Hiroshi Saruwatari

12K posts

Hiroshi Saruwatari banner
Hiroshi Saruwatari

Hiroshi Saruwatari

@hsaruwatari727

音メディア信号処理に関する研究と教育を生業にしております。

Katılım Kasım 2009
43 Takip Edilen2.2K Takipçiler
Hiroshi Saruwatari
Hiroshi Saruwatari@hsaruwatari727·
In addition, we have 3 journal-paper presentation. YANG: Speaker-conditioned phrase break prediction for TTS IMAMURA: Stride conversion for sampling-frequency-independent DNNs SEKI: TTSOPS: A Closed-loop corpus optimization for training multi-speaker TTS models from dark data
Hiroshi Saruwatari@hsaruwatari727

The following 4 papers are ACCEPTED in ICASSP2026. YANG: Layer-wise self-distillation for MOS prediction SEKI: Learning spatially-aware audio-text embeddings IMAMURA: Dissecting performance degradation in sampling-mismatch SS NAKATA: Fast & Robust Multilingual Speech Restoration

English
0
3
8
2.1K
Hiroshi Saruwatari
Hiroshi Saruwatari@hsaruwatari727·
The following 4 papers are ACCEPTED in ICASSP2026. YANG: Layer-wise self-distillation for MOS prediction SEKI: Learning spatially-aware audio-text embeddings IMAMURA: Dissecting performance degradation in sampling-mismatch SS NAKATA: Fast & Robust Multilingual Speech Restoration
English
0
5
20
3.5K
Hiroshi Saruwatari retweetledi
Kentaro Seki / 関健太郎
Kentaro Seki / 関健太郎@trgkpc·
Our paper Spatial-CLAP has been accepted to #ICASSP2026! Many thanks to my amazing co-authors. Also, our IEEE TASLP paper (TTSOps) has been accepted for Journal Presentation. See you in Barcelona! 🇪🇸
Kentaro Seki / 関健太郎@trgkpc

New on arXiv: Spatial-CLAP 🎧📝 Can CLAP represent stereo audio even under multi-source conditions? Yes, Spatial-CLAP enables it. And, our spatial contrastive learning enforces robust and generalizable representations. 🔗 Full details in our preprint: arxiv.org/abs/2509.14785

