gaudio

뒤로가기back

HEY MA&PA - An unforgettable day at work with our kids

2024.09.30 by Heidi Hwang

 

 

We organized the "HEY MA&PA" event for Gaudio members and their children, who were on summer break, to spend some quality time together.

 

The event featured a variety of activities designed for both parents and kids, allowing families to bond and create lasting memories. Most importantly, we wanted the children to experience their parents' workplace firsthand, strengthening family ties and instilling pride in the work their parents do.

 

At first, the kids shyly held their parents' hands, but as the day went on, their excitement grew, and they became more engaged. Curious about what made this event so memorable for them? Let’s take a look!

 

 

 

A Day at Dad’s Workplace!

 

 

우리 아빠네 회사로 소풍가요!

 

우리 아빠네 회사로 소풍가요!

 
 

The theme of the event was "A Picnic at Mom and Dad's Office 🧺." With picnic chairs, checkered mats, and delicious food set up in our lounge area, the day kicked off with a fun adventure for the kids. Gaudio Lab’s lounge, usually a relaxation spot for employees, was transformed into a cozy picnic area where families could unwind. The kids had the chance to explore the office and see what their parents' workplace is like up close.

 

 

일등으로 도착한 Jayden의 자녀 신영이

<Jayden's child, Shinyoung, first to arrive>

 

 

 

Exploring the Office

 

 

 

As soon as the children arrived, they embarked on a series of missions, guided by our staff. These missions took them to various parts of the office, giving them a glimpse of where their parents spend most of their day.

 

Gaudio Lab’s office, modeled after locations in Jeju Island, is designed to support sound research and development, making it a unique space for the kids to explore.

 

At the "Bijarim" area, the kids got to experience Gaudio Lab’s audio technology, taking on the role of junior engineers. They also posed for pictures at their parents' desks, tapped away on keyboards, and explored the workspace while enjoying snacks. Some even checked out their parents' New Year’s resolutions, with a few children candidly revealing whether their parents were sticking to them! 😆

 

After completing all the missions, the kids received family t-shirts and cute stickers as rewards, and the day ended with a family photo in front of our large media wall, "Ora".

 

 

<비자림에서 가우디오랩의 공간 음향 기술을 직접 체험하는 아이들>

<Children experiencing Gaudio Lab's spatial sound technology at Bijarim> 

 

 
<엄마 아빠 자리에서 무슨 일 하는지 설명듣는 아이들>

<Kids learning about what their parents do at work>

 

 
<포토존에서 Johnny네 가족사진 찍기>

<Johnny's family taking a photo at the photo zone>

 

 

 

CEO Henney's Special Mission: Keeping the Kids Engaged

 

 

헤니 사장님의 회사 소개 (특별 미션: 아이들의 집중력을 지켜라!)
 

 

No company event would be complete without a welcome speech!

Our CEO, Henney, was tasked with introducing Gaudio Lab to the children in an exciting way. By demonstrating Gaudio Lab’s music separation technology using the kids’ favorite songs, Henney kept their attention while proudly highlighting the important work their parents do. It was a proud moment for both the kids and the parents. 😃

 

 

 

Bath Bomb Making Time

 

To give the children a memorable experience in the “Aewol” meeting room—typically reserved for client meetings—we decided on a bath bomb-making activity. This hands-on project was chosen because it was something kids of all ages could enjoy. We brought in an instructor, and the kids threw themselves into the creative process, crafting their own bath bombs with enthusiasm. It was heartwarming to see them proudly showing off their creations to their parents 🥰. Meanwhile, the parents enjoyed a well-deserved moment of relaxation.

 

 

입욕제 만들기 시간
 
입욕제 만들기 시간

<Fun bath bomb making activity> 

 

 

 

The Grand Finale: A Group Photo

 

To wrap up the event, we gave each family matching t-shirts, which they wore for a group photo. Seeing everyone smiling and wearing their shirts together was the perfect way to close the day. The memories captured in these photos will be treasured by all who attended.

