Multilingual speech datasets: a technical intake guide Validate language labels, transcripts, scripts and model inputs before comparing multilingual speech corpora. Includes an illustrative normalization test and intake record. A multilingual speech dataset needs more than a list of supported languages. An engineering intake should verify the language and locale labels the source can support, the written reference convention, code-switching treatment, audio preparation and independent evaluation groups. This guide focuses on testing files and labels before training. It complements a sourcing specification: a broad language catalog is useful for discovery, while a small reproducible intake test reveals whether a particular delivery fits your pipeline. Primary decision Can this source be ingested and scored correctly? Keep distinct Language, locale, script and speaker metadata Text versions Source transcription and documented normalization Evaluation Language-specific counts and errors Test the representation before comparing corpus totals A buyer may receive audio labeled “Spanish” or “Arabic” without enough information to interpret the exact text targets. The immediate engineering questions are whether the labels match the recordings, whether the writing convention is explicit and whether the data loader preserves the intended characters. A corpus can have many hours and still fail this first intake. Take a deliberately varied sample before bulk processing: different files, source families, speakers, language labels and known difficult text patterns. Use the same small fixture through decoding, normalization, tokenization, training-data preparation and scoring. Record where the representation changes instead of assuming every stage treats the text identically. Inspect a public card as a schema example, not a supplier guarantee [Google’s FLEURS dataset card](https://huggingface.co/datasets/google/fleurs) exposes language metadata and both raw and normalized transcriptions. These fields demonstrate useful distinctions to preserve when inspecting another source. They do not establish a universal schema or guarantee that another collection has the same coverage. A language identifier, a recording location and a speaker’s self-described language background are different fields. The archive may support some and not others. Preserve unknown values and document how a label was assigned. Do not fill a locale field by guessing from an accent or treat a language code as proof of regional representativeness. A small normalization test with large consequences The following authored test fixture uses hypothetical source text; it is not a transcript from an available dataset. Its purpose is to surface choices that can otherwise remain hidden until evaluation. Test string or event Decision to document What can go wrong “12/03/2026” spoken as a date Locale, intended spoken form and reference representation A text-normalizer imposes the wrong day/month interpretation A product name embedded in another language Preserve the name, mark language change if supported, define scoring A normalizer translates or deletes the term the product must recognize A word with a meaningful diacritic Character preservation, Unicode normalization and tokenization Different text encodings create apparent recognition errors Two valid written variants Allowed variants and adjudication policy One acceptable output is counted as wrong without a stated convention An incomplete self-correction Verbatim treatment and uncertainty marking A cleaned target erases words that remain audible Interpret the fixture as a contract between pipeline stages If one preparation tool writes a number as words while the scorer expects digits, the resulting error may be a convention mismatch rather than a recognition failure. Decide which representation belongs in the source transcript, model target and scoring view. Keep the original available so the transform is reversible or at least explainable. The [Transformers ASR task guide](https://huggingface.co/docs/transformers/en/tasks/asr) illustrates model-specific preparation, including a particular vocabulary convention. Follow the selected processor’s requirements rather than copying a case transformation into every language. What is harmless for one tokenizer can damage another script or task. Keep audio input choices separate from language claims Inspect the actual source sample rate, channel layout and encoding. Prepare the model input as required while retaining the source identity and timing relationship. Resampling should not silently rewrite metadata to claim that the source was recorded at the derived rate. For code-switching, decide whether the example has one dominant language label, span-level labels, multiple labels or an explicitly unknown status. Those choices support different tasks. A mixed utterance should not disappear from evaluation merely because the loader assumes exactly one language code per file. An illustrative intake record The record below is an authored schema example. It keeps what the source says separate from what the preparation pipeline does. Its locale is intentionally unknown; inventing a plausible region would make the record less trustworthy. { "segment_id": "sample-017", "source_language": "es", "source_locale": null, "label_method": "source-provided language label; sample review pending", "raw_text": "La entrega es el 12 de marzo.", "normalized_text": "la entrega es el doce de marzo", "normalizer_version": "project-rule-set-1", "source_sample_rate_hz": 48000, "model_input_sample_rate_hz": 16000, "source_group": "recording-family-04", "split": "development" } Score each language with its own assumptions visible [Whisper’s official documentation](https://github.com/openai/whisper) reports substantial variation in recognition performance by language. A single aggregate score can therefore conceal both source gaps and scoring errors. Report the metric, text convention, example count and relevant source conditions for each tested group. Where word segmentation is ambiguous or inconsistent, define a suitable scoring unit with qualified language review rather than assume space-delimited English-style WER is automatically comparable across scripts. Keep the same evaluation identities through revisions. Do not interpret a small language slice as a precise population estimate. Release an intake report that another engineer can repeat The useful output is not a reassuring language checklist. It is a record of what passed, which transformations were applied and what remains unresolved. A failed test should lead to a correction, an explicit exception or a decision not to use that subset for the intended task. Identify the sampled source groups, language labels and unsupported metadata. Retain raw text, normalization rules and processor/model versions. Record decode failures, transcript mismatches and script-handling defects separately. Define treatment of code-switching, variants, uncertainty and non-speech events. Report scores and sample counts by language and relevant conditions. Agree which unresolved fields or errors block a larger delivery. Is language count a reliable measure of multilingual quality? It measures one aspect of breadth. It does not describe independent speakers, locales, domain vocabulary, recording conditions or label quality. Check the source design and the task-specific sample. Should all transcripts be translated into English? Only if translation is the intended target. Recognition, language identification and translation are different tasks. Preserve the source-language reference when the task requires it. Can the same normalization rule be used for every language? Do not assume so. Test scripts, numbers, abbreviations and accepted variants with the selected model and qualified reviewers. Version the rules so the score can be reproduced. Research method: Fiund research desk; public sources, explicit worked-example labels and project-specific availability review. Editorial contact: hello@fiund.com. Sources FLEURS dataset card: https://huggingface.co/datasets/google/fleurs Transformers automatic speech recognition task guide: https://huggingface.co/docs/transformers/en/tasks/asr Whisper official repository and model documentation: https://github.com/openai/whisper