FAQ – Frequently Asked Questions

🔹 Do you implement or integrate your suggestions into our TTS system?

No. I don’t write code or follow up on technical implementation. My role is to provide clear, strategic linguistic insight that your development team can apply on their own.

🔹 How can I apply this linguistic research to my TTS or speech tech project?

If you're developing a voice AI, TTS, or speech-based interaction system, my work can help you model not just how speech sounds, but what it's doing — its communicative intent.

Thanks to the prosodic-functional analysis, you'll be able to:

  • Enrich your datasets with utterances labeled by function (e.g., [Greeting], [Thanking], [Requesting]), beyond emotional tone.

  • Model prosody based on measured contours (in semitones, decibels, and duration) for more intentional, natural-sounding speech.

  • Use textual markers as input cues (e.g., “Ciaoooo” for lengthened greetings).

  • Adjust post-processing of audio to match natural intonation patterns.

  • Validate if your TTS output matches human expectations based on actual linguistic use.

This knowledge integrates smoothly into modern TTS pipelines like Tacotron, FastSpeech, VITS, or Bark — either in training, preprocessing, or postprocessing stages.

🔹 What do you need from us to deliver an analysis?

For the basic Strategic Module, I only need three types of utterances that your system struggles with, along with seven examples of each (audio + transcription). That’s all.

🔹 What kind of transcriptions should we send?

Standard orthographic transcription is enough. If you already have phonological transcription, that’s welcome, but not required.

🔹 Do we need to understand phonetics or linguistic theory to work with you?

Not at all. If you’re comfortable with the terminology, I’ll use it. But if not, I’ll explain everything in clear and accessible terms.

🔹 Is your model language-specific?

My analysis is based on a model developed from Spanish spoken in Mexico, but the prosodic patterns and structure types I work with are transferable to other Western languages. I adapt the guidance to the language your system is working with.

🔹 Will other clients receive the same examples?

No. Each project receives a tailored analysis based on your own corpus and specific challenges.

🔹 What’s the difference between the $1,500 and $5,000 packages?

The $1,500 package gives you prosodic analysis of 3 representative examples. The $5,000 package includes that, plus a full corpus classification recommendation to help you identify and fix larger data-structuring issues in your training.

🔹How can I pay for the consulting service?

Payment can be made via:

  • PayPal

  • Direct bank transfer within Europe (IBAN)

  • Wise (for international transfers)

For the Strategic Module Analysis (USD $1500), full payment is required upfront before the project begins.

For the Corpus Diagnosis + Strategic Module (USD $5000) and the Full Prosodic Mapping (USD $12,000), payment is split into two phases:

  • 50% upfront, before project initiation

  • 50% before final delivery of the report and visualizations

If you require an invoice, please indicate this at the time of booking. I will include all necessary tax information as a self-employed consultant registered in Austria.

🔹Do you provide invoices? How are your consulting services billed?

Yes. Every client receives a PDF invoice upon confirmation of the project.

Consulting services are billed under Dr. Lesly Widner, independent consultant registered in Austria (Selbständige Tätigkeit).
All invoices include my Austrian tax number and are issued in accordance with Austrian freelance regulations.