Is xml header necessary in ssml request in the SSML Specification?

Question

I wonder if we strictly need to specify xml header in our requests to ssml synthesizers like the following:

<?xml version=\"1.0\"?><speak>hello world</speak>

or does the SSML standard also allow this?:

<speak>hello world</speak>

Answer 1

IN SSML specification there is a section https://www.w3.org/TR/speech-synthesis/#S2.1 that says:

A legal stand-alone Speech Synthesis Markup Language document must have a legal XML Prolog [XML §2.8].

For XML 1.0 both <?xml version=\"1.0\"?><speak>hello world</speak> and <speak>hello world</speak> are well-formed, so it's not strickly needed to specify a xml header.

Answer 2

I use the Azure and Google TTS engine in dot.net Core and for Azure in the line

synthesizer.SpeakSsmlAsync(SSML)

I can send SSML text without the XML header

I think the answer is now

Azure TTS : No Google TTS: No

<speak version='1.0' xmlns='http://www.w3.org/2001/10/synthesis' xml:lang='en-US' xmlns:mstts = 'http://www.w3.org/2001/mstts'> 
        <voice name = 'en-US-BrandonNeural'>
         Hello Stack Overflow
        </voice>
        </speak>

Answer 3

Azure TTS: Yes, xml header required

Google TTS: No