Channel
@speechinfo
On this record: Growth · Engagement · Reactions · Posts · Citations · Cite this entry
1,320subscribers
+4 since we began measuring on 6 August 2026
Risers and fallers across the register · movement among entries of 1,000–3,162.
| Telegram ID | -1002605317562 |
|---|---|
| Type | Channel |
| Username | @speechinfo |
| Created | Between 1 March 2025 and 31 July 2025— estimated from Telegram’s id allocation, not measured. How this range is calculated. |
| First recorded | 6 August 2026 |
| Last confirmed live | 13 August 2026 |
| Measurements held | 5 |
| Confirmed unchanged | 1 time, most recently 13 August 2026 |
| On Telegram | t.me/speechinfo |
| Measured (UTC) | Subscribers | Change |
|---|---|---|
| 13 Aug 2026, 00:03 | 1,320 | +2 |
| 10 Aug 2026, 00:02 | 1,318 | +3 |
| 6 Aug 2026, 21:21 | 1,315 | -1 |
| 6 Aug 2026, 14:01 | 1,316 | no change |
| 6 Aug 2026, 13:53 | 1,316 | first reading |
13 posts held, back to 29 May 2026 — the reader has not yet reached the start of this channel’s public history, so older posts may sit further back, unread. Read across 1 pageof Telegram’s post history, 20 posts per page.
ERR is average views per post over the last 30 days divided by subscribers, the definition TGStat uses, so this figure is comparable with the one you will see elsewhere. It falls structurally as a channel grows: a high ERR on a small channel and a low one on a large channel describe reach mathematics, not quality. We publish the figure and the sample it came from and pass no verdict on it.
ER is defined industry-wide as (forwards + reactions + comments) ÷ views— note the denominator is views, not subscribers. Telegram’s public web preview carries views and reactions but not forward or comment counts, so the reaction rate above is the reactions term only and is therefore a floor: the true ER for this channel is higher by an amount we have not measured and will not estimate.
| Window | Rolling 30 days · latest post in window 5 August 2026 |
|---|---|
| Posts held | 13 (29 May 2026 – 5 August 2026) |
| Views total | 3,805 |
| Reactions total | 134 |
| Forwards / comments | not exposed by the public surface — not measured, not estimated |
| Readings taken | 6 Aug 2026, 14:01 UTC |
Views are a single reading per post, taken at the time above. A post published in the last day or two is still accumulating views, which pulls the 30-day average down slightly. That is a property of the standard definition rather than a fault in it, so we keep the definition rather than “correcting” the number into something nobody can reproduce.
Precision. Telegram publishes view counts on its public widget in short form — 8.12K, 3.7M — so any reading at or above 1,000 reaches us rounded to three significant figures, and only counts below 1,000 are exact. Averages and rates derived from them are shown to the same precision rather than to the unit: a figure like 3,701,250 would assert digits nobody measured.
Reaction counts are published per emoji and rounded the same way, so a total below 1,000 is exact and a larger one is a sum that may carry a rounded component from each emoji above 1,000. Because it is a sum, it does not look rounded — read a large reaction total as three significant figures per contributing emoji rather than as the figure it prints.
365 reactions across 13 posts, in 13 distinct kinds. The most used accounts for 35.1% of them.
| Reaction | Count | Share | Share, drawn |
|---|---|---|---|
| ❤ | 128 | 35.1% | |
| 🔥 | 106 | 29.0% | |
| 👍 | 99 | 27.1% | |
| 👌 | 9 | 2.47% | |
| 💯 | 5 | 1.37% | |
| 🎅 | 4 | 1.10% | |
| 🤓 | 4 | 1.10% | |
| 🤩 | 3 | 0.822% | |
| ❤🔥 | 2 | 0.548% | |
| 👏 | 2 | 0.548% | |
| 🆒 | 1 | 0.274% | |
| 🕊 | 1 | 0.274% | |
| 🥰 | 1 | 0.274% |
No sentiment is inferred, and none should be read in. This table is ordered by count and by nothing else. Emoji do not carry stable meaning across languages or communities — 🙏 is thanks in one channel and mourning in another — so we publish which ones were pressed and how often, and pass no judgement on what an audience meant by them.
Precision. Telegram publishes reaction counts per emoji and short-forms each one — 4.34K, 1.2M — so any single kind at or above 1,000 reaches us at three significant figures, and only counts below 1,000 are exact. The shares above are ratios of those figures and carry the same error. This is also why the total here can differ slightly from a reaction total printed elsewhere on the page: both are sums of the same rounded parts, taken over samples with different edges.
