Georgia Tech Voice+ Research Lab的动态

Language and music, fundamental elements of human culture, exhibit vast diversity worldwide. A recent study in Science Advances delved into understanding the shared and distinct features of speech and song across cultures, unveiling significant cross-cultural regularities in song acoustic features. The study was motivated by curiosity surrounding the evolutionary roles of language and music, both utilizing rhythm and pitch in vocal communication. Previous research has explored neural mechanisms, but empirical data on global similarities and differences between music and language has been lacking. This new study aimed to bridge this gap by analyzing acoustic attributes of speech and song across diverse cultures, seeking to unveil universal patterns and unique distinctions.

Harmony and Discord: The Impact of Leadership Dynamics on the Group Climate of Jazz Ensembles ?? Read more: https://lnkd.in/djnj5MAC This paper investigates the influence of different layers of leadership on the group climate within two jazz ensembles, consisting of conductors, instrument section leaders, and music students preparing for live concerts. Using data from 58 short semi-structured interviews, we examine how sentiments and emotions within these relationships affect group climate. We use applying both traditional coding by human researchers and coding by large language models (LLMs) with zero-shot learning. Our key finding is that the hierarchical relationships between conductors, section leaders and students impact the group's climate. This finding suggests that leadership plays a crucial role in the success of jazz ensembles. Additionally, our analysis demonstrates that LLMs, and particularly their average, can effectively code group climate, sentiments and emotions in small datasets which highlights the potential of LLMs in qualitative research. #ces_ethz #Jazz #leadership #ethzurich

Researchers from #NYUTandon led by Rose T. Faghih — inspired by the Yerkes-Dodson law in psychology, known as the inverted-U law — investigated the arousal-performance link throughout a cognitive task in the presence of personalized music. In this study, participants selected music with calming and exciting music components to mimic the low and high-arousing environment. To decode the underlying arousal and performance with respect to everyday life settings, they used peripheral physiological data as well as behavioral signals within the Bayesian Decoders. Click the link to learn more about this research. #NYUTandonMade

A very interesting study showing that singing is likely a universal human behavior. Not really a surprise, but this study gives validity to that intuition. So, if singing is universal, it follows that it is an evolved behavior. The larger question is “why?” Why do we sing? What is the evolutionary advantage? The study only points to various theories and identifies two variables in song (pitch and tempo) on which to focus future investigations. Personally, I lean toward song (and each art in general) as a pre-language expressive mode that communicates identity (“it’s me!”), location (“I’m here!”), status (“I’m excited!”), and action-readiness/initiation/direction/duration (“Let’s go, slowly, this way, for a minute!”). What I find exhilarating is that studies like this are well-constructed, thoughtful, inclusive, and humble. And they are using huge data sets and algorithms to help us analyze and understand who we were, who we are, and who we are becoming. And that can help us learn, teach, and live more effectively, sustainably, and—as it were—harmoniously.

Sheet Music - The Bar and The Full Score Repurposing The American Towers and The American Mines of Extraction A Metaphor used as a literary device to construct a conceptual frame through linguistics and cognitive psychology to communicate abstract concepts and skills where authors may re-enforce conceptual frames with Visual Rhetoric based upon Theories of Gestalt and Semiotics. Silicon Substrates A Descriptive Metaphor by definition can be used in a classroom with new learners. Although not a literary device, a descriptive Metaphor along with speaking and hand gestures will quickly construct a contextual frame for a Lecture while leaving time for more complex topics or procedures that require advanced cognitive psychology time and effort, both by the Lecturer and the new learners. For example, a Sheet of White paper used to describe a Substrate and its material construction, plus the Added Ink for the printed text and image as a Metaphor for the semiconductor Dopants of material form of p-type and n-type compounds that are applied to Silicon Substrates. The Responsibilities surrounding Operations. Remember! The Repurposing of Local Towers from Offices and Executive Operations straight into The Executive Living Quarters. Technolgy Influencers and Publishing Networks capture The Idea and Its Identity that stem from The Arts and Sciences.?

Do you want Jerry Garcia to read you a story? ElevenLabs and the Garcia estate have created something remarkable - an AI recreation of the Grateful Dead legend's voice, now reading in 32 languages through the ElevenReader app. The music industry faces a persistent challenge: how do you keep an artist's legacy vibrant and meaningful for new generations? Traditional approaches like remastered tracks or archived footage only tell part of the story. This AI voice recreation opens new doors. Garcia's distinctive sound - that warm, familiar tone that defined an era - now connects with audiences in unprecedented ways. From Spanish to Mandarin, his voice transcends both time and language barriers. Consider the possibilities: ? Narrating stories in multiple languages ? Reading poetry and literature ? Creating educational content The Garcia estate's move signals a shift in how we think about artistic legacy. Other music estates are taking notice - this could become the new standard for preserving iconic voices. The technology serves as a bridge, connecting longtime Deadheads with new listeners who never had the chance to experience Garcia's presence firsthand. Will you be tuning in?

