Why do you choose a device voice?

Going back to what people were sayin' before about prefs between male & female voices, I really think it depends.
Espeak, eloquence, Dectalk and Keynote Gold all are primarily male, at least by default. I don't really think of them as male or female though. I've heard people say in the past, "I'm working on the pc and he said 'tree view level 1,' so tab til he says 'list view.'" To me it's not a gendered thing. It's just the voice I'm using that's synthetic.
Zoe and Samantha etc I consider to be female. I Guess because most of the fully synthetic/formant voices sound distinctly unnatural and like a computer with few anthropic attributes. The Vocalizer stuff is still based on human speech, so it's more believable, even in compact. I say this as someone who prefers the pure formant stuff like Eloq and Dectalk. Probably because I .... grew up with it. lol. :)

I don't want some AI voice to read to me on my phone, at least not in fast screen reading applications. The voice should be void of distracting features like breath or vocal fry, etc,. Perhaps that's why I like Zoe so much; she's easy to understand at fast rates and has less of an edge than Sam compact.

I might try rolling tgSpeechBox or eSpeak on everything again for a while. Fast, lightweight, works on everything and no external installations.

I tried to like ESpeak. Really, I tried...
I even ran it on my copy of Windows 11 with NVDA for a time, using the New York accented voices.
In the end, ESpeak just gives me a headache. I cannot do long form reading with it. At all.
Even IBM TTS, I'm able to tolerate, at least for coding. 🤯

So just released an app to the App Store that has about 17 different human voices to announce things. I personally think no synthesised voice can ever beat a human voice, but that’s just me

OK, well technically I still do. I really want like a Samantha doll, a Karen doll and a doll of like any other voices that I like. The only issue is as I really wouldn’t know what kind of face to give them or anything. I mean these voices don’t really look like anybody so yeah I don’t know what kind of doll I don’t want them to be in but yeah, I mean I’m picturing like actual like toy plastic dog kind thing and then I have a voiceover Voice inside of it. I have no idea who would make the surface whatever be a thing, but I totally want that to happen. Yeah so I was thinking of something else but of course now I’m not remembering. Oh yeah, well voices can totally be hot. So like somebody that sound hot I don’t even know how to really describe it. But somebody that I consider hot for the iPhone it’s probably like Jamie in like all of his variance and maybe Lee and his variance and guess maybe Ava because Ava sounds like one of those like a meditation voices so she’s just super calm and soothing. I have a hard time using her on a screen reader full-time unfortunately because your voice is just kind of low for me and I don’t really feel like the enhanced premium variance of her can handle pitch changing very well without getting distorted. I’m not talking about the pitch change typing. I’m talking about pitch change in sliders when you actually change the percentage. Although I honestly find this to be true with any of the enhanced in premium voices. They just do not handle pitch changing well at all. I have to keep everything kind of about the default 50% pitch changed. Has anybody else noticed this? I really do feel like the regular variance of the voices handle changing the pitch sliders a lot better. For example, I’ve definitely noticed this with the regular Karen and regular Samantha. Also, does anybody know why superstar is considered an novelty Voice? She totally sounds like a regular Voice as far as I can tell. She’s not one of those singing voices or anything I was just talking to one of my friends a couple weeks ago that we were saying that it would be really cool to hear some of the vocalizer voices be able to sing. No it hasn’t happened yet, but we were saying we still want to hear it.

She was both, well she still and regarding Lynda I have the ringtone of wonder woman. I also saw her when I was sighted and she was wonder. Same for Joan.

This was an old idea. Warning! I think this thing wants to eat your soul.

https://www.youtube.com/watch?v=_bgXH7U2Ja0

That's a talking doll using a cylinder record put out by Edison. It is creepy as hell.

