So you could sort of "turn table" these tracks a little bit by just controlling the rate with your hand on the spinning arrow. You could make them go slow or in reverse. That's where the real good demonic stuff's at....
I’m blind and I’m using voiceover on my phone right now. Or I should say, I’m using dictation at this moment, and then I will use voice over after I’m done. This is fascinating! I’ve been using synthesized speech in conjunction with screen readers since I was about five years old. A lot has changed.
What came before them were even worse. You heard a bit of the Voder there at the end, but at that time there were a few other versions that were pure nightmare fuel.
the sam synthesizer sounds almost exactly like that one horror game called FAITH, I am also just realizing that the game tried to mimic what commodore 64 games looked like
When I was a kid Bonzi Buddy fascinated me for the same reason. You could make him say anything and that alone provided hours of entertainment to my young mind.
Oh, this is great! My dad used to have a speech synthesiser on the computer when my brother and I were little, and would like to say to my brother, “The computer wants to speak to you” before having it say stuff like “Clean your room, Boogerboy” (that was one of my brother’s least favourite nicknames).
I struggle with reading huge blocks of text, and that Natural Readers program you used near the end actually made my life so much easier in school, as I'd use it to listen to assignments and readings for school as well as look at them. It's crazy to see how far speech synthesis has come.
All your videos are alike - I don't care about the title but I click it because it brings back some memory. Then I watch the entire thing, impressed by the video. Good work.
They'll never do it in time! The code the code, figure out the code!! Uhh... what? I gotta diffuse this bomb! It won't be easy! Replace this fourth, this first, this third, this second. OSHIT OSHIT!!! OH GAAAAAAAAH!!! **KABOOOOM**
Mike's Mac Shack Brilliant acting by Techmoan there... Slight tease in his last video too at 8-Bit Guy’s channel. All in good fun... ruclips.net/video/EKl6-fpW7_o/видео.html
Russia can’t even afford to feed its troops; they honestly eat ONE MEAL every 3 days if they are •in• combat and only twice a week when they are not in combat! Yes, Russian military members eat only two MEALS every seven days when they are not fighting! Not two as in breakfast, lunch and dinner for two days but only two servings of food COMBINED! Those two Russian “meals” are one cracker and 64 grams (half cup) of unflavored oatmeal! That is a lightweight breakfast snack in NATO but for Russia that is their entire meal that is supposed to last them three entire days! THREE WHOLE DAYS! I am not joking nor exaggerating! Russia puts on a show on the world stage but in reality, they are a washed up ex-superpower on it’s knees about to collapse, AGAIN! They were NEVER a super power country to begin with; they have been living in the shadows of the United States since before the age of timeso technically they can’t be washed up because they were never there to begin with and the good people of America need to expose these fake Russians for whom they are. America should invade Russia, seize their oil and timber industries and use Russia as their garbage dump.
@@legalizerapingrussianbroad8299 Epic trolling man, well done. I truly believe you are being honest especially because you created your channel 2 weeks ago.
I found _Nostalgia Nerd_ first, then _Technology Connections_ before finding _Techmoan._ That chain eventually led me to *The 8-Bit Guy* with stops at _The Gaming Historian_ from time to time.
5 лет назад+190
Kraftwerk got me into using SAM for my 80s electro tracks.
@@simonzinc-trumpetharris852 What is S&S ? Sorry if it is in the video, I watched it some days ago but still have the tab open but too lazy to watch it again
Back in 1984, I built my own speech synthesizer based off the SPO256 for my Atari 800. I got the PCB design and code from ANTIC magazine. Etched my own PCB. I got the SPO256 and other parts from Radio Shack.
The talking cars actually came in two possible technologies: Tech similar to the Speak-and-Spell (Electronic), and tech similar to the See-and-Say (Analog)
The phonograph mechanism from the see-n-say, if I recall correctly, originated in the Chatty Cathy doll. It got re-used over and over in a massive number of products. In Chatty Cathy, the entry groove selection was somewhat randomized by the string pull / spring mechanism. See-n-say was a refinement to that & as you showed, it made it all the way to automotive tech.
Yep. Used to put in nonsensical words that it would try and pronounce, as long as you placed in a vowel, otherwise it would recite the individual letters! And three full stops gave you "and so on". So ........................... would give you many "and so on's."
It's a reference to some RUclips-Poops that TerriblePerson made: ruclips.net/channel/UCMgZV74ir2PfQChBH_gkWjQvideos (and spin-offs that others made ruclips.net/user/results?search_query=8-bit+guy+techmoan ).
