Datasets:
audio audio | id string | text string | duration_s float64 | categories list | target_terms list | prompt_id string | prompt_source string | mode string | edited bool | device string | environment string | sample_rate int64 | recorded_at string | microphone string | prompt_origin string | hebrew_pronunciation string | variant_groups list | variant_spelling string | speaker_id string | conditions list | imported_from string | original_text string | text_before_normalization string | text_he string | noise_level string | background_speech_languages list | notes string | model_hypothesis string | asr_model string | speaker string | speaker_gender string | noise_flags list | clipped bool | task string | spoken_text string | aside_text string | aside_position string | verbatim_text string | prompt_categories list | source_prompt_id string | prompt_kind string | prompt_note string | term_language string | paragraph_text string | contributor string | bit_depth int64 | audio_format string | native_rate int64 | capture_mode string | tts_eligible bool | tts_split string | transcript_status string | prompt_set string | topic string | style string | style_instruction string | written_form_uncertain string | session_id string | room_tone_id string | lead_trim_s float64 | peak_dbfs float64 | floor_dbfs float64 | speech_dbfs float64 | snr_db float64 | clipped_samples int64 | edge_head_dbfs float64 | edge_tail_dbfs float64 | browser_processing bool | room string | priority_word string | location string | location_name string | location_note string | latitude string | longitude string | location_accuracy_m string | text_full string | fallback_ok bool | background_speech_kinds list | extra string |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
20261015T055645Z_u24ade | Can you send me the kishur for the meeting? I'll look at it tonight, and then we can talk properly machar morning. | 7.078 | [
"message",
"code_switch"
] | [
"kishur",
"machar"
] | p_b4a691da6f | generated:anthropic/claude-sonnet-5.5 | bank | false | OnePlus Nord 3 5G | quiet-indoor | 16,000 | 2026-10-15T05:56:45.768Z | internal | generated | null | [] | null | 8992 | [
"background-noise",
"people-talking"
] | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | |
20261015T055700Z_v6yzls | Rothschild boulevard is lined with beautiful old buildings. | 2.87 | [
"code_switch"
] | [
"Rothschild"
] | p_eec49e0fd3 | english-hebrew-mixed-sentences | bank | false | OnePlus Nord 3 5G | quiet-indoor | 16,000 | 2026-10-15T05:57:00.403Z | internal | dataset | null | [] | null | 8992 | [
"background-noise",
"people-talking"
] | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | |
20261015T061540Z_vtfjbq | Hey, can you set up the sensor gateway for the Mitzpe Ramon site before Thursday? Then I'll take the spare router over to Holon. | 7.35 | [
"message",
"names"
] | [
"Mitzpe Ramon",
"Holon"
] | p_8f5d399a81 | generated:anthropic/claude-sonnet-5.5 | bank | false | OnePlus Nord 3 5G | quiet-indoor | 16,000 | 2026-10-15T06:15:40.412Z | internal | generated | null | [] | null | 8992 | [
"background-noise",
"people-talking"
] | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | |
20261015T062011Z_fzz4qk | Tomer Yosef is going to be here in one daka, so tell Savta we're ready to leave and she should wait by the door. | 6.95 | [
"message",
"code_switch",
"names"
] | [
"daka",
"Tomer",
"Yosef"
] | p_fde50e288b | generated:anthropic/claude-sonnet-5.5 | bank | false | OnePlus Nord 3 5G | quiet-indoor | 16,000 | 2026-10-15T06:20:11.974Z | internal | generated | null | [] | null | 8992 | [
"background-noise",
"people-talking"
] | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | |
20261015T062020Z_2fdrwx | Is anyone up for kaduregel on Friday morning at the park near the house, or are we all too tired from the week? | 5.758 | [
"message",
"code_switch"
] | [
"kaduregel"
] | p_a8ecaf6469 | generated:anthropic/claude-sonnet-5.5 | bank | false | OnePlus Nord 3 5G | quiet-indoor | 16,000 | 2026-10-15T06:20:20.864Z | internal | generated | null | [] | null | 8992 | [
"background-noise",
"people-talking"
] | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | |
20261015T062028Z_dzzldl | Is your company structured as a ba'am? | 3.046 | [
"code_switch"
] | [] | p_c448755d18 | english-hebrew-mixed-sentences | bank | false | OnePlus Nord 3 5G | quiet-indoor | 16,000 | 2026-10-15T06:20:28.067Z | internal | dataset | null | [] | null | 8992 | [
"background-noise",
"people-talking"
] | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | |
20261015T062040Z_mhna1y | The doctor sent the mirsham electronically to the pharmacy. | 4.19 | [
"code_switch"
] | [
"mirsham"
] | p_ca25e185f3 | english-hebrew-mixed-sentences | bank | false | OnePlus Nord 3 5G | quiet-indoor | 16,000 | 2026-10-15T06:20:40.664Z | internal | dataset | null | [] | null | 8992 | [
"background-noise",
"people-talking"
] | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | |
20261015T062225Z_hrt3os | I tried the new build on Linux and af ehad from the team has hit that crash yet, so I think we're fine. | 6.974 | [
"tech",
"code_switch"
] | [
"af ehad",
"Linux"
