Datasets:
audio audioduration (s) 8.57 26.3 | id stringclasses 5
values | text stringclasses 5
values | duration_s float64 8.57 26.3 | categories listlengths 1 1 | target_terms listlengths 0 0 | prompt_id stringclasses 5
values | prompt_source stringclasses 2
values | mode stringclasses 1
value | edited bool 1
class | device stringclasses 1
value | environment stringclasses 1
value | sample_rate int64 16k 48k | recorded_at stringclasses 5
values | microphone stringclasses 1
value | prompt_origin stringclasses 2
values | hebrew_pronunciation stringclasses 0
values | variant_groups listlengths 0 0 | variant_spelling stringclasses 0
values | speaker_id stringdate 8992-01-01 00:00:00 8992-01-01 00:00:00 | conditions listlengths 0 0 | imported_from stringclasses 0
values | original_text stringclasses 0
values | text_before_normalization stringclasses 0
values | text_he stringclasses 5
values | noise_level stringclasses 0
values | background_speech_languages listlengths 0 0 | notes stringclasses 0
values | model_hypothesis stringclasses 0
values | asr_model stringclasses 0
values | speaker stringclasses 1
value | speaker_gender stringclasses 1
value | noise_flags listlengths 0 0 | clipped bool 1
class | task stringclasses 1
value | spoken_text stringclasses 5
values | aside_text stringclasses 0
values | aside_position stringclasses 0
values | verbatim_text stringclasses 0
values | prompt_categories listlengths | source_prompt_id stringclasses 0
values | prompt_kind stringclasses 2
values | prompt_note stringclasses 2
values | term_language stringclasses 0
values | paragraph_text stringclasses 5
values | contributor stringclasses 1
value | bit_depth int64 16 16 | audio_format stringclasses 2
values | native_rate int64 48k 48k ⌀ | capture_mode stringclasses 1
value | tts_eligible bool 1
class | tts_split stringclasses 0
values | transcript_status stringclasses 1
value | prompt_set stringclasses 0
values | topic stringclasses 0
values | style stringclasses 0
values | style_instruction stringclasses 0
values | written_form_uncertain bool 0
classes | session_id stringclasses 2
values | room_tone_id stringclasses 0
values | lead_trim_s float64 0 0 ⌀ | peak_dbfs float64 -2.5 -1.7 ⌀ | floor_dbfs float64 -51.3 -50.6 ⌀ | speech_dbfs float64 -17.1 -16.3 ⌀ | snr_db float64 34.2 34.4 ⌀ | clipped_samples int64 0 0 ⌀ | edge_head_dbfs float64 -49.6 -25 ⌀ | edge_tail_dbfs float64 -42.6 -38.1 ⌀ | browser_processing bool 1
class | room stringclasses 0
values | priority_word stringclasses 0
values | location stringclasses 0
values | location_name stringclasses 0
values | location_note stringclasses 0
values | latitude float64 | longitude float64 | location_accuracy_m float64 | text_full stringclasses 0
values | fallback_ok bool 0
classes | sentence_id stringclasses 5
values | pronunciation stringclasses 1
value | extra stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
20261021T114057Z_vd8qv4 | Hi Sarah, thanks for sending over the draft contract yesterday. I've had a chance to read through it properly and overall it looks good. There are two things I'd like to change. The payment terms should be 30 days rather than 60, and the termination clause needs a notice period on both sides, not just ours. If you can ... | 8.574 | [
"paragraphs"
] | [] | paragraphs_e16e8f6378 | generated:claude-self-written | bank | false | OnePlus Nord 3 5G | quiet-indoor | 16,000 | 2026-10-21T11:40:57.700Z | internal | generated | null | [] | null | 8992 | [] | null | null | null | Hi Sarah, thanks for sending over the draft contract yesterday. I've had a chance to read through it properly and overall it looks good. There are two things I'd like to change. The payment terms should be 30 days rather than 60, and the termination clause needs a notice period on both sides, not just ours. If you can ... | null | [] | null | null | null | Daniel Rosehill | male | [] | false | paragraphs | Hi Sarah, thanks for sending over the draft contract yesterday. I've had a chance to read through it properly and overall it looks good.
There are two things I'd like to change. The payment terms should be 30 days rather than 60, and the termination clause needs a notice period on both sides, not just ours.
If you ca... | null | null | null | null | null | email | null | null | Hi Sarah, thanks for sending over the draft contract yesterday. I've had a chance to read through it properly and overall it looks good.
There are two things I'd like to change. The payment terms should be 30 days rather than 60, and the termination clause needs a notice period on both sides, not just ours.