 

 

 

For the children, this day was likely an unforgettable experience spent with their parents. And for the parents, sharing a part of their daily work life with their kids made the day truly special. This event wasn’t just a family outing—it was an opportunity for the children to better understand and take pride in their parents’ work. Gaudio Lab is dedicated to creating more moments like this for families in the future.

A big thank you to everyone who participated, and we look forward to seeing you at the next event! 🥳

 

 

Related article: https://www.econovill.com/news/articleView.html?idxno=663840

pre-image
Exploring ICML 2024: Latest Advances in AI and Audio Research

Introduction   Hello again, it’s Kaya back with more updates!As a researcher focused on audio AI at Gaudio Lab, I often get the chance to attend various conferences. You might remember my recent post on ICASSP 2024 & Gaudio Night.   This time, I’m excited to share my experience attending the ICML (International Conference on Machine Learning) 2024, which was held in beautiful Vienna, Austria.   ICML is one of the most prominent AI conferences in the world, along with ICLR and NeurIPS, and Gaudio Lab makes it a point to attend every year. It’s an incredible opportunity for researchers like me to connect with others in the field and discover the latest breakthroughs.   Unlike my last post, which focused on the event atmosphere, this one will dive into the innovative ideas and insights I gained from ICML 2024!   The venue of ICML 2024 held in Vienna, Austria. Wien~       Highlights from ICML 2024   Spotlighted Papers   Let’s start with a few standout research papers I encountered. The first one that caught my eye was "Discrete Diffusion Modeling by Estimating the Ratios of the Data Distribution." This study introduces a novel approach to handling discrete data, showcasing a method called SEDD that outperforms existing SOTA models like GPT-2 in the natural language processing field. This research significantly broadens the applicability of diffusion models and has made a notable impact in the AI community.     Lou, Aaron, Chenlin Meng, and Stefano Ermon. "Discrete Diffusion Language Modeling by Estimating the Ratios of the Data Distribution."       Another intriguing study was titled "Debating with More Persuasive LLMs Leads to More Truthful Answers." This paper presents experimental evidence that AI models can enhance their accuracy through debates—a concept that once seemed confined to the realm of imagination. This research could play a crucial role in addressing issues with misinformation in models like ChatGPT.       Khan, Akbir, et al. "Debating with More Persuasive LLMs Leads to More Truthful Answers."         The Introduction of “Position” Papers   A new category of papers made its debut at this year’s ICML: ✨Position✨ papers. These papers don’t necessarily propose new models or ideas; instead, they challenge current academic norms and provoke deep reflection.   One particularly compelling paper was titled "Position: Measure Dataset Diversity, Don’t Just Claim It." This study argues that simply asserting dataset diversity isn’t enough. The authors analyzed 135 image and text datasets to offer a fresh perspective on the topic, reminding us as AI researchers to deeply consider the fairness and inclusiveness of datasets.       Trends in Audio Research   ICML 2024 also featured a number of exciting papers on audio AI.   The current trend in this field is a focus on more sophisticated generative models, which is evident across areas from music generation to general-purpose audio synthesis.   One of the most impressive studies was "DITTO: Diffusion Inference-Time T-Optimization for Music Generation." This paper introduces a technique that allows for precise control over the intensity, melody, and structure of generated music, marking a significant step forward in music AI. The future of music generation looks incredibly promising with such advancements.       Novack, Zachary, et al. "Ditto: Diffusion inference-time t-optimization for music generation."       Additionally, Video-to-Audio Generation has emerged as a hot topic. With tools like OpenAI’s “Sora” pushing the boundaries of video generation, the creation of matching audio for these videos has become a critical research area. Google proposed a model called "VideoPoet" that generates both video and audio simultaneously, while Adobe introduced "Masked Generative Video-to-Audio Transformers with Enhanced Synchronicity," focusing on syncing sound effects with video actions.           These developments naturally reminded me of Gaudio Lab’s own sound effects generation model, FALL-E. It’s a technology that even caught the attention of Microsoft’s CEO, Satya Nadella, at CES. A few months ago, we demonstrated how FALL-E could generate sound effects perfectly synchronized with Sora videos. Seeing these trends at ICML reinforced my pride in how quickly Gaudio Lab is catching on to industry trends and reaffirmed our team’s research direction.😎       Conclusion   In this post, I’ve shared some of the diverse trends and fascinating research from ICML 2024. I was so engaged in exploring all the new studies that I found myself running around the conference hall to keep up with everything! 🏃‍♀️     One particularly fun experience was a poster session where researchers were given questions about large language models (LLMs) and then encouraged to debate, turning the session into a lively and spontaneous event.     Attending ICML 2024 and absorbing all the new trends and research has left me feeling inspired and ready to return to Korea to continue developing smarter AI models. I’m excited to apply the knowledge and inspiration I gained from the conference.   Hopefully, at the next conference, we’ll see Gaudio Lab’s research featured as a spotlighted paper! 💪 Until then, this wraps up my ICML 2024 recap. 😁    

