenAI, L.L.C.
How Race, Age and Gender Shape Attitudes Towards Mental HealthThinkNow
Mental health has been in the news quite a bit lately. Dozens of U.S. states are currently suing Meta for contributing to the youth mental health crisis by inserting addictive features into their products, while the U.S. Surgeon General is touring the nation to bring awareness to the growing epidemic of loneliness and isolation. The country has endured periods of low national morale, such as in the 1970s when high inflation and the energy crisis worsened public sentiment following the Vietnam War. The current mood, however, feels different. Gallup recently reported that national mental health is at an all-time low, with few bright spots to lift spirits.
To better understand how Americans are feeling and their attitudes towards mental health in general, ThinkNow conducted a nationally representative quantitative survey of 1,500 respondents and found some interesting differences among ethnic, age and gender groups.
Technology
For example, 52% agree that technology and social media have a negative impact on mental health, but when broken out by race, 61% of Whites felt technology had a negative effect, and only 48% of Hispanics thought it did.
While technology has helped us keep in touch with friends and family in faraway places, it appears to have degraded our ability to connect in person. Staying connected online is a double-edged sword since the same news feed that brings us pictures of the grandkids and fluffy kittens also feeds us news about the wars in Israel and Ukraine, the dysfunction in Washington, the latest mass shooting and the climate crisis.
Hispanics may have a built-in defense against the isolation technology breeds, owing to their large, multigenerational households, strong social support systems, and tendency to use social media to stay connected with relatives abroad.
Age and Gender
When asked how individuals rate their mental health, men rate it higher than women by 11 percentage points, and Baby Boomers rank it highest at 83%, saying it’s good or excellent vs. 57% of Gen Z saying the same.
Gen Z spends the most amount of time on social media, so the notion that social media negatively affects mental health appears to be correlated. Unfortunately, Gen Z is also the generation that’s least comfortable discussing mental health concerns with healthcare professionals. Only 40% of them state they’re comfortable discussing their issues with a professional compared to 60% of Millennials and 65% of Boomers.
Race Affects Attitudes
As seen in previous research conducted by ThinkNow, Asian Americans lag other groups when it comes to awareness of mental health issues. Twenty-four percent of Asian Americans believe that having a mental health issue is a sign of weakness compared to the 16% average for all groups. Asians are also considerably less likely to be aware of mental health services in their communities (42% vs. 55%) and most likely to seek out information on social media (51% vs. 35%).
AI Trends in Creative Operations 2024 by Artwork Flow.pdfmarketingartwork
This article is all about what AI trends will emerge in the field of creative operations in 2024. All the marketers and brand builders should be aware of these trends for their further use and save themselves some time!
A report by thenetworkone and Kurio.
The contributing experts and agencies are (in an alphabetical order): Sylwia Rytel, Social Media Supervisor, 180heartbeats + JUNG v MATT (PL), Sharlene Jenner, Vice President - Director of Engagement Strategy, Abelson Taylor (USA), Alex Casanovas, Digital Director, Atrevia (ES), Dora Beilin, Senior Social Strategist, Barrett Hoffher (USA), Min Seo, Campaign Director, Brand New Agency (KR), Deshé M. Gully, Associate Strategist, Day One Agency (USA), Francesca Trevisan, Strategist, Different (IT), Trevor Crossman, CX and Digital Transformation Director; Olivia Hussey, Strategic Planner; Simi Srinarula, Social Media Manager, The Hallway (AUS), James Hebbert, Managing Director, Hylink (CN / UK), Mundy Álvarez, Planning Director; Pedro Rojas, Social Media Manager; Pancho González, CCO, Inbrax (CH), Oana Oprea, Head of Digital Planning, Jam Session Agency (RO), Amy Bottrill, Social Account Director, Launch (UK), Gaby Arriaga, Founder, Leonardo1452 (MX), Shantesh S Row, Creative Director, Liwa (UAE), Rajesh Mehta, Chief Strategy Officer; Dhruv Gaur, Digital Planning Lead; Leonie Mergulhao, Account Supervisor - Social Media & PR, Medulla (IN), Aurelija Plioplytė, Head of Digital & Social, Not Perfect (LI), Daiana Khaidargaliyeva, Account Manager, Osaka Labs (UK / USA), Stefanie Söhnchen, Vice President Digital, PIABO Communications (DE), Elisabeth Winiartati, Managing Consultant, Head of Global Integrated Communications; Lydia Aprina, Account Manager, Integrated Marketing and Communications; Nita Prabowo, Account Manager, Integrated Marketing and Communications; Okhi, Web Developer, PNTR Group (ID), Kei Obusan, Insights Director; Daffi Ranandi, Insights Manager, Radarr (SG), Gautam Reghunath, Co-founder & CEO, Talented (IN), Donagh Humphreys, Head of Social and Digital Innovation, THINKHOUSE (IRE), Sarah Yim, Strategy Director, Zulu Alpha Kilo (CA).
