Datasets:

---

ddamianos

/

hparl

Like 0

Hello,

Could you please clarify how the values in the utt_id column are constructed?

For example, the dataset contains identifiers such as:

aikaterinh_papanatsiou-olomeleia_11_02250000_02253000_1_1
aikaterinh_papanatsiou-olomeleia_11_02245000_02248000_2_1

I assume that aikaterinh_papanatsiou is the speaker name and that olomeleia_11 identifies the source recording or parliamentary session. However, I am not certain what the remaining components represent.

In particular:

• Are 02250000 and 02253000 start and end timestamps? If so, what unit and reference point are used?
• What does the number in olomeleia_11 represent?
• What do the final two values, such as _1_1, _2_1, or _2_2, represent?
• Do the timestamps refer to the exact audio sample, or to a larger parent window from which the sample was extracted?

A description of the complete utt_id format would be very helpful.

Thank you for making this dataset available!