A while back, I played around a bit, with a Japanese voice synth; (Vocaloid) however, as the Japanese language is mostly spoken in syllables, the synth's system is based on syllables, rather than vowels and consonants.
vocaloid is a great software for many reasons, including accessibility. for japanese beginners or people who get a grasp of hiragana or romaji/romanized japanese syllables and wanna make a cover or original song in the japanese language (or any language. ive seen people use the japanese voice databases for multilingual songs many times and it is interesting! it isnt just limited to japanese if you can get it to sound right.), inputing the syllables are easy for them due to how vocaloid seperates japanese syllables in a really easy way. i enjoy it too. :3
It's great, but made for singing. Doesn't produce natural sounding speech - maybe, if you tweaked a lot of settings. There are similar programs made for speech specifically.
@@chills_tiny_mom Voiceroid, Voicepeak, & Voicevox come to mind, and there are probably others, too. Voiceroid has some familiar characters' voices avaliable, such as Yuzuki Yukari and Tsurumaki Maki (i.e. I don't recognise any of the other two softwares' characters from anywhere else). And like singing synths can be tweaked to produce some sort of speech, these speech synths can be tweaked to produce some sort of singing (or rapping, probably with fewer tweaks). At least, I've read about this being possible with Voiceroid; I don't know about the other two. Quick edit: Oh and there's also CeVIO, which among other voices has IA & ONE from the Aria On The Planetes project, as well as - apparently - also Yuzuki Yukari and Tsurumaki Maki.
@@5ucur thx for the quick response 😇 damn why didn’t I think of voiceroid I knew what that is 😭 also yes I have tried to tune singing synths on vocaloid to sound like they’re talking but it gets quite hard 😅 and yeah yukari is on cevio
“When we were ten years old” i knew people in high school that would loose their shit from doing that kind of stuff with the google text to speech things lol
Who else hoped for any sort of mention of Miku / Vocaloid?^^ I know that technically isn't speech synthesis but... well singing synthesim and also he of course mostly covered old stuff, but since the end of the video had a look in the current times it certainly would have been a neat addition. Still an interesting video about how it all started!
he actually completely fuxed it. Those old chryslers used a TI chip, and the TI chips used true synthesis - just like any synthesizer they only had a limited number of stored voices, it wasnt recorded sound in the ROMs but the parameters for a synthesizer.
So, that old Chrysler you showed actually uses a Texas Instrument chipset that's very closely related to the Speak n spell. GM used a similar unit as well. That record player one is from the pre-1984 Nissans. The later Nissan ones actually used digital samples to play the messages, so it was an electronic version of that little record. There were also various other 80's cars from Germany, France, and Japan that used different kinds of speech synthesis but I don't have much information on them because I have been unsuccessful in finding them.
He also utterly fuxed everything about the TI system - it actually models a voice box and what is on the ROMs is modelling data, the equivalent of a voice on a normal synth. I am guessing he didnt know the difference between LPC speech synthesis and LPCM wavefile audio. en.wikipedia.org/wiki/Linear_predictive_coding
5 лет назад+1
In France, there's been for exemple an option on some Renault 25s that provided speech for some functions.
@@peshozmiata The Quattro I had was fitted with a small dictaphone size cassette mechanism which qued up to the message, then played it and rewound, different cassettes gave different languages.
Yeah, it didn't show him dialing the whole number, but I wonder if he did actually call Mat in England...or if he has actually talked to him on the phone at any time. Seems like it would cost a pretty penny.
Allophones aren't really the "building blocks of speech" Phonemes are, and allophones are variations of a single phoneme with in a language (not as in dialectal variation, but as in variation based on phonetic context) For example, most speakers of American English say "h" quite differently in the words "huge" and "whole," although they may never notice. This means that the phoneme "h" in American English has (at least) two allophones.
Thank you. Was just about to say this.these phonemes are written as IPA (funny looking letters and such in a dictionary entry, for pronounciation purposes)
I had S.A.M for my Atari 800 when I was a teen in high school. You could give it phonetic codes for further tuning what you wanted to say and even how to say it such as speech inflection.
I indeed derived hours of entertainment from making SAM say all sorts of ridiculous things on my Commodore 64 as a kid. Its phonics rules were atrocious, so the challenge was to figure out a phonetic spelling that would make it say what you wanted it to say correctly.