Though I am far from being an expert, tomorrow I'll share my thoughts on the super-hot topic of AI's possible impact on our craft. Is live music, and the jobs of musicians in danger? Short answer is yes, of course, the longer one is: no, of course not. What do I mean? Find out tomorrow at the AI Summit in Budapest, from 3.30 PM at the brand new building of Museum of Ethnography.

It's been one year into my PhD, and whenever I meet non-PhD people, they always ask how a PhD is and what exactly it involves. I always struggled to give the most appropriate answer. After months of pondering, I think I finally have an answer. I feel PhD students are like music artists. For example, a musician first brainstorms and thinks about the idea, subject, or theme of the song. Once they figure that out, they start experimenting with different melodies, spending a lot of time finding the perfect one. Once they have the perfect melody, they spend time writing lyrics to match it. They also collaborate with other musicians, brainstorming on how to complete the jigsaw and make everything sound perfect. After working on it for months, they finally produce the song. They then have to show that music to the record company, and if they like it, they publish it. Once the song is out there, it is there for a lifetime. Now, let's understand how this is very similar to a PhD. PhD students also start by brainstorming an idea, subject, or theme on which their PhD will revolve. Once they figure that out, they start experiments to support their idea. A PhD student spends a lot of time, similar to a musician finding the perfect melody, to get the experiment working. Once the PhD student is successful in their experiment, they spend time writing a paper for it, similar to how musicians spend time writing lyrics. PhD students also work in groups or solo, just like musicians, each bringing their specialty. Once the student has worked on the experiments and written the paper, they need to publish it in a journal or conference. These also have rankings; the higher the rank, the more difficult it is to publish, but it's also highly regarded. Similarly, musicians need to publish their song with a music label company, and it's difficult to release it with a highly regarded one. Once the music is out, it is there for a lifetime. Even when the artist dies, we still listen to their songs and remember them. For example, we still listen to Michael Jackson's songs. Similarly, once the research is published, we still read about it even when the author has died. I strongly feel a PhD is like a song, and we are the musicians. #music #Phd #Researcher #Researchers #PHD #DataScience #AiResearcher #progress #careers #motivation #AdobeResearch #GoogleResearch #MicrosoftResearch #MetaResearch #OpenAIResearch #NetflixResearch #SpotifyResearch #ACML #Journal #Conferenece #ACL #IEEE

?? ?????? ???????????????????? ?????????? ?? I'm excited to share our project: Fine-tuning Large Language Models for Music Generation. Our process initiates with the conversion of MIDI files into a text-based representation utilizing Simplified Midi, a representation used by our tutor Ga?tan Hadjeres[1]. This representation encodes musical notes as quadruplets, each separated by a colon (:), with the variables representing: - ??: Pitch - ??: Velocity - ??: Duration - ??: Time before next note begins Following the conversion of MIDI files to text, we fine-tuned the Llama2 model, using Unsloth AI for faster and memory-efficient fine-tuning. Utilizing datasets such as GiantMIDIPiano and DadaGP, converted into our proprietary syntax, we developed two specialized models specifically designed for the generation of piano and guitar compositions. The models are capable of generating structured and contextually relevant responses to provided prompts, thereby having the potential to assist musicians in finding inspiration for their compositions. This project would not have been possible without the dedication of my team members: Marc Janthial, Charles Benichou-Chaffanjon, Jules Cognon Damien Fromilhague, Cyprien Laruelle. Special thanks to our project coordinator El Mahdi EL MHAMDI for his time and patience in evaluating our project. The complete code for this project is accessible on GitHub at the following URL: https://lnkd.in/epAFzzrS. Blog post : https://lnkd.in/dJ9pHNnA [1] Ga?tan Hadjeres and Léopold Crestel. The piano inpainting application, 2021. #MusicGeneration #LLM

I was 35 years old before I published anything significant. I am pound to say that my son, Spencer Kuhn just co-published a fine paper in Laser Journal linked below. It is an investigation of how sound waves are perceived as pleasant or unpleasant. Traditionally, scientists thought "smoothness" of sound waves determined how pleasing they sound. This study explored if that's true for dissonant sounds, not just perfect tones. The results were mixed: For some simple chords, the "smoothness" matched how pleasant they sounded. For other chords, especially complex ones, the "smoothness" didn't predict how pleasant they were. This suggests other factors, like how familiar a chord is to listeners, might also play a role in how pleasant it sounds. Overall, the study shows that understanding how pleasing music sounds is more complex than just analyzing the "smoothness" of sound waves.