Mac OS X 10.4 Tiger's "VoiceOver" is an embarrassment!!
That is if the demo on Apple's web site is to be belived (click on "play here now" to see/hear a quicktime demo) 
The voice used in the demo is absolutely horrendously bad and technically outdated - it's the exact same voice as used back on the original mac in 1984, with no (or very little) noticeable improvements! That's bad! And it's totally unacceptable in 2005.
Now, Apple may have some surprises up their sleives and actually update the build-in voices before Tiger is released, but if not, they should be ashamed.
(PS: I'm not dissing OS X 10.4 as a whole of course - I'm sure it will be a wonderful OS overall and I have every intention of buying it, but this particular "feature" is not worthy of such an otherwise great OS)
The voice used in the demo is absolutely horrendously bad and technically outdated - it's the exact same voice as used back on the original mac in 1984, with no (or very little) noticeable improvements! That's bad! And it's totally unacceptable in 2005.
Now, Apple may have some surprises up their sleives and actually update the build-in voices before Tiger is released, but if not, they should be ashamed.
(PS: I'm not dissing OS X 10.4 as a whole of course - I'm sure it will be a wonderful OS overall and I have every intention of buying it, but this particular "feature" is not worthy of such an otherwise great OS)
Comments
Originally posted by Power Apple
The voice used in the demo is absolutely horrendously bad and technically outdated - it's the exact same voice as used back on the original mac in 1984, with no (or very little) noticeable improvements!
Um, no. That voice is Bruce, and dates from, at the earliest, System 7, 1991 or so. (Actually, I believe Bruce is one of the Enhanced Voices introduced in OS8.x(5?), along with Victoria, in about '97.)
You can also select the voice you want. Haven't played with spoken text much before I take it? Go to the Speech pane in System Preferences, select the Default Voice tab, and play away.
Originally posted by Kickaha
Um, no. That voice is Bruce, and dates from, at the earliest, System 7, 1991 or so. (Actually, I believe Bruce is one of the Enhanced Voices introduced in OS8.x(5?), along with Victoria, in about '97.)
Allright, so it's from 1991 instead of 1984, does that change my point? Even if it's from 1997 it's still 8 years old technology.
You can also select the voice you want. Haven't played with spoken text much before I take it? Go to the Speech pane in System Preferences, select the Default Voice tab, and play away.
I'm aware of that! But none of the voices I can select are significantly better in any way.
You haven't adressed any of my points though
btw, THIS is the way the voices ought to sound in Mac OS X 10.4 anno 2005, to put things into perspective!
Originally posted by Power Apple
Allright, so it's from 1991 instead of 1984, does that change my point?
It certainly changes your credibility.
Even if it's from 1997 it's still 8 years old technology.
Yes, and the Unix basis at the core of OS X originally dates from 1969.
How new a technology is has zip to do with its quality. If you want to talk quality, that's one thing, but don't bring up specious arguments based on incorrect facts, it doesn't mean anything or support your position. *shrug*
I'm aware of that! But none of the voices I can select are significantly better in any way.
You haven't adressed any of my points though
You didn't have any really to address other than incorrect facts to correct. I mean really, it boiled down to "It sucks!" + wrong info. What did you want me to do, respond with "Does not!"?
btw, THIS is the way the voices ought to sound in Mac OS X 10.4 anno 2005, to put things into perspective!
Those are definitely nice, and it's obvious that ATT (now SBC it looks like) has moved the state of the art a ways. How much for a consumer package to use them? Or maybe for a company to license them?
Originally posted by Kickaha
How new a technology is has zip to do with its quality. If you want to talk quality, that's one thing
I have been talking about the quality all along, what did you think I was talking about? You are right, who cares about of the age of the origin of technology, but the speech tech has improved a lot in resent years. That was my point and still is. If that was not clear enough then I apologise (and for getting one fact wrong)
Those are definitely nice, and it's obvious that ATT (now SBC it looks like) has moved the state of the art a ways. How much for a consumer package to use them? Or maybe for a company to license them?
Apple is advertising "VoiceOver" as one of the major features of Tiger, they should at least be able to improve the quality, perhaps not to match ATT (or SBC) but at least narrow the gap.
More recently, some of the PlainTalk voices got much larger, and better capable of handling nuances in language.
So they've been looking at the problem. I'm sure that if they do come up with a really good solution—in house, or licensed, or purchased—we'll hear about it. But the fallback seems to be an effort to update the PlainTalk voices.