Making these speech synthesizers speak in Polish was a great fun, especially that I didn't speak English back then, so figuring out how to use letters so they sound like my language felt like coding.
Really appreciate your videos! I grew up in the 80's-90's and didn't really realize my interest in computers in general until I was older. You're like a technology historian :) Looking forward to watching your other videos
hahaha as a mostly blind user, and using every speech synth since the apple II addon box, I miss always typing in a phonetic spelling to make it sound good. Then the Kurzweil personal reader changed all that, with some pretty great human voices for their time!
@@Bacon420 Did you get them fixed? Cataracts aren't really a reason to be blind to the point of needing braille or aids if you're living in a first world country.
@@gregorymalchuk272 My parents in 1975 decided not to get lens transplants, so I just lived with shitty vision forever haha. Due to a few other issues, there was too much scar tissue to ever do it again. I do read braille still, which is rare these days! I had lots of magnification devices through school.
My native language is not English, and I am not into movies or a fan of Potter, so I was actually as mistaken as the synthesizers - totally though her name sounded "Herr-mi-on" all those years. (so that part of the video didn't immediately make sense to me)
Roger Waters’ album Radio Kaos used a.speech synthesiser on a BBC Master 128 for the character of Billy, and there was a speech synth ROM for the original BBC B featuring the voice of Kenneth Kendall, the BBC News reader. Both of these were early to mid ‘80s
David, You actually didn't explain how the speech synthesis works. You did explain how the fakes work. Looking forward for part two where you go deeper with how the allophones sound and how they are joined together!
the thing is he was completely wrong all round - those so called fakes are true speech synths - what they have stoired are synthesis parameters not recordings - a quick google would have told him that
@@johnfrancisdoe1563 and that isnt at all how the TI based stuff works, including the Chrysler cars. They use LPC speech synthesis, which 8bitguy has sadly confused with LPCM audio, they have nothing in common beyond the word linear. The speak and spell uses a TI TMC0280 speech synthesiser which is fed parameters for a basic model of voice box which it then models, it doesnt just play back recorded sound en.wikipedia.org/wiki/Linear_predictive_coding They are very much real synthesizers and what he claimed was like saying a synthesizer isnt a real synthesizer because it uses stored voices and midi. He utterly fuxed the video WRT TI based stuff, and with regards what is and isnt speech synthesis.
@@johnfrancisdoe1563 Also what he claimed with the shitty commodore magic voice was correct - but it is STILL a synthesizer as much as wave table synthesizers and samplers are synthesizers (he certainly wouldn't argue a Fairlight CMI or an EMU on an AWE32 isnt a synthesizer).
I had an Amiga 2500 as a kid and the word processing program had a speech synthesizer to play back what you had typed. It actually provided me with so much entertainment as a kid.
the pronunciation quirks around the 12 minute mark are why if you see quotes from Stephen Hawking that aren't corrected for spelling, some words are intentionally spelled incorrectly so the voice box he had would sound correct saying them.
@15:32 "When we were 10 year old kids...the favorite thing we liked to do with them was to make them say dirty words." Yeah, when we were 10 we liked that...and then we stopped, and haven't liked it since. No way it's our favorite thing to do at 36. No, just when we were 10... right guys?
My brother and I called a radio station once with Sbatso, spelled wrong it was the speech synthesis for the Sound Blaster sound card, and we asked the DJ to play a song. The guy said, "I'll see what I can do man." :) If he only knew what he just talked to. :)
There's a ancient DOS program called _TRAN.EXE_ that can do true speech-synthesis. It might be interesting to disassemble it to figure out how it works; it's only 47KB.
@@KRAFTWERK2K6 The same pack included SBTALKER which was a straight TTS program without the Eliza. And the speech synthesizer it used (judging from the startup message printed by the driver) was First Byte Software's "SmoothTalker" engine.
it also had that Parrot that supposedly reacted to what you said through a microphone. Never got it to react to anything except pressing the Space Bar ("Ouch!").
David, I am a huge fan of your work. I dropped everything the second i was notified of your video. But i dont think the word you want is "allophones". Allophones are not the basic sound unit. Phonemes are the basic sound unit. Allophones are sounds in any given language that natives see as the same sound, but they might be distinct in another language. For example, the p in pin is not the same as the p in spin, even though we would all think it is. the p in spin is said with a puff of air, and apparently that makes a difference in other languages, so we would say they are allophones in english.
That's not really Dave's fault: he was just quoting the manual that came with the device. Still, what we're dealing with here is a bit lower-level than phonology, is actually _formants_.