The voiceover Voice is gonna be inside one of those plastic dolls. I know which doll you’re talking about with the pole string. My aunt actually has a chatty Cathy doll. Pretty sure that would’ve been using a human voice of the time. Nothing is really old. I think she said it said 1961 on the box I think I don’t remember exactly. So yeah if you go on YouTube, you can actually see somebody taking apart one of those CSA toys and they actually took the record out of it and stuck on a normal record player and were able to play it. But I don’t really even need those. I’m good with the iOS voices we have now. I seriously hope that those will always be available to us. I definitely don’t want any super kind of human sounding voice reading my phone. I don’t know. I never put the speaking right past 50% anyway. But did you guys ever notice what I was talking about earlier on this thread about how the enhanced and the premium voices can’t really handle the pitch sliders going up any higher than their default to 50%. They get to start getting distorted like right away. Their regular variance don’t seem to have that issue. I don’t know why it is, but it seems to have always been a thing ever since at least I’ve had an iPhone.

Does anyone remember Furbies? They, to this day, scare the hell out of me.

Below is a link to learn how to play with Furbies. The audio isn't great, but it shows what people got up to back in the late 90s. These things are still creepy as hell to me, however.

https://youtu.be/3f63h1-U_Oo

someone wrote an NVDA add-on so the Braille light voice can be used with it. there is another add-on that brings the old Macintalk voices, and Alex to NVDA. it would be cool to play around with those on my Windows 11 virtual machine, using AVM. it works well, to a point. once i got narrator talking, was going to install jaws,BUT NARRATOR WOULDN'T TURN BACK ON, AND UNLIKE INTEL, jaws for Apple Silicon VMs, does not talk in the instalation process for some reason.

I remember the Furby, unfortunately. That thing is scary as hell. He’s a creepiest voice too. I’m honestly surprised that kids didn’t get freaked out by that. The best way I could just grab the Voice of those things is like the alien voices in Toy story. Like I do kind of like that Voice by those little Furby things like the way that they look they’re creepy. And now I don’t mean visually look here I mean like I’ve literally touched the Furbies. They have the creepy looking eyes and those weird looking ears. I mean they were just so weird looking. I wouldn’t mind that voice, though that voice is kind of funny. I would never ever want to hear the Braille voices again. Unless it was the double top ones. Not the Arctic speed chip. Talk voices in windows. I mean I’m just wondering how legal that is? I mean, they are Apple voices. But yeah, that would just be a weird situation for me because I couldn’t really get used to Apple voices coming out of a Windows computer. I mean it was the same thing that kinda happened to me once eloquence became available for the iPhones. I just couldn’t used to getting hearing that voice coming out of my phone because I was so used to being in windows in jaws. No, I’m thinking if I want any kind of doll with the voiceover Voice, I’m going to have to make something like Build A Bear and use their option to record something and I’ll maybe put like a recording of Samantha or Karen or something inside of one of the Build A Bear animals. I could record like the samples of the voices or something so that at least will say something press the button. So yeah that’s probably the easiest thing at this point. I’ve probably just record the actual samples that you get you know of them saying their name because at least it would say something when he pushed the button. I actually do have a Karen and a Samantha stuffed animal from Build A Bear but maybe I will go back to Build A Bear and then see if I can add a recording sound to those. If not, then I’ll make a separate Build A Bear animal and maybe give them like you know different place of Voice name or something and then put a sample in there for those maybe Allison

I would kill to have the Snow Leopard version of Alex for my copy of NVDA. That was, IMHO, the best version of the Alex voice.

Oh, and@OldBear, don't pretend you don't love Furby! 😛

Had one and a vocalize. The sound was good. Better than some voices now. Still think talking Tina was great voice. Picture it, you turn your phone on and talking Tina starts talking. "Hello I am talking Tina and you better be good to me".

I don’t remember that version of Alex. But then again I never had a Mac so I guess it might’ve been on there. The only version of Alex I know was the ones left in the iPhone. Because I didn’t have a Mac van and I also don’t have one right now so I guess what you’re saying is. They really were different versions of Alex‘s voice too? That’s interesting. Can you find that as a demo somewhere? I’m just curious. Because Alex on the phone isn’t that bad but then again I’d really don’t have anything to compare him to. I use that voice every once in a while.