] | p_135e9cf120 | generated:anthropic/claude-sonnet-5.5 | bank | false | OnePlus Nord 3 5G | quiet-indoor | 16,000 | 2026-10-15T06:22:25.553Z | internal | generated | null | [] | null | 8992 | [
"background-noise",
"people-talking"
] | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | |
20261015T062236Z_24zf5u | Hi Sigal, thanks for asking. The parsha is not something I know much about, but I can ask my father-in-law, who reads it every week. | 9.142 | [
"email",
"code_switch",
"names",
"religious"
] | [
"parsha",
"Sigal"
] | p_249b817524 | generated:anthropic/claude-sonnet-5.5 | bank | false | OnePlus Nord 3 5G | quiet-indoor | 16,000 | 2026-10-15T06:22:36.230Z | internal | generated | null | [] | null | 8992 | [
"background-noise",
"people-talking"
] | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | |
20261015T062247Z_avfnbz | For this architecture, we're considering either Linkerd or Istio as our service mesh solution, with Envoy handling the proxy layer. | 10.022 | [
"tech"
] | [
"Linkerd",
"Istio",
"Envoy"
] | p_826e2748e5 | tech-sentences-for-asr-training | bank | false | OnePlus Nord 3 5G | quiet-indoor | 16,000 | 2026-10-15T06:22:47.932Z | internal | dataset | null | [] | null | 8992 | [
"background-noise",
"people-talking"
] | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | |
20261015T062300Z_sc1aro | Remember to bring a clean kos for the Kiddush tomorrow, the nice glass one from the cabinet, not the plastic ones. | 8.574 | [
"message",
"code_switch",
"religious"
] | [
"kos"
] | p_b2b1141423 | generated:anthropic/claude-sonnet-5.5 | bank | false | OnePlus Nord 3 5G | quiet-indoor | 16,000 | 2026-10-15T06:23:00.463Z | internal | generated | null | [
"kiddush"
] | null | 8992 | [
"background-noise",
"people-talking"
] | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null |
Background Speech
Every take from the main dataset with another voice audible behind the speaker: someone talking in the room, a market crowd, a recorded PA message, TV or radio. Kept apart for work on speaker isolation and target-speaker ASR, where the transcript is the main speaker's words only and the background speech is not transcribed. background_speech_kinds says what the speech is (conversation, crowd, pa-announcement, tv-radio, other), background_speech_languages its language (en, he, ar, ru, other); both are empty when not tagged. The main speaker is always Daniel; the voices behind him are other people's and are not in the transcript. Takes in this set also appear in Hebrish/Noisy-Condition-Sentences.
Website: hebrish.app: get the Hebrish keyboard for Android and the desktop apps, see the results, and find every model, dataset and demo in one place.
This is a subset of danielrosehill/Read-Aloud-Voice-Sentences, the full, undifferentiated
dataset. It contains every phone recording there with the people-talking condition ("Background speech" in the app) or a background speech language; a take later labelled by listening counts as labelled (background speech language or level, or a PA announcement: in; judged clean: out). Rows are copied
with the same ids and train/test split, plus the columns named above, by scripts/sync_subsets.py in the
project repository. Do not edit this dataset directly: fix the row in the main dataset and re-sync.
The subset grows as new takes are published.
Speaker: Daniel Rosehill, the dataset's author, who was born in Ireland and speaks English with an Irish accent, and who lives in Jerusalem. Every recording is his voice. Deliberate variation: he sometimes changes his speaking speed on purpose (fast and rushed, or slow) and sometimes says a word the way people commonly mispronounce it, to make the models robust to real speech. The transcript is always the sentence as written, so a take whose speed or pronunciation sounds off against its card is usually intentional, not a recording error.
All audio was recorded on a OnePlus Nord 3 5G phone using its internal microphone, with browser echo cancellation, noise suppression and auto gain turned off.
Layout
data/train-*.parquet one row per take: `audio` (the FLAC or WAV file, embedded) and every column
meta/train.jsonl index: the same rows without audio, plus `shard`, `audio_path` and the audio's byte range
meta/schema.json column types
data/test-*.parquet, meta/test.jsonl held-out set, never trained on
Audio is 48 kHz mono 24-bit FLAC for rows from 2026-10-22 and 16 kHz mono 16-bit WAV before.
The test split is term-disjoint: a held-out term never appears in train.
load_dataset("Hebrish/Background-Speech") works directly.
Columns
The columns are the same as in the main dataset: text (ground truth, punctuated and cased like
Parakeet v2 output), text_he (the same transcript with its Hebrew terms in Hebrew script, for
Hebrew-capable models), categories, target_terms, prompt_source, prompt_origin (manual / dataset / generated), edited, duration_s,
environment, conditions, device, speaker, speaker_gender, contributor, hebrew_pronunciation, variant_groups, variant_spelling and recorded_at. The main dataset's card describes them.
Licence
CC-BY-4.0, with sentence-text attribution per prompt_source as described in the
main dataset's licence section.
- Downloads last month
- -