2024.08.30
after-image
Dissonance, Mahler, and Beyond - Part 1

Hello! I’m Ste, and I’m researching voice AI at Gaudio Lab.   Lately, I’ve been diving into the music of the classical composer Gustav Mahler here at Gaudio Lab!Mahler is known for his music, which is full of strong dissonances and intricate structures.   To express the emotions of love and pain he felt throughout his life, he used dissonance as a key element in his compositions. Let’s explore these dissonances from an acoustical, musical, and historical perspective!     1 Dissonance   Music is an art that uses sound as its material. Just like how meat or seafood is essential to make a flavorful soup, harmonious chords are the base of music. But, just like those basic ingredients alone can’t make a rich soup, the addition of dissonance, with its strong personality, is what gives music its depth and variety. A piece of music made only from harmonious tones would sound monotonous, like a simple broth made only from meat. Dissonance, on the other hand, can provide stimulation and create tension, making the music more compelling—like adding seasoning to a dish.     1.1 Complex Integer Ratios   How does dissonance occur? It has to do with the ratio of frequencies between two notes. Sound is created by vibrating objects, and when two notes vibrate at a simple integer ratio, we hear them as consonant. But when the ratio between their frequencies is more complex, we hear them as dissonant. Here’s a table showing the ratios and consonance/dissonance of different intervals based on the note A:   Table 1: Intervals, Integer Ratios, and Consonance/Dissonance Based on A   For example, the perfect octave (2:1 ratio) and the perfect fifth (3:2 ratio) sound harmonious to the ear. These ratios form the foundation of basic harmony. As the frequency ratio becomes more complex, dissonance emerges. For example, intervals like the minor second (16:15) and the augmented fourth (45:32) sound dissonant. These dissonances are useful in music to express tension or emotional shifts. But what happens when the ratio is extremely complex, like 441:440, where the two notes’ frequencies differ only slightly?     1.2 Beating   Another way to understand dissonance is through a phenomenon called "beating." This occurs when two frequencies that are very close to each other vibrate together, causing their amplitudes to modulate periodically. If this modulation happens quickly, it can create an unpleasant tension—almost like a drum being struck rapidly. This effect is called "beating." According to the physicist Helmholtz, the tension caused by beating is most noticeable in the 30-40Hz range. [2] The principle behind beating is simple, and can be derived from the addition formula of trigonometric functions.   If we combine two sine waves with frequencies f1 and f2, the resulting wave can be expressed as:   Using the trigonometric addition formula, we get:   The cosine term represents the center frequency, vibrating at (f1+f2)/2, while the sine term represents the beating frequency, vibrating at |f1−f2|/2. This is fascinating: when we combine two waves with different frequencies, a new center frequency appears, and amplitude modulation happens, causing the sound to either amplify or dampen. I remember when I was younger, I thought that when I played C and E on the piano, I would hear the note D in between, but I was confused when this didn’t happen. It’s amazing to think that this phenomenon happens when the sound sources are the same and frequency differences are so small!   Figure 1: Combining the A tone of 440Hz with 441Hz, Bb, and C#     Figure 1 shows the result of combining two different frequencies in a graph. (a) and (b) show the combination of 440Hz and 441Hz. Despite the complex ratio, this combination sounds consonant, with a center frequency of (440+441)/2 = 440.5Hz and a beating frequency of (441−440)/2 = 0.5Hz. This results in a slow modulation, producing a "waa-waa" effect without creating the percussion-like tension we associate with dissonance.   (c) and (d) show a dissonant minor second, with a beating frequency of 29.33Hz. This rapid beating creates tension, like a drum being struck about 30 times per second, which makes the sound dissonant. (e) and (f) show consonant tones of A and C#, with a beating frequency of 110Hz. This frequency is too fast for us to perceive as rhythm, so we hear the notes as independent and consonant.   