English
1
19
65
15K
Hiroshi Saruwatari
Hiroshi Saruwatari@hsaruwatari727·
C君、学振採択おめでとうございます！
日本語
0
0
10
1.1K
Hiroshi Saruwatari retweetledi
C君
C君@__Ckun__·
学振DC2（情報学）二次採用内定でした 自信無くなってたので一安心
C君 tweet media
日本語
1
4
191
22.8K
Hiroshi Saruwatari
Hiroshi Saruwatari@hsaruwatari727·
Our paper titled "Excitement-inducing commentary text-to-speech system for fighting game video scenes" has been ACCEPTED for publication in IEEE ACCESS. Conglaturation, Iura-san and Saito-sensei!
English
0
3
8
851
Hiroshi Saruwatari retweetledi
Kentaro Seki / 関健太郎
Kentaro Seki / 関健太郎@trgkpc·
Our paper titled "Toward Data-Efficient Speech Synthesis: Active Learning–Based Corpus Construction for Multi-Speaker Text-to-Speech Synthesis" has been ACCEPTED for publication in IEEE Access! Many thanks to great co-authors!
English
0
4
22
6.5K
Hiroshi Saruwatari
Hiroshi Saruwatari@hsaruwatari727·
弊研M1の錦織君が日本音響学会2025年秋季研究発表会における以下の発表に関して、学生優秀発表賞を受賞いたしました。おめでとうございます！ 錦織広尚 他「空間正則化付きILRMAと雑音事前分布ありランク制約付き空間共分散行列推定法を用いたドローン搭載マイクロホンアレーによる音声抽出」
日本語
0
5
19
1.4K
Hiroshi Saruwatari
Hiroshi Saruwatari@hsaruwatari727·
Our paper titled "Stride Conversion Algorithms for Convolutional Layers and Its Application to Sampling-Frequency-Independent Deep Neural Networks Signal Processing" has been ACCEPTED for publication in Signal Processing (Elsevier). Congratulations, Imamura-kun!
English
0
4
8
1.1K
Hiroshi Saruwatari retweetledi
Yuki Saito
Yuki Saito@ysaito_human·
The following paper has been accepted for the Speech Communication journal: D. Yang et al., "Speaker-Conditioned Phrase Break Prediction for Text-to-Speech with Phoneme-Level Pre-trained Language Model" Congrats!!👏👏👏
English
0
5
15
1.9K
Hiroshi Saruwatari
Hiroshi Saruwatari@hsaruwatari727·
Our paper titled "TTSOps: A Closed-Loop Corpus Optimization Framework for Training Multi-Speaker TTS Models from Dark Data" has been ACCEPTED for publication in IEEE Trans. on Audio, Speech and Language Processing. Congratulations, Seki-kun!
English
0
7
15
10.3K
Hiroshi Saruwatari retweetledi
k_imamura
k_imamura@imamura_asp·
Journal presentationで発表しました！ 論文の方も読んでいただけるととても嬉しいです🙇‍♂️ nowpublishers.com/article/Detail…
k_imamura tweet media
日本語
0
3
15
1.7K
Hiroshi Saruwatari
Hiroshi Saruwatari@hsaruwatari727·
ミニムーグの価格高騰がすごい。都内のシンセ専門店では80万円を超えている。ヤフオクでも60万円超え。私が20年前に買ったときは30万円で買えた。1990年代の音楽雑誌の広告を見ると10万円台後半でも買えたらしい。基本、電子楽器は年とともに価格下落するけど、こういう銘器は別ですね。
日本語
0
0
2
693
Hiroshi Saruwatari
Hiroshi Saruwatari@hsaruwatari727·
今日久しぶりに近所のハードオフ行ってきたけど、ギターコーナーは「ハードオフドリーム（そこそこ価値のある楽器がなぜか破格で売られている事）」がありそうだったけど、キーボードは全くそんなことないんですよね。ミニムーグとかプロフェット5とかが1万円なんてのは見たことがないですよねえ。
日本語
0
0
1
861
Hiroshi Saruwatari
Hiroshi Saruwatari@hsaruwatari727·
Cont. 5. Continuous Function Approximation of Convolutional Kernels for Sampling Frequency Adaptation of Pre-trained Source Separation Networks 6. Analysis of a dataset for evaluating semantic relevance between text and audio
Hiroshi Saruwatari@hsaruwatari727

We present papers in ASA-ASJ Joint Meeting: 1. Low-latency real-time BSS using asymmetric window 2. Switching distortionless BSS in underdetermined scenarios 3. Real-time hearing assistance system combining BSS and VC 4. TTS by perceptual rating parallel iterative decoding

English
0
2
4
1.7K
Hiroshi Saruwatari
Hiroshi Saruwatari@hsaruwatari727·
We present papers in ASA-ASJ Joint Meeting: 1. Low-latency real-time BSS using asymmetric window 2. Switching distortionless BSS in underdetermined scenarios 3. Real-time hearing assistance system combining BSS and VC 4. TTS by perceptual rating parallel iterative decoding
English
0
3
10
2.8K
Hiroshi Saruwatari retweetledi
Kentaro Seki / 関健太郎
Kentaro Seki / 関健太郎@trgkpc·
#yans2025 での下記発表につきまして、PKSHA Technologys様（@PKSHA_saiyo）よりスポンサー賞をいただきました。 [S1-P07] ステレオ信号に対する空間情報を伴う音響キャプショニング 本研究の実装上の意義を評価頂いたとのことで、大変光栄です！🥳 （arXiv版も是非ご覧ください！） #YANS2025
Kentaro Seki / 関健太郎@trgkpc

New on arXiv: Spatial-CLAP 🎧📝 Can CLAP represent stereo audio even under multi-source conditions? Yes, Spatial-CLAP enables it. And, our spatial contrastive learning enforces robust and generalizable representations. 🔗 Full details in our preprint: arxiv.org/abs/2509.14785

日本語
0
8
59
7.2K

Keşfet

@PKSHA_saiyo @elonmusk @BarackObama @taylorswift13 @cristiano @BillGates @NASA @nikifrancismediavine