Help talk:Audio pronunciations
- I found that my recordings had a chopping sound at the start and finish, no matter how I cropped the selection. To solve this I "normalized" the selection, only removing the DC offset, not normalizing the maximum offset. This removed the chopping noise. - TheDaveRoss
- No idea. The Shtooka Recorder only operates in Windows, so Audacity is still necessary for those of us who don't run Windows. --EncycloPetey 19:35, 17 June 2008 (UTC)
What about using a text-to-speech engine to generate pronunciations in bulk? I would assume commercial tools licenses would be incompatible with Wiktionary. Has anyone looked into this? John1deer 15:06, 1 March 2009 (UTC)
- Yes, we've looked into it. There are no engines capable of doing this correctly. The pronunciation of a word varies from region to region, sometimes a little but sometimes quite a lot. An engine would merely tell users how a particular algorithm interprets the pronuncation, not how a person would say it. --EncycloPetey 16:19, 1 March 2009 (UTC)
Language-country-word, right? For Tagalog-Philippines-word, would the resulting file be something like tl-ph-merienda.ogg? --Icqgirl 09:05, 13 May 2009 (UTC)
I cannot hear them!
from sarri.greek 2017.10.26. υπ' όψιν/attn: TWIMC, @Saltmarsh. Here are a naïve's first impressions:
- All wikipedia, wiktionary oggs: i click on that player, and i hear nothing. I download them and play them with my VLC or WindowsMediaPlayer: it cuts off the last syllable in all short mp3s, oggs, etc. Apparently there is something wrong with me not the oggs. But still, I go to external pages and i hear fine. (e.g. cambridgedictionaries)
- i need a confirmation that the voice is of a native reader, (accent is described, ok. But is he native?)
- i need medium speed of utterance (Normal speech may be too fast for a learner/nonspeaker. Slow speed is unnatural)
- i would like female and male voice.
- i would also like audio recordings for the examples!
- i am happy to make recordings, but the procedure is too complicated. (Here is my voice [theta@webtopos]) I hear some last syllables cut off here too
naming audio pronunciation files
(for eng and non-english words)
Request for expansion of 'naming' at paragraph #Procedure.
attn: Beer parlour
At Help:Audio pronunciations#Procedure I read: Call it ll-cc-word.ogg, where ll is the language code (en for English), cc = your country (us, uk, fr, etc.), and the word that you have pronounced.
I also checked commons:Cat:Pronunciation to see how editors name their files.
- language: ll (e.g. en- fr-) All the audio files i see have capital first letter: Ll-.... Xx(languageCode) (e.g. En- Fr-), apparently a Commons practice...
- country: At most non-english languages at commons:Cat:Pron I see the pattern Xx-word.ogg (no country mentioned). Even for english, the pattern ll‑word‑cc would index A...Z the words in a very helpful way. A suggestion for non english languages discussed (in Greek) with @Xoristzatziki is:
- ll - word - ccoptional - style, intonationoptional
- how to write non-latin script words: commons suggested native script, not transliterations of any kind commons/HelpDesk 2018.06.26.)
Could someone please, add a sub-heading with a bit more about naming the Files? Perhaps also include an Upload paragraph with info:
- Upload at Commons: the Upload Wizard has easy steps for your 'own work': 1. tick own work, 2. name the ogg, choose date, write a description: e.g. Pronunciation of the Xxxxxlanguage word xxxxxx (trans.english). Native speaker, female/male, accent.
- EL-ωωωωω.ogg If further transliteration is given, robots cannot add them automatically at pages: EL-ωωωωω-transliteration (or other comment).ogg. See Cat:Greek pronunciation Talk#logbook for naming files etc, and examples at Sarri.greek/contributions@commons for greek audio pronunciation. sarri.greek (talk) 15:13, 24 November 2018 (UTC)