If you ca... | Daniel Rosehill | 16 | wav | null | null | null | null | script | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | p_bef8c2d108 | [] | null | |
20261021T114138Z_k2rfvl | Hi Tom, just a quick update on the server migration. Everything moved over on Saturday night and the new machine has been running without problems since. The only thing still outstanding is the backup job. It's configured, but I want to see three clean nightly runs before I switch off the old box. I'll let you know on ... | 8.894 | [
"paragraphs"
] | [] | paragraphs_d7c6a2cfcb | generated:claude-self-written | bank | false | OnePlus Nord 3 5G | quiet-indoor | 16,000 | 2026-10-21T11:41:38.398Z | internal | generated | null | [] | null | 8992 | [] | null | null | null | Hi Tom, just a quick update on the server migration. Everything moved over on Saturday night and the new machine has been running without problems since. The only thing still outstanding is the backup job. It's configured, but I want to see three clean nightly runs before I switch off the old box. I'll let you know on ... | null | [] | null | null | null | Daniel Rosehill | male | [] | false | paragraphs | Hi Tom, just a quick update on the server migration. Everything moved over on Saturday night and the new machine has been running without problems since.
The only thing still outstanding is the backup job. It's configured, but I want to see three clean nightly runs before I switch off the old box.
I'll let you know o... | null | null | null | null | null | email | null | null | Hi Tom, just a quick update on the server migration. Everything moved over on Saturday night and the new machine has been running without problems since.
The only thing still outstanding is the backup job. It's configured, but I want to see three clean nightly runs before I switch off the old box.
I'll let you know o... | Daniel Rosehill | 16 | wav | null | null | null | null | script | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | p_c52530606f | [] | null | |
20261021T115304Z_f06vjg | Dear Ms. Clarke, I'm writing about the invoice dated 14 September for the annual software licence. The amount charged was for 25 seats, but our agreement covers 15. I've attached the original order confirmation, which shows the correct number. Could you please issue a corrected invoice? I'm happy to pay the difference ... | 26.338 | [
"paragraphs"
] | [] | paragraphs_f6103d46b1 | generated:claude-self-written | bank | false | OnePlus Nord 3 5G | quiet-indoor | 16,000 | 2026-10-21T11:53:04.145Z | internal | generated | null | [] | null | 8992 | [] | null | null | null | Dear Ms. Clarke, I'm writing about the invoice dated 14 September for the annual software licence. The amount charged was for 25 seats, but our agreement covers 15. I've attached the original order confirmation, which shows the correct number. Could you please issue a corrected invoice? I'm happy to pay the difference ... | null | [] | null | null | null | Daniel Rosehill | male | [] | false | paragraphs | Dear Ms. Clarke, I'm writing about the invoice dated 14 September for the annual software licence.
The amount charged was for 25 seats, but our agreement covers 15. I've attached the original order confirmation, which shows the correct number.
Could you please issue a corrected invoice? I'm happy to pay the differenc... | null | null | null | null | null | email | null | null | Dear Ms. Clarke, I'm writing about the invoice dated 14 September for the annual software licence.
The amount charged was for 25 seats, but our agreement covers 15. I've attached the original order confirmation, which shows the correct number.
Could you please issue a corrected invoice? I'm happy to pay the differenc... | Daniel Rosehill | 16 | wav | null | null | null | null | script | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | p_af30faf85b | [] | null | |
20261119T143444Z_2abuti | I thought the film about the lighthouse was far too slow. But the acting from the two leads was truly excellent. And the ending left me thinking for days. So I would still say it is worth your time. | 11.423 | [
"paragraphs"
] | [] | paragraphs_42827b357b | deepseek-v4.1-flash | bank | false | OnePlus Nord 3 5G | quiet-indoor | 48,000 | 2026-11-19T14:34:44.555Z | internal | dataset | null | [] | null | 8992 | [] | null | null | null | I thought the film about the lighthouse was far too slow. But the acting from the two leads was truly excellent. And the ending left me thinking for days. So I would still say it is worth your time. | null | [] | null | null | null | Daniel Rosehill | male | [] | false | paragraphs | I thought the film about the lighthouse was far too slow. But the acting from the two leads was truly excellent. And the ending left me thinking for days. So I would still say it is worth your time. | null | null | null | null | null | paragraph | conjunction | null | I thought the film about the lighthouse was far too slow. But the acting from the two leads was truly excellent. And the ending left me thinking for days. So I would still say it is worth your time. | Daniel Rosehill | 16 | flac | 48,000 | stt | true | null | script | null | null | null | null | null | 20261119T141658Z_rcgu96 | null | 0 | -1.7 | -51.3 | -17.1 | 34.2 | 0 | -25 | -42.6 | false | null | null | null | null | null | null | null | null | null | null | p_f46abbe4f0 | [] | {"mic_link":null,"bt_codec":null,"mic_name":"CPH2493","mic_usb_vendor_id":null,"mic_usb_product_id":null,"mic_type":"builtin_mic","mic_chain":null,"activity":"seated","background_speech_kinds":[],"pronunciation_variants":[],"ui_note":null,"source_url":null,"check_verdict":null,"check_errors":null,"check_error_types":nu... | |