I just want to say that years ago my grandmother passed away in the University of Michigan's hospital. One small thing that made the awful weekend more bearable for my wife and I was that the elevators spoke to you. "Floor two." The voice was EXACTLY that of Mr. Spell from Toy Story. Dead perfect. Edit: OMIGOSH I'm so so happy you featured the Tandy speech cartridge. Me and my sister played with that thing for HOURS when we were kids!! One thing we did notice though is that after a while the computer would start to overheat, and as a result the speech output would be different letters than what were typed. After a while we started to notice what letters were connected to other letters. It was weird.
"Hello, my name is Dr. SBAITSO. I am here to help you. Please tell me your problems." I loved this thing, mainly to play with different combinations of words and letters to make it pronounce things properly or say goofy things. But the simple "AI" interactions and conversation was pretty fun for that era.
Sorry, but the TI stuff was true speech synthesis. It did NOT play back recordings of words. It played series of speech sound commands. It just doesn’t have a “plain English”-to-speech-sound-commands converter. So the extra ROMs simply added these mappings. (And provided an extra revenue source.) It’s very similar to the VODER. One can think of the mappings of speech sound commands as being “sheet music” for the VODER player.
@@Gabrol No, they where more like stringing together phonemes. You could technically have built a text to speech synthesizer with one, but the technology at the time was primarily restricted by the many ways given letters are pronounced in different words such that the look up tables would have been huge The processing power available was too low and data base would have been too large to be practical in consumer applications. Therefore given words were analyzed to determine the proper sequence of the phonetic parts of a word. and those pieces were strung together in code. The device also could determine on its own how to blend the phonetics together so the words were more fluid and natural. This was something I think they called 'linear predictive encoding' which allowed it to sound more natural than starting with recordings of people pronouncing a bunch of phonemes and then trying to string them together. The TI chip did not depend at all on any prerecorded sounds or words. You could technically have taken the phonetic representation of a word from a dictionary and used that to code a fairly accurate sounding word (though there was a little more to it than simple common phonetic definitions). I actually worked at TI with the people who had developed the Speak-and-Spell. The synthesis was done on a special speech chip, but the control of the system was a simple 4-bit TMS1000 microprocessor.
Finally my search ended! I was wondering for years since I heard “ its my life - dr. Albun” intro. It had that robot sound. Today I came to know its the SAM! Thank you! You always bring something new from the past and make everyone feel younger. 🙏🏻❤️
I'm a bit shocked you left out the TI-99/4a speech synthesizer, it was really good for the early 80's, It also had full text to speech with the Terminal Emulator II cartridge. The TI-99/4a in extended basic had a set list of words. Some games, like Parsec, for example, had different voices that were recorded, but really really well for the time frame. Very good video explaining the progression of synthesized speech.
It was a terrible vid in that everything about the TI was wrong. The same system was used on the 99/4a - it actually models the voice box and what is fed to it (either form the limited vocab on ROM or from a set of phonemes on the 99/4a or the BBC micro or many other computers) are modelling parameters for that virtual voice box, not recordings at all. It is more true synthesis than any other system you could find. It seems like he has mixed up LPC speech synthesis and LPCM audio recording.
I remember finding an old Apple II in the store room at a mainframe manufacturer I was working for in about 1988ish. It had a speech synthesizer and totally blew me away even then. To be honest I don't feel we've come that far since the 80's
Thanks for the walk down memory lane. It brought back memories of being in the Explorer Scouts. The computer oriented version of Boy Scouts hosted by AT&T. The scout leader took us to the lab where they were working on a speech synthesizer. To all our amazement the leader made it say "I am a shit head".
@@danielsjohnson You know, in polish language the sentence "zażółć gęsią jaźń" is used to test. It's not really related to your comment. The sentence simply looks weird.
