In our Text to Speech service, credit usage is calculated based on the phonemes and punctuation marks:
| Phoneme Count | Credit Consumption |
|---|---|
| 1 phoneme | 1 credit |
1 credit equals 1 phoneme synthesized.
Each punctuation mark (comma, period, etc.) within or at the end of a sentence consumes 1 credit.
Example 1
Sentence: "Hello, my name is Jessica."
Phoneme breakdown:
hAlo mY nHm iz cEsikA
| Word/Punctuation | Credits | Explanation |
|---|---|---|
| Hello | 4 | 4 phonemes: hAlo |
| , | 1 | 1 punctuation mark |
| my | 2 | 2 phonemes: mY |
| name | 3 | 3 phonemes: nHm |
| is | 2 | 2 phonemes: iz |
| Jessica | 6 | 6 phonemes: cEsikA |
| . | 1 | 1 punctuation mark |
| Total | 19 |
Example 2
Sentence: "Hello, my name is Jessica"
Phoneme breakdown:
hAlo mY nHm iz cEsikA
| Word/Punctuation | Credits | Explanation |
|---|---|---|
| Hello | 4 | 4 phonemes |
| , | 1 | 1 punctuation mark |
| my | 2 | 2 phonemes |
| name | 3 | 3 phonemes |
| is | 2 | 2 phonemes |
| Jessica | 6 | 6 phonemes |
| Total | 18 |
Note: In the second example, since the sentence does not end with a period, the system adds one for a smooth ending, but this added period does not consume a credit.
Language-Specific Notes
Phoneme counts vary by language.
For example:
- In English:
"Hello"= 4 phonemes - In Turkish:
"Merhaba"= 7 phonemes (almost one phoneme per character)
This means that credit consumption may differ depending on the input language, even if the word length appears similar.
By understanding these rules, you can better estimate the credits your text-to-speech tasks will consume.
Version History
| Version | Date | Description |
|---|---|---|
| 1.1 | 25 Sep 2025 | Added phoneme breakdown examples and note on language-specific differences |
| 1.0 | 22 May 2024 | Initial document created |