Trends In Paid Search: Navigating The Digital Landscape In 2024Search Engine Journal
The search marketing landscape is evolving rapidly with new technologies, and professionals, like you, rely on innovative paid search strategies to meet changing demands.
It’s important that you’re ready to implement new strategies in 2024.
Check this out and learn the top trends in paid search advertising that are expected to gain traction, so you can drive higher ROI more efficiently in 2024.
You’ll learn:
- The latest trends in AI and automation, and what this means for an evolving paid search ecosystem.
- New developments in privacy and data regulation.
- Emerging ad formats that are expected to make an impact next year.
Watch Sreekant Lanka from iQuanti and Irina Klein from OneMain Financial as they dive into the future of paid search and explore the trends, strategies, and technologies that will shape the search marketing landscape.
If you’re looking to assess your paid search strategy and design an industry-aligned plan for 2024, then this webinar is for you.
5 Public speaking tips from TED - Visualized summarySpeakerHub
From their humble beginnings in 1984, TED has grown into the world’s most powerful amplifier for speakers and thought-leaders to share their ideas. They have over 2,400 filmed talks (not including the 30,000+ TEDx videos) freely available online, and have hosted over 17,500 events around the world.
With over one billion views in a year, it’s no wonder that so many speakers are looking to TED for ideas on how to share their message more effectively.
The article “5 Public-Speaking Tips TED Gives Its Speakers”, by Carmine Gallo for Forbes, gives speakers five practical ways to connect with their audience, and effectively share their ideas on stage.
Whether you are gearing up to get on a TED stage yourself, or just want to master the skills that so many of their speakers possess, these tips and quotes from Chris Anderson, the TED Talks Curator, will encourage you to make the most impactful impression on your audience.
See the full article and more summaries like this on SpeakerHub here: https://speakerhub.com/blog/5-presentation-tips-ted-gives-its-speakers
See the original article on Forbes here:
http://www.forbes.com/forbes/welcome/?toURL=http://www.forbes.com/sites/carminegallo/2016/05/06/5-public-speaking-tips-ted-gives-its-speakers/&refURL=&referrer=#5c07a8221d9b
ChatGPT and the Future of Work - Clark Boyd Clark Boyd
Everyone is in agreement that ChatGPT (and other generative AI tools) will shape the future of work. Yet there is little consensus on exactly how, when, and to what extent this technology will change our world.
Businesses that extract maximum value from ChatGPT will use it as a collaborative tool for everything from brainstorming to technical maintenance.
For individuals, now is the time to pinpoint the skills the future professional will need to thrive in the AI age.
Check out this presentation to understand what ChatGPT is, how it will shape the future of work, and how you can prepare to take advantage.
A brief introduction to DataScience with explaining of the concepts, algorithms, machine learning, supervised and unsupervised learning, clustering, statistics, data preprocessing, real-world applications etc.
It's part of a Data Science Corner Campaign where I will be discussing the fundamentals of DataScience, AIML, Statistics etc.
Time Management & Productivity - Best PracticesVit Horky
Here's my presentation on by proven best practices how to manage your work time effectively and how to improve your productivity. It includes practical tips and how to use tools such as Slack, Google Apps, Hubspot, Google Calendar, Gmail and others.
The six step guide to practical project managementMindGenius
The six step guide to practical project management
If you think managing projects is too difficult, think again.
We’ve stripped back project management processes to the
basics – to make it quicker and easier, without sacrificing
the vital ingredients for success.
“If you’re looking for some real-world guidance, then The Six Step Guide to Practical Project Management will help.”
Dr Andrew Makar, Tactical Project Management
Unlocking the Power of ChatGPT and AI in Testing - A Real-World Look, present...Applitools
During this webinar, Anand Bagmar demonstrates how AI tools such as ChatGPT can be applied to various stages of the software development life cycle (SDLC) using an eCommerce application case study. Find the on-demand recording and more info at https://applitools.info/b59
Key takeaways:
• Learn how to use ChatGPT to add AI power to your testing and test automation
• Understand the limitations of the technology and where human expertise is crucial
• Gain insight into different AI-based tools
• Adopt AI-based tools to stay relevant and optimize work for developers and testers
* ChatGPT and OpenAI belong to OpenAI, L.L.C.