I like the robotic voices a lot for the main reason they don't slow down the device that's why I won't use AI or human voices unless it's my own of course cause I am goated! I wish there was a way I could downgrade my Google TTS on my phone to use an older version of the speech engine before the AI voices came in. But on my MacBook, I'm using the Karen vocalizer compact voice Australian English with speak speaking rate at 85%. On my android I'm using eloquent voice read. I would use eloquence you speak, but it's not good with punctuation and symbols. e speak works decently outside of that. but i just wouldn't use it because of the issues with symbles being red rongly

I would never want to hear my own voice coming out of my phone. That would just freak me out. Yeah no thank you. Hey I mean if it works for people that’s great. I know when it first came out it was crashing voiceover a lot. I have no idea if that’s gotten better now but I know when it first became available in iOS 17 for use with voiceover it was crashing a lot of people thrones. Yeah I’d much prefer to have more synthetic voice for our phone. I can totally see while using the robot voices is a lot more responsive for sure. I haven’t used the regular canned Voice for a while. I think I should start trying to do that again. I can’t even use it in windows because something happened with the vocalizer. Voice is coming to jaws and windows where the regular car Voice trunk gets so many words that she’s barely unusable and freedom scientific is literally doing nothing about it because they don’t make the voices and they’ve known about it for years and they haven’t done anything about it so I’m not even going to bother telling them anymore so the only Karen variant I can use in Jaws as premium high. Like I’m not complaining it’s awesome. But I can start using regular Karen again on my devices. I might do that.

@ singerGirl, basically the versions of Alex have the same voice data but Apple did some tweaking of how it reads. Without going into the tech side of it you could have the same sentences read with the various Alex versions, and it would sound more or less natural. Yay concatenated speech voices. Lol.

I still don't get why so many people like Alex. It's a good voice, but still suffers from the same problems every concatenated voice has. Slurring, random pitching that is not how real people speak, and so on. But I guess that's why there's so many choices. If Apple ever made a HMM-based Alex compact (not just smaller voice data), I'd try it.

The aren’t any computerized voices that speak like real people speak. That’s why they’re called computerized voices. Smiley face lol but yeah, I know exactly what you’re talking about. Alex does not sound human. And the thing really like make it where he breathes is kind of weird in a way because he only does it sometimes. I mean, people don’t really sound like that that frequently either. I mean, you might hear stuff like that once in a while, but not like the way that they do it with they make it look so pronounced. Yeah I don’t love Alex, but I will use the Voice every once in a while. They’ve been tweeting the voices these phones for a while. I mean, there’s definitely differences between like the iOS 13 voices that I started with an iPhone to the current ones we have an iOS 26.7.1 but I’m currently running. So even the vocalizer voices are being tweaked for years. Right now I’m trying out regular care and again in my SE phone. She sounds OK but I kind of realized why I’ve been going more towards the enhanced premium voices. Very a bit more expressive. The regular Karen voice doesn’t seem to have very much expression. But she’s still cool because she’s Karen and I’ll probably keep her here for a little bit longer. I mean, she’s preinstalled like I can delete her or anything which is good.

Because @Singer Girl wanted to hear a sample, and because I am a twisted mista.
I present, "Hot and Cold," performed by Alex:

https://youtu.be/UxD-laySIk4

Oh my gosh, Brian, this is hilarious. Yeah it’s like the best sample you could’ve possibly found. I’m just saying I totally prefer this version of the song over the original. That’s great. Like you know, they really be getting the technology working once they can get Alex actually sing. I really don’t hear that much of a difference in this version of Alex to the version of Alex that we currently have. I mean, maybe the inflection is a little bit more even but I really can’t tell. It does seem to be a little bit clearer though. Because now there’s almost like this kind of whistling thing going on with Alex that I don’t think was around before so yeah there’s definitely slight differences. But oh my gosh, that sample is just perfect. I was literally like dying laughing when I heard that. I love that somebody actually figured out how to put Alex speaking the lyrics to the rhythm of that that’s amazing. Some freaking awesome sound editing right there. Who wants an album of that. We need like an album of the synthesized voices speaking different songs like that’s perfect. Now I need my two favorite voice to sing me something. I was literally just talking to one of my friends about this couple weeks ago. You were saying how cool it would be for the vocalizer voices to be able to sing. You mean granted it’ll be something like this but it’s still hilarious. I love it.