OK, speech synthesis nerd here, so...brace yourself. Pretty sure there's a vocoder in the "War Games" voice chain. Something about the very high end of it and how it doesn't get sibilance quite right. I think the talking car in the first video (a Chrysler, I believe?) was a Texas Instruments synth, similar to the speak and spell. In fact Plogue posted some videos of them emulating it (or something similar) as part of their research for their plug-in "Chipspeech". However, many other talking cars did use the plastic record mechanism. You're right in that the Speak and Spell, Magic Voice etc work from a predetermined word bank that was sourced from recordings of human voices. However, there was no way to hold raw audio data of intelligible speech of any length with the memory constraints at the time, so the voice recordings were analyzed by a computer program that translated them into a series of "keyframes" that were used to control a speech synthesizer chip. This is called "synthesis by analysis", and the audio encoding for C64 Ghostbusters worked a similar way for a similar reason. This is why there's a garbled, almost auto-tuned pitch "stepping" as the voice sweeps downward. Fun fact, the chip used in the Currah cartridge is extremely similar to the one in the Intellivoice (both in the General Instruments SP0256 family), except its internal ROM has allophones instead of synthesis-by-analysis data. So really, it's doing the same sort of "stringing together" separate recordings, although on a more granular scale. (This is also why Voice 0 speaks more slowly than Voice 1-in reality they're the same data playing at different speeds! In fact the CoCo synth uses the EXACT same chip, playing at yet ANOTHER speed-although it does appear that the text-to-allophone software routine is significantly superior.) I hope I'm not coming off TOO pedantic here, I'm SUPER happy to see you making a video about speech synthesis because it's an incredibly fascinating topic (at least to me). :)
IMO you arent coming off near pedantic enough when he actually plays 8 bit synths and calls himself a synth specialist etc. He is completely arse backward about what speech synthesis is, and those TIs are true speech synthesis because that what they have stored are the parameters for wave generation, not recorded waveforms. What he said is the equivalent of saying a synthesizer isnt a real synthesizer unless it can play every tune ever made from the name, and that if it needs midi or a keyboard it isnt a real synthesizer.
Wow! I love that you gave us all this information seriously I'm fascinated by this. I've been blind all my life so I use text to speech everyday, and just hearing about how all this works is so fascinating it's amazing. Thank you so much for telling us about all this, I love it it's awesome.
6:32 These speaking clocks are a bit past my time, but I still get nostalgia from them for a different reason. At the beginning of Half Life 1, there's an automated voice in the intro train ride that reads the time aloud in a very similar way. I have a lot of fond memories of that game and it's remake, Black Mesa, so I still get a nostalgia hit, just from a different source :p.
The Navy originally devised computationally efficient rules for automatically converting plain English text to a sequence of phonetic commands for a speech synthesizer.
Speech synthesis is the one thing which gives me life. I use it everyday since I'm a blind. It enables me to work. It helps me to entertain myself. It helps me to communicate so in a nutshell it gives me life. I am using speech synthesise in various forms for the last 15 or so years. It is really fascinating to see how it evolved and I'm glad for what I'm having right now
There was also a DOS program called tran.exe that was a true voice synthesizer for the PC speaker, and then there was Dr. SBAITSO (Sound Blaster Acting Intelligent Text to Speech Operator), a mock psychologist program with a text parser and true voice synthesizer that came with our Sound Blaster card.
Dr Sbaitso was really just an audio demo program built ontop of ELIZA. ELIZA was already able to fool and convince people (or rather, let them fool and convince themselves) that it was "real" and intelligent long before Sound Blaster added a voice.
Cool. I have a Speak & Spell knocking around. I also still have the family Amiga 500. Hours was spent using the text to speech program found within Workbench.
Nice video as always. However, it doesn't really explain *how* they actually work. This is more like "how have they evolved over the years" type of video.
Seippolf Yes you can have opinions, but don’t tell me you don’t see where it starts to derail and show him very subtlety insulting Paul midway in his comment.
When I was a teenager (in the 90s), I found a spot in the circuit panel of the Speak 'N Spell, where if you jammed a flathead screwdriver between two of the solder points of two different chips, it would cause most of the keys to spit out garbled speech, and one key to say "relief." I really thought I was on to something there...
There was a practical use for speech synthesis back during this period of computing. Anyone who was vision impaired could not use a monitor with any practical expectations. But the speech synthesis modules could speak to them and tell them what they were missing. A good friend of mine had a series of speech synthesis devices from the C64 on up through the development of the PC. The more modern ones had a headphone jack so that he could do personal business over his computer without the fear of someone looking over his shoulder. Sadly he passed away a few years ago from Cancer but his legacy as a sightless genius lives on.
Replace the car record with the see and say record!
*Opens door*
"The cow says moo!"
shut up and take my money!
That would be an _excellent_ prank!
Omg I want this
Ahh old tech. I remember swapping the records between my sister's laugh and cry doll and my neighbors soldier.
I so wanna see video of this if anyone ever does it, lmfao
8:04 *The rooster says... *demon screeching sounds**
Creeper_Playz says the Sound Garden Fan!