It may very well be that they have improved speech in the wings for Tiger, and haven't unveiled it yet. Or it may be that they may license technologies from other sources. Or Bruce may be what will remain for the time being. Who knows?
I agree that the ATT voices sound very nice, but I'm not sure I'd go so far as to term the current voices an embarrassment. What sort of quality of voice is shipped for free with Windows XP for voice-directed UI navigation? Maybe there's one for a Linux distro? (No, I know that these are silly questions, that's rather the point.)
Originally posted by Amorph
Steve said some years ago that Apple was looking around for voice tech to acquire. That was pretty much the last of that.
More recently, some of the PlainTalk voices got much larger, and better capable of handling nuances in language.
So they've been looking at the problem. I'm sure that if they do come up with a really good solution—in house, or licensed, or purchased—we'll hear about it. But the fallback seems to be an effort to update the PlainTalk voices.
I hope they find a good forward-looking solution soon. If they are thinking of aquiring some speech tech, now may be a good time with all the cash they have in the bank and the high market value...
Anyhow, I see it like. Mac OS X (not least Tiger I pressume) is a very polished and highly integrated system, with so much work put in to even tiny little details. That's part of what is so great about the OS. But then there is the ugly voices, they don't fit in such a modern and polished system.
An anology could be if Mac OS X had all the functionality and stability of today, but with the look of Windows 3.1. Not a pretty thought
I know the quality of the voices doesn't matter for the majority of people (as they will not be using that feature), but for those who will be using it, it does matter. Besides, as I said, Apple advertices it as a major feature.
Originally posted by Kickaha
It may very well be that they have improved speech in the wings for Tiger, and haven't unveiled it yet. Or it may be that they may license technologies from other sources. Or Bruce may be what will remain for the time being. Who knows?
I certainly hope they have improvements in mind. But right now, all I can go from is that they use Bruce for demoing the new VoiceOver tech. I would assume if they had something better they would want to use that in the demo and advertice it as well.... anyway, here's to hoping....
I agree that the ATT voices sound very nice, but I'm not sure I'd go so far as to term the current voices an embarrassment. What sort of quality of voice is shipped for free with Windows XP for voice-directed UI navigation? Maybe there's one for a Linux distro? (No, I know that these are silly questions, that's rather the point.)
For a company (and CEO) as perfectionist as Apple/Jobs, I would call it an embarrassment. I don't know the quality of the Windows XP voices, but I remember reading somewhere that Microsoft was putting some effort into that field (could be wrong, don't shoot me). But regardless, I wish Mac OS X to be better than XP wherever possible
Originally posted by Power Apple
I know the quality of the voices doesn't matter for the majority of people (as they will not be using that feature), but for those who will be using it, it does matter. Besides, as I said, Apple advertices it as a major feature.
Agreed - if I had to rely on it, I would want the best voice I could get. But I also would realize that for a free version of such included with the OS, it's pretty frickin' amazing. If the voice bothered me that much, I'd look into commercial products to enable similar behaviour, but also be looking at spending a considerable chunk of money.
But regardless, I wish Mac OS X to be better than XP wherever possible.
I can't seem to find anywhere that XP has a similar technology. Perhaps you could point me to where it has the same sort of voice feedback included?
Originally posted by Placebo
Victoria in 10.3 is markedly better than Bruce, as reflected by a 100MB voice file.
Maybe so, but still not anywhere near the quality of the ATT voices.
Originally posted by Kickaha
I can't seem to find anywhere that XP has a similar technology. Perhaps you could point me to where it has the same sort of voice feedback included?
Windows XP has something called Text-to-speech. But remember that I'm not talking about the functionality of "VoiceOver", only the quality of the voices.
From the Microsoft site:
Text-to-speech (TTS) is the ability of the operating system to play back printed text as spoken words. An internal driver, called a TTS engine, recognizes the text and, using a synthesized voice chosen from several pre?generated voices, speaks the written text. A TTS engine is installed with the Windows XP operating system. Additional engines are also available through third party manufacturers.
Originally posted by Power Apple
Maybe so, but still not anywhere near the quality of the ATT voices.
Which is why I was asking how much it costs to be able to use those voices, as a consumer. How much do I have to pay to be able to have those on my computer, Mac or PC?
http://support.microsoft.com/default...b;en-us;306902
Can you say 'complicated'? I knew you could.
Their equivalent of VoiceOver is Narrator, and the TTS page above states:
Narrator is designed to work with the Notepad, WordPad, Control Panel programs, Microsoft Internet Explorer, the Windows desktop, and Windows Setup. Narrator may not read words aloud correctly in other programs.