Coverage. Reactions were read on 13 of the 13 sampled posts in this sample. Summed by Telegram’s own count on each post — not by adding up the per-emoji breakdown above — those same posts carry 365reactions in total: the kind of figure the paragraph above means by “a reaction total printed elsewhere on the page”.
Measured over the 13 most recent posts we hold, published 29 May 2026 to 5 August 2026, using the newest reading held for each. Telegram Stars are excluded: they are a payment, not a reaction, and they have their own section.
Interaction Models: A Scalable Approach to Human-AI Collaboration Каскад из VAD, ASR, LLM и TTS всё ещё остаётся одним из самых понятных способов собрать голосового агента. Компоненты можно независимо улучшать и менять распознавание, языковую модель, поиск, инструменты, и, конечно же, синтез. Но за модульность приходится платить — задержки компонентов складываются, а взаимодействие обычно остаётся пошаговым: сначала…
🔥16❤8👌4👍2🤩1
Mega-ASR: Towards In-the-wild^2 Speech Recognition via Scaling Up Real-world Acoustic Simulation Разбираем статью со скромным названием Mega-ASR. Основной вклад работы — улучшение распознавания сложных аудио. Также авторы собрали и выложили свой датасет — VOICES-IN-THE-WILD-2M. В реальных условиях, где есть разные шумы, ASR работает хуже. Так происходит, потому что шумов нет в данных. Кто-то пробует добавлять в дан…
❤11👍9🔥7🤓4🆒1
A review on subjective and objective evaluation of synthetic speech Сегодня разбираем обзор 2024 года, авторы которого попытались охватить 40 с лишним лет эволюции оценки синтеза речи — с 80-х и до наших дней. При этом статья не такая уж большая: 27 страниц, 7 из которых — ссылки на множество упомянутых работ. Первая часть, пожалуй, самая увлекательная — историческая справка о том, как оценивали синтез речи с самог…
❤10🎅4👍4🤩2❤🔥1💯1🔥1
Работы о голосовых технологиях на ICML 2026 [2/2] Продолжаем подборку работ от Максима Борисова. В первой части — общие впечатления о конференции и несколько интересных статей. CoCoEmo: Composable and Controllable Emotional TTS via Activation Steering Эмоция в речи часто состоит из нескольких, порой противоречивых оттенков, которые могут не совпадать с текстом, но TTS обычно навязывают одну эмоцию на всё высказыва…
🔥11👍8❤7
Работы о голосовых технологиях на ICML 2026 [1/2] С 6 по 11 июля в Сеуле проходила International Conference on Machine Learning, на которой побывал наш коллега Максим Борисов. Он поделился впечатлениями и подборкой работ на тему Speech. По спичу представленность была довольно ограниченной — большинство релевантных работ можно было найти на постерах, а не в основных треках и oral-докладах. Зато именно постерные сесс…
👍10🔥6❤5❤🔥1
Voxtral Realtime Разбираем статью о модели Voxtral Realtime, в которой предложили ещё один способ, как стримить speech-to-text. Авторы утверждают, что у них получилось сделать модель, в которой можно контролировать латенси и которая при задержке в полсекунды имеет паритет по качеству с Whisper. Если увеличить задержку до секунды, то модель начинает превосходить Whisper, ещё больше — например, 2,5 секунды, — качеств…
❤8👌5👍5💯2
UALM: Unified Audio Language Model for Understanding, Generation and Reasoning В более ранних статьях аудиопонимание и генерация традиционно шли параллельно и не пересекались. Но, если задуматься, человек, решая задачу в области аудио, одновременно мыслит словами и воспринимает звук, постоянно переключая в голове эти модальности — например, так происходит, когда композитор пишет музыку. Объединить аудиопонимание, г…
❤8🔥7👍4💯2🕊1
Ускорили перевод видео в Яндекс Браузере — задистиллировали диффузионный декодер TTS Сегодня делимся свежей хабростатьёй о том, как ускорили синтез речи при переводе видео в Яндекс Браузере. С чего стартовали Внутри TTS — каскад из трёх частей: 🔴языковая модель предсказывает аудиотокены по тексту; 🔴диффузионный декодер восстанавливает мел-спектрограмму из латентов; 🔴вокодер превращает её в звуковую волну. После т…
❤20🔥11👍6
Chunk-wise Attention Transducers for Fast and Accurate Streaming Speech-to-Text В NVIDIA есть несколько сотрудников, которые стабильно пишут интересные статьи об ASR в целом и RNN-T в частности. Примеры таких работ — FastConformer, TDT, WIND. Сегодня расскажем о CHAT, суть которого также в улучшении RNN-T. Но сначала вспомним, что это такое. Recurrent Neural Network Transducer — архитектура для распознавания и пере…
👍16❤11🔥7👏1🥰1