5. Selecció de
característiques
Selecció de
classificadors
Configuració
Corpus
Configuració d’escenaris
Sistema de reconeixement
automàtic
Anàlisi de resultats
6. Intensitat i volum
Intensitat: quantitat absoluta
Volum: valor subjectiu
Altura: freqüència
Escala Mel
Duració
Timbre: permet diferenciar el mateix so produït per
dos instruments diferents
7. Informació que s’extreu de cada un dels petits
fragments en que es talla el senyal d’àudio
En el domini freqüencial
Centroide
Extensió
Coeficients MFCC
Roll-Off
En el domini temporal
Taxa de pas per zero
Energia
8. Model de mescles Gaussianes (GMM)
Probabilístic i sense supervisió
K-veí més proper (k-NN)
Basat en distància
Molt senzill d’implementar. Alt cost de computació
Support Vector Machines (SVM)
Basat en distància
Dissenyat per a discernir entre dues classes.
Versions multi-classe.
9. TunediT
115.000 mostres (no es disposa de l’àudio original)
70 – 30 % (entrenament – test)
18 instruments (2 famílies – 8 subfamílies)
Generació pròpia
18 instruments (2 famílies – 8 subfamílies)
Entrenament: Gravació de cada una de les notes de
l’extensió de l’instrument
Test: Gravacions amb fraseig per cada instrument.
10. Funcions relacionades amb l’extracció de les
característiques acústiques
cAudioList = TFCLoadAudioList (filename)
Cfeatures = TFCExtractFeatures (cAudioList,
windowSize)
TFCExportFeatures (audioListFile,
featuresFile, windowSize)
11. Funcions relacionades amb la classificació
vWhat = [x x x x x x x x x x x x x x x x x x]
MFCC1 … MFCC13
cResults = TFCGMMCheck (cFTraining,
vWhat, scale, classType)
CResults = TFCGMMClassify (cFTraining,
cFTest, vWhat, scale, classType)
Centroide
Extensió
RollOff
ZeroCross
Energia
19. Arxiu Instrument frames Classe Percent. Classe Percent. Encert
violoncel.wav violoncel 344 violoncel 64.83% contrabaix 23.26% SI
contrabaix.wav contrabaix 214 contrabaix 98.60% guitarra 1.40% SI
viola.wav viola 353 viola 82.44% violi 10.76% SI
violi.wav violi 138 viola 81.16% violi 17.39% NO
tubasolo.wav tuba 390 tuba 66.92% fagot 32.05% SI
celosolo.wav violoncel 683 violoncel 63.54% trompa 18.89% SI
clarsolo.wav clarinet 320 violoncel 32.81% clarinet 17.81% NO
violsolo.wav violi 633 violoncel 33.81% trompa 15.48% NO
vilasolo.wav viola 738 violoncel 37.53% violi 24.66% NO
clarinet.wav clarinet 341 clarinet 79.18% trombo 7.92% SI
saxofon.wav saxofon 229 saxofon 81.22% piano 18.78% SI
flauta.wav flauta 203 flauta 81.28% trompa 8.87% SI
corn_angles.wav corn_angles 215 corn_angles 36.28% trompeta 27.91% SI
Test gravat amb
les mateixes
condicions que
aprenentatge
Test procedent la
enciclopèdia
musical Microsoft
Musical
Instruments
Test format per
notes dobles o
triples del mateix
instrumments
Dades originals 1er candidat 2on candidat
20. Implementació funcions per a:
Extreure característiques acústiques
Classificar e identificar mostres
Analitzar resultats
Experiments que mostren entre altres:
El nombre de mostres afecta de diferent manera als
resultats segons el mètode de classificació
La rellevància de les característiques acústiques depèn del
mètode de classificació
Hi ha instruments més fàcils d’identifcar amb un mètode
que amb un altre
El fraseig harmònic redueix el percentatge d’encerts
Sóc Juan Castro Mayorgas i aquesta es la presentació del treball de fi de carrera d’Enginyeria Tècnica de Telecomunicacions especialitat Telemàtic amb el títol: “RECONEIXEMENT AUTOMATIC D’INSTRUMENTS MUSICALS”
Aquesta és l’organització de la presentació:
En primer lloc, es descriu el context en què el reconeixement automàtic d’instruments musical es desenvolupa i els objectius marcats en aquest Treball.