But what does the fox say?
@@ErikvanderKolk ...go home.
So you could sort of "turn table" these tracks a little bit by just controlling the rate with your hand on the spinning arrow. You could make them go slow or in reverse. That's where the real good demonic stuff's at....
The rooster says... [dog abuse]
I’m blind and I’m using voiceover on my phone right now. Or I should say, I’m using dictation at this moment, and then I will use voice over after I’m done. This is fascinating! I’ve been using synthesized speech in conjunction with screen readers since I was about five years old. A lot has changed.
I hope you're doing okay right now
@@Kai_On_Paws_4298 I am. I hope you’re doing fine too.
I think its so awesome you are blind and still using RUclips to listen to things and interact with people!!
@@ChairmanMeow1 I agree, Its incredible how accessible things are nowadays
@@ChairmanMeow1there's multiple blind RUclips content creators like Tommy Edison
Who would win:
-The most advanced voice synthesizer at the time
-Hermione Granger
Actually the synthesizer pronounced it right... in french x)
@@alduin240 and knowing that Hermione is a french name...
hERMY oNE gRAAAANGER.
@AlexH27 that's actually how to say her name the correct way
*haarrrmeeneee grraaannnggggeerrr*
Some of those speech synthesiser voices are nightmare fuel.
Yeah like the CD/EBS/EAS voices for a start...
What came before them were even worse. You heard a bit of the Voder there at the end, but at that time there were a few other versions that were pure nightmare fuel.
Faith used them extremely well.
This.
The voder is the worst
First you get him pressure washing records, now you're prank calling him... Poor Techmoan :-)
That right there, was a YTP in itself, lolz
have you seen the 8bit guy tech moan memes where david is constantly pestering him! check it out then you will know
the sam synthesizer sounds almost exactly like that one horror game called FAITH, I am also just realizing that the game tried to mimic what commodore 64 games looked like
Because it is the exact same sound
Also trollge
@@theheavy6430 no.
@@Deezboyofficial yes
I think it's cause it was heavily inspired by the games of that time period! And I also recognized it from there too. Good game!
Now you can also do singing synthesizer, like the Vocaloid software.
Yeees! I want episode about Vocaloid
#MakeVocaloidGreatAgain
Yesssss I hope he can cover Defoko! She's the default UTAU and is entirely computer generated (no human vocals whatsoever)!
Great to see a video of mikus great grandparents
Or dectalk I love hearing dectalk sing,I've tried to learn how to do it but every time I do it it crashes. I must be typing in the code wrong.
@@manoerinafanchannel3196 ah, darn. I must have heard wrong ;w; Thanks for letting me know!
When I was a kid Bonzi Buddy fascinated me for the same reason. You could make him say anything and that alone provided hours of entertainment to my young mind.
16:24
*eEEh tHerE TecHMOAn! Oy jUst KeLL dOo DeLL U daD yOR YOOtoOb ChanNL eEs toAdL cRaaAp!*
Cracked me up quite hard. xD
Thank you for this time stamp I can’t stop rewatching it
Specially the "craaap" segment 😂😂
LOL
Oh, this is great! My dad used to have a speech synthesiser on the computer when my brother and I were little, and would like to say to my brother, “The computer wants to speak to you” before having it say stuff like “Clean your room, Boogerboy” (that was one of my brother’s least favourite nicknames).
That's awful
So your dad used his computer to bully your brother? And that's great?
that's hilarious hahahah
I struggle with reading huge blocks of text, and that Natural Readers program you used near the end actually made my life so much easier in school, as I'd use it to listen to assignments and readings for school as well as look at them. It's crazy to see how far speech synthesis has come.
All your videos are alike - I don't care about the title but I click it because it brings back some memory. Then I watch the entire thing, impressed by the video. Good work.
You know what's more impressive? Of course 火锅大王 XD
Fan passed by.
@@diggerszhang Tiananmen Square
@@crazilynoobz Hahahah, funny that you'd say it here
@@diggerszhang yeah i just wanna point out he's selling out to the ccp
Were you inspired by the YTPs made of yourself and Techmoan for that prank call segment? Amazing.
When is the next YTP?
Yo it's my man CS188, if you are wondering a few weeks ago he made a livestream making a YTP of the Galaxy Flip keynote
You're everywhere, too, right?
hey
I always read your name in Obama's voice, cs188!
The Intellivision part reminded me of the AVGN review of the synthesizer, "BEEEEEE SEVENTEEEEN BOHMBERRRRRRRR".