Figure 2: Dissonance within an Octave     [2] Helmholtz suggested this analysis of dissonance through beating, and [3] Plomp and Levelt, after experimenting with human subjects, created a graph of how different ratios within an octave affect the perception of dissonance. If we reproduce this graph based on intervals, we get Figure 2. From the graph, we can see that intervals like the perfect unison, major/minor thirds, perfect fourth, perfect fifth, and major/minor sixth have low dissonance compared to neighboring intervals. The dissonance increases between the minor second and perfect unison due to the tension created by beating. However, as the interval nears a perfect unison, the dissonance quickly decreases because the beating frequency slows down, no longer creating that percussive tension.     1.3 Non-Chord Tones   As mentioned earlier, dissonance is essential to music because it adds variety and depth. In music, the relationship between intervals becomes more structured, forming a hierarchy of harmonic tones that make up the system of harmony. For example, the C-E-G chord forms the harmony of the C major chord. The G note, a perfect fifth above C, is consonant, and the E note, also relatively consonant, fits in between. Once this harmony is established, all other notes, like D, F, A, and B, are considered non-chord tones, as they create dissonances with C, E, or G.   One of the key features of classical music, such as Mozart’s, is the clear hierarchy between chord tones and non-chord tones. Non-chord tones create tension with the other chord tones, and thistension is resolved when non-chord tones move to a chord tone. This process is known as "resolution" in musical terms. Within a clear harmonic structure, this tension-resolution pattern moves the music forward, like fitting pieces of a puzzle together. It’s similar to how in language, nouns are modified by adjectives and verbs, and adjectives and verbs are further modified by adverbs. Music without non-chord tones, consisting only of harmony, might sound innocent and naive, like a child speaking only in nouns, without adjectives or verbs.   Figure 3: Mozart’s String Quartet No. 19 "Dissonance" – Red lines show dissonance, green lines show resolution     Figure 3 shows how Mozart creates dissonance in his string quartet and elegantly resolves it. The passage begins with the cello playing C notes in succession. After two beats, the viola introduces an Ab note, creating a sense of tension due to the use of a first inversion chord. Normally, music starts with a root-position chord, and we expect the perfect fifth to be present. Soon, the second violin adds an Eb, forming an Ab chord and slightly alleviating the tension.   However, this is quickly undone by the A♮ played by the first violin on the next beat. Even though the viola moves to a G note to soften the dissonance, the G note continues to create discomfort, as it forms a major second interval with the A♮. Additionally, the Eb from the second violin, which was aligned with the Ab major chord, now forms an augmented fourth when paired with the A♮, creating the dissonance known as the "devil's interval." Eventually, the viola’s A resolves to F#, and the second violin’s Eb resolves to D, completing the resolution.   In this work, Mozart uses counterpoint, inherited from 16th-century church music, to carefully manage intervals, skillfully moving between consonance and dissonance. However, he stays within the limits of classical harmony and counterpoint, experimenting only within the boundaries of traditional tonality, leaving the development of functional harmony to his successors.     Dissonance, Mahler, and Beyond – Part 2 to follow     References [1] Jens Malte Fischer. Gustav Mahler. Yale University Press, 2011. [2] Hermann LF Helmholtz. On the Sensations of Tone as a Physiological Basis for the Theory of Music. Cambridge University Press, 2009. [3] Reinier Plompand Willem JM Levelt. “Tonal consonance and critical bandwidth”. In: Journal of the Acoustical Society of America38(1965), pp. 548–560.

2024.11.06