Как устроена голосовая активация в Яндекс Дропс Недавно Яндекс запустил свои первые ИИ-наушники — Яндекс Дропс. В числе прочего они умеют распознавать обращение «Алиса», а отвечает за эту способность компонент, который мы внутри называем «споттером» (чуть подробнее писали о споттерах тут). И если с голосовой активацией в колонках всё плюс-минус понятно, то перенести её в наушники — это челлендж. О том, что было сло…
🔥16❤11👍11👏1
Reward-Driven Interaction: Enhancing Proactive Dialogue Agents through User Satisfaction Prediction Разбираем статью об улучшении диалоговых агентов с помощью «проактивности». Речь о способности системы в нужный момент задать уточняющий вопрос, если она понимает, что пользователь, скорее всего, останется недоволен ответом. Применяют обычный для голосового ассистента каскадный пайплайн: отдельный ASR, переформулиров…
❤10👍8🔥7
Три работы о том, как сделать речь полноценной модальностью для LLM В сегодняшней подборке — три любопытные идеи: от генерации голосового ответа с ризонингом без лишней задержки до более компактных речевых представлений и подготовки аудиоданных для мультимодального претрейна. STITCH: Simultaneous Thinking and Talking with Chunked Reasoning for Spoken Language Models Статья Microsoft о том, как добавить ризонинг в …
❤13🔥11👍7
Showing the 12 most recent of 13 posts we hold for @speechinfo. View and reaction counts are the latest single reading for each post, not a live figure, and a recent post is still accumulating both. A view count marked ≈ was rounded by Telegram before we ever saw it — t.me prints views in full below 1,000 and to three significant figures above, so ≈1,200,000 means somewhere between 1,150,000 and 1,249,999. Unmarked counts are exact. Text is reproduced from the public post preview and truncated for length.
Citation-graph rank — 61,494 of 1,160,990entries in the measured graph. A weighted position computed from the forward and mention edges below — republished posts weigh more than named mentions — and recomputed periodically, over the whole graph. Published only as this ordinal position, never as a score: a position is a fact, and a score printed beside one channel’s name would read as a verdict this register does not make. The two counts beneath stay separate for the same reason mentions are never summed with forwards anywhere else on this page — a named-by count costs nothing to manufacture. The top 100 by this measure, or how it is computed.
Republished by
Channels on the register that have forwarded this channel's posts into their own feed.
Built only from forwarded posts we have actually read, on both sides. Coverage is early and deliberately incomplete: a missing link means we have not read the post that would prove it, never that the relationship does not exist. Counts are distinct forwarded posts observed, so they only ever go up as we read more.
Named by 1 registered channel — every channel on the register whose own posts have named this one, by its current username or any other username it currently holds, merged from two separately captured readings of the same fact so a namer caught by only one of them is not missed and a namer both caught is not counted twice. A username this channel has since dropped is not matched — that handle may belong to someone else now, and crediting today’s namer to yesterday’s owner would misattribute it.
Named by
Channels on the register whose posts name this channel's handle.
Names
Channels on the register whose handles appear in this channel's posts.
A mention is a weaker signal than a forward and is counted separately for that reason — naming a channel is not republishing it, and a handle in a post body is easy to place deliberately. The post counts beside each row below are distinct posts in which the handle appeared, from posts we have read on both sides — the “Named by N registered channels” figure above is a different count, of distinct NAMING CHANNELS rather than posts, and is not the sum of the rows under it.
A live page changes as we take new readings, so a citation should name the measurement it is based on, not just the URL. The line below cites the subscriber count as measured 13 August 2026 — this entry's latest reading, not the date you are reading this.
“Speech Info” (@speechinfo), 1,320 subscribers as measured 13 August 2026. Telegram Register, tgregister.com/channel/speechinfo.
Full measurement history, CC BY 4.0. Every reading this register holds for this entry, not just the latest one, as a dated, downloadable record: CSV · JSON. Free to use with attribution to tgregister.com. Each file carries its own generation timestamp, which is the figure to cite for exactly when the data was retrieved.