Now we just need Karen, Samantha, or maybe Zoe to sing us, "One Way, Or Another," by Blondie.
Back me up here, OldBear!

I've never thought of the device or or voice as my friends, they're tools for me. I generally prefer somewhat robotic voices, as a lot of these human sounding voices sound rather one note for me. Take UK Daniel for example. He has a very chipper, happy voice which works very well if you read a message wishing you a happy birthday. But if you're reading the latest news of carnage in some war or a horror novel, then it just sounds weird. Which makes sense, as I believe a lot of these voices were developed more for customer service and use in kiosks, with screen readers as sort of an afterthought; so they're primed more to greet a customer and say thank you for your business, rather than reading Lovecraft's call of Cthulhu. The only one among them I can tollerate is AVA, which is what I use on the iPhone because eSpeak and elequence don't work as well on the iPhone as on the PC. On the other hand some of the robotic voices like Sam, mess up a lot of words and don't sound very clear. That's why I prefer like the ones I mention above, because I can project what every emotion I want to project onto their blank slate.

For me, I like Eloquence because it is predictable in how it says things. If there's an exclamation point or a question mark, Eloquence will inflect accordingly whereas the more natural synths, in my experience, do not. "Good morning!" and "Good morning." are read differently, for example, and punctuation like that can impact how a message is received.

I wasn’t sure if you meant Sam from Microsoft for the more robotic Voice or Samantha from vocalizer. Just curious. Yeah the Microsoft Voice is I don’t really even know where to start with those. I’m not really the biggest fan of any of those except for like Zara and maybe Mark and that’s like it. If you met vocalizer Samantha, well you guys already know what I think about her. Yeah I kind of get what you mean about like wanting to project a certain emotion through robotic voice. Because yeah, they really don’t distinguish between a?! Period. I didn’t actually know that the vocalizer voices were made for kiosks first. That’s interesting info. And yes, we totally need Karen, Samantha, or Zoe to do one way one way or another by Blondie. Definitely. And yes, Katy Perry is totally auto tuned. Glad you picked up on that.

I’m definitely not a fan of auto tune. Artificially corrects pitches. Makes the singer sound like robots. It’s horrible. If you want that just make a synthesizer thing. You don’t need to make a human voice sound like a robot. That’s what they did with the live beauty and the beast 2017 for Emma Watson. It’s completely obvious. I hate listening to that.

It is hard now to listen female voices unless they are not to soft. I used Karen when apple came out with VO. Now is tom.

Did the volume of certain female voices change for you? I was just curious. Because you said you used to use Karen when you first had voiceover and that switch to Tom. So maybe there’s been a reduction in speech volume a little bit? I’m not sure. I haven’t had an iPhone as long as some of the some of you guys are here so not totally sure about that. Tom is a pretty cool voice. The only thing with Tom is he makes me think of calling the Social Security office and he’s also the same Voice they use for Cvs Pharmacy. That’s a store in the US you can get your medicine and things from. They use that voice to welcome people to their Pharmacy so yeah that’s one thing with Tom but yeah, he’s a pretty good voice. I might start using him again. Enhance though.

I can definitely understand that. I were hearing aids now too. I’ve had them for a little bit over a year. So I can definitely understand not liking that too much. Yeah I think we could totally make all of the iPhone voices saying if you put enough autotune on them. I really do sometimes wish the auto tune was never invented though. You really do. It makes like anybody be able to have that kind of a Voice not really be great at singing and it’s just kind of sad I think. If there’s a point that you can never use it to where it’s not like overly done I guess it’s OK but yeah, I’ve heard like a little clips of Katy Perry‘s live voice and it’s completely different than the autotune stuff in the studio albums so yeah it’s definitely been heavily processed. That’s why I said earlier that if you really want that kind of a voice you could just get a computerized voice from one of our phones put that on there and it will still be able to sing. That kinda does need to happen now. I still love that video with Alex though I played it again last night. And I was still like laughing just as much it was perfect. Totally made my afternoon when I first heard it and it totally made my night last night. Thanks again, Brian.