Després, s’exposen els fonaments teòrics que són el punt de partida de tota la part d’implementació.
Segueix amb la descripció dels Corpus que es fan sevir pels experiments i de les funcions més importants implementades.
Finalment, amb l’anàlisi dels experiments realitzats, es fa un resum del que han sigut les aportacions d’aquest treball.
El reconeixement automàtic d’instruments musicals és una de les àrees de treball dintre de l’àmbit multidisciplinari de la Recuperació d’Informació de la Música (de l-anglès Music Information Retrieval) que també recull camps com la notació automàtica, la identificació de cançons, la classificació de gèneres musicals i els sistemes de recomanació basats en paràmetres de similitud.
Per a entendre els objectius d’aquest treball, és apropiat veure primerament quin és el principi en el que es basa un sistema de reconeixement d’instruments musicals.
Un sistema d’aquest tipus consta, bàsicament, de dos mòduls. El primer d’ells consisteix en treure del senyal d’àudio d-entrada un conjunt de propietats acústiques que el caracteritzin. El segon mòdul utilitza aquestes característiques per a alimentar un sistema d’aprenentatge que, tal com es veu en la figura, serveixi com a model de referència en la identificació dels instruments de posteriors senyals f-audio
Aquest esquema reflexa els objectius del treball. El que s’ha volgut fer es estudiar el comportament d’un conjunt de caracteristiques acustiques i tècniques de classificació en el reconeixement automatic d’instruments musicals. Es a dir, es vol obtenir respostes a preguntes tals com:
Hi ha carateristiques que son mes rellevants en un sistema de classificacio que un altre?, Com varia el resultat segons el nombre de mostres?, etc.
Per assolir aquests objetius, s’ha hagut també de desenvolupar tota una serie de funcions d’alt nivell que implementin els mòduls d’un sistema de reconeixement automatic i que , a més, ofereixen les eines que facilitin configurar els escenaris i l’anàlisi posterior dels resultats.
La percepcio del so es basa en quatre atributs:
La intensiitat i el volum informen sobre la quantitat de so que es genera i es percep. La intensitat es un valor que es pot mesurar quantiativament i el volum es un valor subjectiu que depen de la sensibilitat uditiva de l’oient.
L’altura correspon a la frequencia de vibració de la font que l’origina. En música, aquestes frequencies estan prefixades a uns valors determinats segons uns criteris d’afinacio. Per exemple, la nota «LA» central d’un piano te una frequencia de 440 Hz.
L’escala mel es una forma de mesurar l’altura dels sons en funció de la percepció que l’oient té de la distància entre d’ells.
La duració és el temps trascorregut desde l’inici dela emissió del so fins la seva finalitzacio.
Finalment, el timbre es l’atribut que permet diferenciar dos instruments musicals que estan tocant la mateixa nota amb la mateixa intensitat i duració.
S’han considerat les següents característiques acústiques per la seva rellevancia:
Centroide: es com el centre de gravetat de l’espectre.
Extensió: Correpon al rang de frequencies ponderat en el centroide.
Coeficients MFCC: com es veura a continuacio, contitueixen una caracerística basica. Es calculen a partir de l’espectre del so.
RollOff: Indica els limits frequencials on resideix el 85% de la energia total de l’espectre.
Taxa de pas per zero: es un indicador del contingut sorollos d’un senyal.
Energia: correspon a l’energia mitjana del senyal. Es un indicador de la seva intensitat.
Son les eines que permeten, per una banda entrenar al sistema i, per l’altre, classificar noves dades segons l’entrenament realitzat.
El model de mescles gaussianes es un metode probabilistic i sense supervisio. Sense supervisio vol dir que les propies mostres d’aprenentatge no son etiquetades i que es el propi classificador qui les classifica.
El metode k-vei mes proper es molt senzill dimplementar pero en contrapartida, te un lt cost de omputacio. Es basa normalment en la distancia euclidiana.
Support Vector machines, es una altra metode basat en la distancia dissenyat inicilament per a la discernit entre dues classes. Es seu fonament es trobar l’hiperpla que optimitzi la distancia que el separa de les mostres de les dues classes. D’aquest meotde existeixen version multiclasse.
Un dels problemes que es presenten en aquest tipus d’estufis és l’obtenciño de suficients dades per a ffer provbes.