Well that's certainly limited. VoiceOver works with any application.
I can't find example of the voices however. I suppose they could be kick ass, but... how sure are you given the above that XP's voices meet the quality of the ATT voices, or are you just grabbing the state of the art, and expecting Apple to meet that, and ship it for free?
If I were vision-impaired, I'd be much more concerned about the functionality of VoiceOver vs. Narrator than the quality of the voices above and beyond the 'clearly usable' point, but that's just me. I find Bruce to be 'clearly usable'.
Originally posted by Kickaha
Which is why I was asking how much it costs to be able to use those voices, as a consumer. How much do I have to pay to be able to have those on my computer, Mac or PC?
The price of liscensing the voices starts at 7.5$ per copy (at 200 copies), but with rebates for volume. I would assume Apple could get it for around 1-2$ per copy of Mac OS X (just a guess).
See this page for pricing.
I can't find example of the voices however. I suppose they could be kick ass, but... how sure are you given the above that XP's voices meet the quality of the ATT voices, or are you just grabbing the state of the art, and expecting Apple to meet that, and ship it for free?
The build-in voice in XP does not meet the quality of the ATT voices. It's not better than the build-in voices in OS X (just tested it in VPC).
I'm not even saying Apple should meet the quality of ATT, just narrow the gap and surpass XP
The speech synthesis Apple uses does suck, hardly better than the original Macintalk. I wonder how that ATT technology differs from Plaintalk. It does seem a lot better, but it doesn't appear to use all that much hard drive space.
Apple has pre-recorded "diphones" which are not just phonemes (all the possible sounds in a language - about 50 in English), but combinations of phonemes. There are a couple hundred I believe.
You can get better quality by using larger samples of recorded human speech, like full words, but obviously that becomes impossible for personal computers, though it might be feasible for a large company with big mainframes spitting out information over the phone or something.
Originally posted by Power Apple
The price of liscensing the voices starts at 7.5$ per copy (at 200 copies), but with rebates for volume. I would assume Apple could get it for around 1-2$ per copy of Mac OS X (just a guess).
Thanks... now, would you be willing to pay an additional 1-2% on price for something that you probably would never use? The vast majority of consumers won't. It'd be something to increase the price without adding a lot of value for most users. That way leads to Featuritis, and an overpriced product.
The build-in voice in XP does not meet the quality of the ATT voices. It's not better than the build-in voices in OS X (just tested it in VPC).
I'm not even saying Apple should meet the quality of ATT, just narrow the gap and surpass XP
Seems to me that even with equivalent voice quality, VoiceOver far surpasses Narrator and TTS in usability.
Originally posted by BRussell
Wherever there is injustice against Apple, Kickha swoops down to save the day!
The speech synthesis Apple uses does suck, hardly better than the original Macintalk. I wonder how that ATT technology differs from Plaintalk. It does seem a lot better, but it doesn't appear to use all that much hard drive space.
Apple has pre-recorded "diphones" which are not just phonemes (all the possible sounds in a language - about 50 in English), but combinations of phonemes. There are a couple hundred I believe.
Actually, they aren't recorded, traditionally, but synthesized. The phoneme approach is correct however - writing a dictionary can be a royal pain.
You can get better quality by using larger samples of recorded human speech, like full words, but obviously that becomes impossible for personal computers, though it might be feasible for a large company with big mainframes spitting out information over the phone or something.
Yup, take a look at the page Power Apple provided, and you'll see that there are two NaturalVoices SDKs, one for Desktops and one for Telephony. I don't believe that they're actually recorded though.
The current research I know of uses wavelet compression and blending for phoneme chaining, but I can't claim to be an expert.
Originally posted by Kickaha
Thanks... now, would you be willing to pay an additional 1-2% on price for something that you probably would never use? The vast majority of consumers won't. It'd be something to increase the price without adding a lot of value for most users. That way leads to Featuritis, and an overpriced product.
You know, actually I would. This would be an extremely cool technology and sometimes (not always of course) I would even use it myself, freeing me from sitting in front of the monitor when reading some news-articles, letters etc.
Besides it would be perfect when demonstrating OS X compared to XP - could possoibly make some PC user's jaws drop
Seems to me that even with equivalent voice quality, VoiceOver far surpasses Narrator and TTS in usability.
I have no doubt it is "usable" and probably much more so than XP. but the sound-quality is still lacking (even though it's not worse than XP). IMO good enough is not good enough for somethings as refined as "Tiger"...