20261119T143908Z_7nhq0j | I've been using the Tallybrook app all week and the daily summary screen is genuinely lovely. But the filter panel is awkward on Android, I keep missing the date row by a few pixels, so maybe just enlarge the tap area. And the export button, I feel like it's buried, so perhaps pin it near the top. Set the default range... | 19.637 | [
"paragraphs"
] | [] | paragraphs_8c6241efe3 | deepseek-v4.1-flash | bank | false | OnePlus Nord 3 5G | quiet-indoor | 48,000 | 2026-11-19T14:39:08.389Z | internal | dataset | null | [] | null | 8992 | [] | null | null | null | I've been using the Tallybrook app all week and the daily summary screen is genuinely lovely. But the filter panel is awkward on Android, I keep missing the date row by a few pixels, so maybe just enlarge the tap area. And the export button, I feel like it's buried, so perhaps pin it near the top. Set the default range... | null | [] | null | null | null | Daniel Rosehill | male | [] | false | paragraphs | I've been using the Tallybrook app all week and the daily summary screen is genuinely lovely. But the filter panel is awkward on Android, I keep missing the date row by a few pixels, so maybe just enlarge the tap area. And the export button, I feel like it's buried, so perhaps pin it near the top. Set the default range... | null | null | null | null | null | paragraph | typology:Product/UI feedback and feature requests on an app or website | null | I've been using the Tallybrook app all week and the daily summary screen is genuinely lovely. But the filter panel is awkward on Android, I keep missing the date row by a few pixels, so maybe just enlarge the tap area. And the export button, I feel like it's buried, so perhaps pin it near the top. Set the default range... | Daniel Rosehill | 16 | flac | 48,000 | stt | true | null | script | null | null | null | null | null | 20261119T143609Z_tdk020 | null | 0 | -2.5 | -50.6 | -16.3 | 34.4 | 0 | -49.6 | -38.1 | false | null | null | null | null | null | null | null | null | null | null | p_508a507fc5 | [] | {"mic_link":null,"bt_codec":null,"mic_name":"CPH2493","mic_usb_vendor_id":null,"mic_usb_product_id":null,"mic_type":"builtin_mic","mic_chain":null,"activity":"seated","background_speech_kinds":[],"pronunciation_variants":[],"ui_note":null,"source_url":null,"check_verdict":null,"check_errors":null,"check_error_types":nu... |
Paragraph Dictation Passages
Audio dataset, split-out. Rows copied from the master, danielrosehill/Read-Aloud-Voice-Sentences, so one task can be browsed alone. Training reads the master, not this copy. Every audio dataset as one tree, with hours: the audio index. How every dataset fits together: the dataset index.
Experimental. Passages of 2 to 4 paragraphs (emails, prompts to AI coding agents, voice notes, messages, meeting notes, blog drafts) dictated in one take by one speaker, pausing naturally between paragraphs. text is the passage on one line, the transcript Parakeet v2 can emit (its tokenizer has no newline). paragraph_text is the same passage with the paragraph breaks as blank lines, the label for a model that puts the breaks back. prompt_kind is the register.
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.
Two tabs. sentences lists every ground-truth sentence with a takes count, so
takes = 0 shows what has not been recorded. takes has one row per recording with its audio, joined on sentence_id; a
sentence can have several takes. load_dataset("Hebrish/Paragraph-Dictation-Passages", "sentences") reads the text; the default config reads the takes.
Takes come from danielrosehill/Read-Aloud-Voice-Sentences, the full dataset, with the same ids and
train/test split, by scripts/sync_subsets.py in the project repository. Do not edit this dataset directly. New takes append and
only the sentence counts are rewritten.
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
sentences/train.parquet one row per sentence: `id`, `text`, `takes`, `prompt_ids`, `served` and the sentence columns
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/Paragraph-Dictation-Passages") works directly.
Columns
A take's sentence_id is its row's id in sentences. A sentence with served: false is no longer offered by the app; it is kept because a take used it. 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
- 102