Amb aquest fianiltat per incorporar de forma addicional un conjunt de mostres procedent de TunedIt.
Son 115000 mostres en una proporcio 70-30 (entrenament-test) corespodnente a 18 instruments que pertaneyen a 2 families i 8 subfamilies.
Com no es tenen els arxius d’audio originals, cada mostra es considera com si fos un unic arxiu d’audio d’un unic frame.
El crpus de generacio propia es molt mes modest (no arriba al 10% en nombre de mostres). Per poder comparar resultats, s’ha generat gravant els sons dels mateixos 18 instruments que figuren el corpus de tunedit.
En canvi, la generacio del corpus propi ha permes experimenar amb la fase d’extraccio de característiques que no hauria sigut necessaria en cas d’utilitzar exclisivament TunedIt.
Ara es mostren les funcinos desenvolupades per a la realiztzacio dels experiments:
La priemra fase es la extraccio de caracteristiques a partir de les gravacions d’àudio. Les funcions principals son:
La funcio TFCLoadAudioList Carrega una llsita ‘arcius d’audio a memoria
La funcion YXExtacyt features treu les features de totls els arxius d’aui de la llista. Aquí ja tenim les carcateriatiques de cada un dels frames de cada un dels arxius d’auidio.
La funcio TFXExport features crea un arxiu a les caracteristiques acustiques de l’arxiu donat per peramentre.
Com es logic, aquesta fase no es necessari aplicar-la al Corpus de TunedIt ja que es disposa directamente de les caracteristiques.
Una vegada s’han extret les caracteristiques s’ha de procedir a la classificacio i identificacio.
Per a permetre escollir el conjut de catacterisitques que es vol implicar en l’estudi, es defineix un vector de 18 posicions. ON cada posicio fa referencia a les caracter9istiques tal com es mostren en l’esqueam. Nome per a aqueslles posicions en que el valor sigui 1, la caraceristica es tindra en compte.
Referent al metode GMM s’han es te les funcions:
TFCGMMCheck per a veure el resultat de la classificacio sobre el propi conjunt d’entrenament (tal com s’ha dit abans, GMM es un sistema sesne supervisio)
TFCGMMClassify retorna el resultat dela classificacio d’un conjunt de mostres TEST, on el model es TRAINING, WHAT indica quines caracteristiques s’han de tenir en compte, scale, imdica si s’han d’escalar en el rang `’1,¿1+, el valors i classtype indica sobre quin atribut s’ha de fer la cassigicacio (familia, subfamila, instrument)=
Ara es mostren les funcinos per als metodes KNN i SVM:
Els parametres TRAINING, TEST, WHAT SCALE i CLASTYPE tenen el mateix significat que l-ex;licat anteriorment.
A KNN s-FEGEIX EL PARAMETRE k QUE SIGNIFICA EL NOMBRE DE VEINS MES PROPERS SOBRE EL QUE L-ALGORISME PRENDRA LA DECISSIO.
I EL PARAMETRE maxdist UE SIGINIFXa la maxoma distancia a la que un vei pot ser consierat com a tal.
De les tre versions implementades, la prmera retorna la clase predominamt dels k-veins resultants. Si no existeix una unica classe predomnianta el resutlat reyornat es de incertesa.
En la segona version no es necessari que hi hagi una unica lcass pred9moniant per a retorna una classe.
En la tercera opcion, en cas de que no tots els k resultats onicideixin en la classe, la decision es pren amb el metode3 SVM en el que les mostes de training son els k anteriors.
EL metode SVM no te el parametre scale per que es implicit.
La funcio reports analutics fa un resym per instrument del nombre d-encert I erros en l-atrubiutd que es volia identifiar. Es un resum a nivell general que no te en compte les posibles mostres wue formen part dun mateix arxiu d-audio.
Aquest es al funcio valida per serutilitzada amb les dades de tunedirt.
En canvi, la fucio reportdecision, si que te en cimpte les mostres qye formen cada arxiu d-aidui. Els resutlatsm en aquests taula resum estan a nivell d-arcxiu on es visylaitzaen els primer I segon candidad amb els seus percetatges,
Una analogia no garie sorprenent quan la guitarra sona com un piano sonre tot a les notes baixes si es punteja. De getr el piano tw unes cordes…
Finaldment es mostren es resultats agrupanbt les mostres en funcion dels arxius d-auido originals.
Aquí acaba la presentació. Moltes gràcies per la vostra atenció.