Premium voices in iOS 5: What went wrong?

Member of the AppleVis Blog Team

When I heard the news of the release of iOS 5, I didn't own an iDevice. However, I was thrilled to know that it'll provide Premium voices as well as the previous ones which are now called Compact voices. So I was looking forward to trying them as soon as I got my iPhone 4S. I used to think that my 4S came with the Premium voices though I didn't have the "Use compact voice" option under Settings->General->Accessibility->VoiceOver -- I was wrong. Honestly, I was mistaken partly because I was pleased with the way the Compact voices were functioning.

At any rate, as I was updating my iOS from V5.1 to V5.1.1, my 4S also downloaded the Premium voice for American English. Again, I can't tell why that didn't happen when I updated from V5.01 to V5.1. Now the point is that I don't like the Premium American English voice at all, but before moving any further let's get this straight: when I refer to Premium voices I primarily mean the American English Female voice, so the experience might be different with other Premium voices. To make sure you have the Premium voice, go to Settings->General->Accessibility->Voiceover, flick right to see if you have the "Use compact voice" option. If you have it and it's set to "Off," you're using the Premium voice for your language. Here it's located to the right of the "Use pitch change" option.

The benefits of Premium voices

As far as the over-all audio clarity is concerned, the Premium voices use a higher bit rate. This is especially evident on calls where it's easier to hear the audio though the bit rate tends to be reduced on calls regardless of the selected voice. Also, the Premium voices make it easier for me to distinguish between "m" and "n" both in typing and in moving around character by character. With the Compact voices sometimes "m" and "n" are announced so similarly that even my supposedly sensitive ears fail to tell which one is which. And, last but not least, the Premium voices apparently have slightly better pronunciations. For instance, the Premium American English voice doesn't pronounce "purchases" as "per cases." That's pretty much all about it.

The Premium voices need an overhaul

The first thing I noticed with the Premium American English voice is that it tends to crackle frequently. To check this in action, just notice how it pronounces lowercase letters "f," "e" and "r" when typing. The Premium voices have been prepared using a volume setting higher than the Compact voices, and that's mostly to blame for the crackle. You'll notice this if you compare it with Compact voices. When using the Premium voice, the equivalent Compact voice is also used to read deleted characters and heading names. So if these voices are to stay, their volume should be decreased a bit.

Second -- and this is my major gripe, the Premium American English voice has serious intonation problems. When reading phrases and clauses its pitch gets badly altered for some words in the phrase, generating very odd pitch problems which make comprehending speech all the more difficult. To check this yourself, make sure you're using the Premium voice, go to Settings->General->Accessibility->VoiceOver and flick right until you hear the following: "VoiceOver speaks items on the screen." Notice how it reads the phrase "speaks items," strangely emphasizing the word "speaks." Flicking further right will move you to the option "Use pitch change" which is again badly read. Also notice how it reads the word, "Off," when an option has an On/Off property. Interestingly the Compact American English voice doesn't suffer from these shortcomings and closely mimics the way English is actually spoken. As such, I can't comfortably use the Premium voice in iBooks, Read2Go, the Email app and the Messaging app.

Third, though the Premium voice has a higher bit rate, it seems to my ears that it unnaturally puts together vowels, consonants and syllables to form words. In short, it lacks the oomph which is the essence of a good TTS engine -- something which doesn't affect the Compact American English voice at all. In general, I have a great habit of coming to terms with new TTS engines and that's what I did when moving from Eloquence to Vocalizer on my 4S, but the updated Vocalizer which comes with iOS 5 and forms the Premium voices is a different story.

Where do we go from here?

As you might imagine, after three days of grappling with the newly downloaded Premium voice I reverted to the Compact voice. It speaks much more naturally, has a more realistic volume, doesn't crackle, and provides a pleasant intonation. I hope Apple takes a closer look at TTS options in iOS 6. Can the Premium voices be improved? If so, it should be done. Also, now that Nuance has purchased Loquendo and SVox, can Apple get better voices from them other than their Vocalizer? That's definitely worth investigating. At the very least, my sincere hope is that Compact voices aren't removed from iOS 6! Also as we're at it, I'd like to see a pronunciation facility in iOS 6.

What do you think? Do you agree with my assessment? Also, do you -- users of British and non-English Premium voices -- have the same opinion?



More Like This