"BEE SZEVENTEEN BAWMBER"!
They'll never do it in time! The code the code, figure out the code!!
Uhh... what? I gotta diffuse this bomb!
It won't be easy! Replace this fourth, this first, this third, this second.
OSHIT OSHIT!!! OH GAAAAAAAAH!!! **KABOOOOM**
I was looking for this comment!! hahaha
I would love to see AVGN and David video .
Yeah.
14:30 Where dubstep wubs happen
Oh hi mio
69 likes
*nice*
Look up The Infinity Project I am Feeling Very Weird at 2:53
It has a small bridge of bass "drop" or rather line where it's the low end c63 voicee
@Gabriel Howell hell yea
@ExDeeXD Music Thath ExDeeXD
Great video. Laughed at the techmoan cameo! Funny stuff right there!
Mike's Mac Shack I wonder if it was inspired by the YTP folks have made with content from the channels.
Mike's Mac Shack Brilliant acting by Techmoan there...
Slight tease in his last video too at 8-Bit Guy’s channel. All in good fun...
ruclips.net/video/EKl6-fpW7_o/видео.html
You have selected Microsoft Sam as the computer default voice
I still got the program
I used to have so much fun when I was younger playing with the tempo of speech of Microsoft Sam lol ... Thanks for the nostalgic moment!
Soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy soy
Edit: Dammit Gboard.
That was my favourite part of Windows XP, not gonna lie.
you still got it i windows 10, just go to start> all apps >windows ease of access> narrator
14:15 finally know from where the voice clips in "FAITH" are from
MORTIS
Thats why i looked in the comments
I KNEW I HEARD THAT SOMEWHERE!
Was about to say this.
That's exactly what I thought as well XD
14:30 “come out and see the sun, the surface is now safe and everything is beautiful”
what
Trollge
Russia can’t even afford to feed its troops; they honestly eat ONE MEAL every 3 days if they are •in• combat and only twice a week when they are not in combat! Yes, Russian military members eat only two MEALS every seven days when they are not fighting! Not two as in breakfast, lunch and dinner for two days but only two servings of food COMBINED!
Those two Russian “meals” are one cracker and 64 grams (half cup) of unflavored oatmeal! That is a lightweight breakfast snack in NATO but for Russia that is their entire meal that is supposed to last them three entire days! THREE WHOLE DAYS!
I am not joking nor exaggerating! Russia puts on a show on the world stage but in reality, they are a washed up ex-superpower on it’s knees about to collapse, AGAIN! They were NEVER a super power country to begin with; they have been living in the shadows of the United States since before the age of timeso technically they can’t be washed up because they were never there to begin with and the good people of America need to expose these fake Russians for whom they are. America should invade Russia, seize their oil and timber industries and use Russia as their garbage dump.
@@legalizerapingrussianbroad8299 Epic trolling man, well done. I truly believe you are being honest especially because you created your channel 2 weeks ago.
@@jcke-2143 I thought SCP-001 “When Day Breaks”
Now I want to see someone replace the internal talking record for a 1980’s car with one from a See N Say
yesssss
"Why does your car Moooooooo?" "Fuck, I forgot my keys" LOL
@@bourdonbt "The duck says quack" "Yeah I know the door is open, shut up car"
"You just hit a cow. Mooooo."
"The cat says, meow".... DAMN IT, that's gonna be expensive.
Gotta love the8bitguy Techmoan crossovers.
The YTP was canon all along.
I found _Nostalgia Nerd_ first, then _Technology Connections_ before finding _Techmoan._ That chain eventually led me to *The 8-Bit Guy* with stops at _The Gaming Historian_ from time to time.
Kraftwerk got me into using SAM for my 80s electro tracks.
80's electro #1
Hell yeah!
Я твои слуга, Я твои работник
Computer World uses S&S extensively.
@@simonzinc-trumpetharris852 What is S&S ? Sorry if it is in the video, I watched it some days ago but still have the tab open but too lazy to watch it again
2020: Speech progam pack
1900: Speech progam *PAK*
Woah the founding fathers really enjoyed those *PAKS* .
P R O G A M
@@kayleyisdisturbing8990 the founding fathers were the 1700s so sadly they weren’t able to enjoy the PAKs :( however, Theodore Roosevelt sure did!
Back in 1984, I built my own speech synthesizer based off the SPO256 for my Atari 800. I got the PCB design and code from ANTIC magazine. Etched my own PCB. I got the SPO256 and other parts from Radio Shack.
why doesn't this have any replies? that's really impressive!
