Voice Speech Clips from Microsoft Sam, Mary & Mike.

You will need a mp3 media player installed on your system to listen to the mp3 clips. Times noted are play time, not the time to load.

Dial-up users will have a delay while the clip loads the larger the clip the longer the delay. Typically dial-up users get 3 to 4 Kb per second, so some delays may be lengthy

Microsoft Sam, Mary and Mike Speech Samples.

Personally I was quite happy with the voices especially Microsoft Mary's, although that may be a english male viewpoint. I accept that at times all voices were distinctly not human when you heard a long session.

The clips read from the list of problems noted on page one are repeated below, so they are pretty much a worst case example of the software. All the voices have the same problems by the way :-

Speech conversion was usually quite realistic, as was the way it handled punctuation.

After quite a few hours of use the few problems I noted were all relatively minor (to me at least): -

          Microsoft Sam, clip = 11 sec, 78 Kb.

  1. Some (relatively few) words are clipped, the word useful comes out as just ful. Two, to and too can also suffer from this.
  2. Sometimes words are given emphasis for no apparent reason.

    Microsoft Mary, clip = 10 sec, 72 Kb.

  3. Numbers used as bullet point marks on web pages were not spoken, try it on this section to see this.
  4. The final point is not necessarily a problem. The salesmen would no doubt call it a feature. The Voice Reader extends some abbreviations so, co becomes company & sep becomes september for example.

    Microsoft Mike, clip = 18 sec, 124 Kb.

June 2005 Update: A non Microsoft voice, CMU Kal (see Download: ByteCool) used with the Natural Voice Reader software didn't extend the abbreviations. Which probably means that its the Microsoft voice extending abbreviations, rather than the Natural Voice software.

Note that the background noise on the clips is a product of the conversion to mp3. The chargeable version produces larger but better quality clips than the ones one this page.

Voice Speed Settings

The software has 9 voice speed setting from slowest -4 to fastest +4. Here the Microsoft Mary voice shows the differences between the voice speed settings in these short examples.

Windows XP Narrator using Microsoft Sam.

If you have Windows XP, Microsoft Sam is already loaded on your machine. A rather basic (and in my opinion not to useful) utility to use it is called the Narrator, it can be accessed through the Utility Manager.

For more information on the narrator see the accessibility page

Software Used to Create the mp3 Clips.

The speech clips on this page were produced using the free version of AT&T's Natural Voice Reader software, together with Audiograbber (see links).

The pro and enterprise versions come with a 'Read to File' option, allowing a mp3 or wav file to be created without extra software.

'Read to File' using the wav file format produced wav files about 10 times larger than mp3's.

Microsoft Sam and Mary Conversations.

This was a slight fiddle as I had to pause the recording midway through to change voices and speed settings.

Sam: 'Mary Mary. Quite contrary how does your garden grow?'
Mary: 'With silver bells and cockle-shells and pretty maids all in a row.'

Clip = 52 Kb - 7 Seconds.

Like some more conversations or other clips? Suggests one! If its interesting or entertaining, I will create a mp3 file and add it here.

Thanks to Paul Simms for this one, from a famous conversion between Winston Churchill and Lady Astor.

Mr. Churchill you are drunk ... Both voices (Mary and Sam) on speed setting 1.
Clip = 115Kb - 16 seconds.

Other MP3 Clips.

  • Mary reads Thomas Hoods 1844 poem November. Clip = 149Kb 21 seconds.

    "No sun, no moon!
    No morn no noon.
    No dawn no dusk. No proper time of day.
    No warmth, no cheerfulness, no healthful ease,
    No comfortable feel in any member.
    No shade, no shine, no butterflies, no bees,
    No fruits, no flowers, no leaves, no birds!


  • Mike reads from Gilbert and Sullivan's The Mikado. Clip = 54Kb 7 seconds, fastest voice setting.

    "To sit in solemn silence in a dull dark dock.
    In a pestilential prison with a life long lock
    Awaiting the sensation of a short sharp shock.
    From a cheap and chippy chopper on a big black block."

Tongue Twister mp3 Clips.

These tongue twister clips make use of Microsoft voices Sam, Mike & Mary. The clips all use the fastest voice speed setting:- Strange really, listening to clips that won't 'twist the tongue' of the electronic voice. Maybe that would make them more real, if they did make an occasional mistake?

External Links.

OnLine Speaking Clock Online Speaking Clock by Bristol Clock.

I used the Voice Reader software with audiograbber (see below) to create the audio clips used by the speaking clock. The clocks use Microsoft Mary, Mike and Sam Voices.

Each use of the clock is about 30Kb to 40Kb so it should be suitable for dial-up users.

Now that the free version no longer outputs directly to mp3 or wav files. To create mp3 (or wav) files from live Microsoft Sam (Mary/Mike) audio, try using Audiograbber's line in sampling, for free software to do this.