Bilingual French and English site: one voice per language, set once
A site published in both French and English does not need two audio installs. How to set a default voice for French and another for English, and what actually changes on an article that is already live.
Plenty of sites publish the same content in French and English: a company selling into two markets, a public body, a technical blog whose English version reaches a wider audience. As soon as you add audio, one question always comes up: do you need two installs, one per language, and do you set the voices twice? The answer is no. The player is installed once, and a per-language setting decides which voice reads what. Here is how it works, and what it does to an article that is already online.
The principle: one default voice, per language
Every language you have enabled on your site gets its own default voice, chosen independently of the others. You pick the French voice for your French content, and you separately pick the English voice for your English content. Those two choices never bleed into each other: setting the English voice leaves the French one untouched, and the other way round. An article published in a given language goes out with the voice meant for that language, without anyone having to say so article by article.
That is exactly what a bilingual site needs. You do not want an English reader landing on English text read in a French voice, or the reverse. The right voice follows the declared language of the content, automatically.
The key point: French and English are both covered
This is worth stating plainly, because it is not true of every language. The voice catalogue currently activates five languages: French, English, Spanish, German and Italian. A French and English site is therefore fully supported, both ways: the French version gets a native French voice, the English version gets a native English voice. Neither one has to settle for a stand-in.
That is a real difference from some other setups. A Belgian site in French and Dutch, for instance, can today narrate only its French side, because there is no Dutch voice in the catalogue, and we say so without dressing it up in our piece on multilingual Belgian sites. The French and English pair has no such gap: both languages are native, and coverage is complete.
How each article's language is recognised
To send each article to the right voice, the tool needs to know which language it is written in. It works that out from what your site already declares: the translation plugin you use (Polylang or WPML), or, failing that, the general site language. You do not tag each piece of content by hand: if your English version is already filed as English in your translation tool, the audio knows.
That is what lets a single install handle both languages. The audio player is inserted once into the site, and it is the per-language setting that routes the voice. You are not managing two players, you are managing one player and a small table of default voices. On WordPress, all of this runs through the plugin described on the WordPress page, which installs once and honours the declared language of every piece of content.
What it changes on an article that is already live
Take a concrete case. You have an article published in both languages, currently read, in both versions, by the same voice because you never set anything. You open the voice settings, choose an English voice as the default for the English language, and save. From then on, the English version of the article, once its audio is regenerated, is read by the English voice; the French version carries on with the French voice. You touched neither the text, nor the layout, nor the install. You changed one default value, and it applies everywhere that language appears.
The same reasoning holds at a finer grain. The choice of a male or female voice, of a timbre, of a pace, is made language by language, and the criteria we lay out in choosing a voice by content apply to each one separately. Nothing forces the English voice to mirror the French one: they are two distinct decisions, each made once.
What this setting does not claim to do
One voice per language handles the language, not the translation. The tool reads English text with an English voice and French text with a French voice: it translates nothing, it rewrites nothing. You supply both versions of the content, exactly as you already do for the written text. The audio follows your versions, it does not invent them.
That is the same line we hold everywhere, and that our piece on what French quality means for a voice tool spells out: say precisely what the tool does, a native voice per language set a single time, rather than imply some magic that would handle translation too. For a French and English site the promise is simple and true: two languages, two voices, one install.
Give your articles a voice with WeDispatch
This blog is itself voiced by WeDispatch. Curious how it sounds on your content?
Book a demo