O
The talking cars actually came in two possible technologies: Tech similar to the Speak-and-Spell (Electronic), and tech similar to the See-and-Say (Analog)
Yep. And the example that said "Don't forget to take your keys." was actually an electronic one, using the same chip as the Speak and Spell.
Yup! The one in the car was the electronic one used in some Chrysler and GM cars and the analog record was used mostly in Nissans, all in the 1980s.
5:48 I was waiting for it to say "it's in the game".
Same.
it kinda sounds like the first ea sports intro
Nice Aphex profile picture
@@gooodels cheers XD much love, have a nice day now
I'm waiting for the Speak n Spell to say "L. I. M. P. Say it. Discover." xD
The phonograph mechanism from the see-n-say, if I recall correctly, originated in the Chatty Cathy doll. It got re-used over and over in a massive number of products. In Chatty Cathy, the entry groove selection was somewhat randomized by the string pull / spring mechanism. See-n-say was a refinement to that & as you showed, it made it all the way to automotive tech.
Love the Techmoan collab! Are you baiting the YTPs who keep the faux-feud going between you two?
I think so.
I'd believe it. They did both express amusement at the YTP videos.
@@RAMChYLD what can you do people are dumb?
Someone needs to take that feud and turn it into a video game.
Patrick Courreges Waiting for that!
"SAY" on the Amiga provided hours of entertainment in the late 80s.
Asterisk asterisk asterisk yoooooyooyooyooyooyooyooyooyooo
Yes, I was hoping for a mention!
Yup..nothing about that...Sad
Indeed. We laughed our butts of the thing trying to pronounce Finnish words!
Yep. Used to put in nonsensical words that it would try and pronounce, as long as you placed in a vowel, otherwise it would recite the individual letters! And three full stops gave you "and so on". So ........................... would give you many "and so on's."
14:24 Bonzi Buddy's For The Commodore 64!
Expand dong
hi there expand dong
@@jeltesmit7255 why
@@jeltesmit7255 ruclips.net/video/wDylYenIJUQ/видео.html
Hello, Expand Dong
The buildup to real voice synthesizers had me at the edge of my seat
Ah man, you don't have BEEE SEVENTEEEN BOOMMMBERRRRR!!!
But bomb squad is fun tho.
Bee sheventeen baawwwmer...
@@Suralin0 THE CODE THE CODE! FIGURE OUT THE CODE!
I guess I gotta defuse the bomb?! Uh...
@@LaskyLabs If only he had used the AVGN clip
𝙸𝚃 𝚆𝙾𝙽'𝚃 𝙱𝙴 𝙴𝙰𝚂𝚈!
AVGN : BUUEEEE SHUVENTEEN BUMBEEEEER !
I laughed so hard at Techmoan's prank call!
It's a reference to some RUclips-Poops that TerriblePerson made: ruclips.net/channel/UCMgZV74ir2PfQChBH_gkWjQvideos (and spin-offs that others made ruclips.net/user/results?search_query=8-bit+guy+techmoan ).
Typing out a prank call to Techmoan - pretty good. Actually getting him to film a clip for it - perfect!
This battle may be over, but the war has just begun.
@@TheBaldr Noice!
Flippin eck!
16:18 OMG ACTUAL TECHMOAN CAMEO WHILE HE WAS EDITING HIS "other methods of cleaning records" VIDEO!
Techmoan is boss, I don't know why 8-Bit Guy hurts him so.
That was hilarious
I look forward to all of the YTP videos that clip will end up in
ruclips.net/video/alTn6j0D8pI/видео.html
It was actually pretty good!
The cameo is spot on!
A while back, I played around a bit, with a Japanese voice synth; (Vocaloid)
however, as the Japanese language is mostly spoken in syllables,
the synth's system is based on syllables, rather than vowels and consonants.
vocaloid is a great software for many reasons, including accessibility. for japanese beginners or people who get a grasp of hiragana or romaji/romanized japanese syllables and wanna make a cover or original song in the japanese language (or any language. ive seen people use the japanese voice databases for multilingual songs many times and it is interesting! it isnt just limited to japanese if you can get it to sound right.), inputing the syllables are easy for them due to how vocaloid seperates japanese syllables in a really easy way. i enjoy it too. :3
It's great, but made for singing. Doesn't produce natural sounding speech - maybe, if you tweaked a lot of settings. There are similar programs made for speech specifically.
@@5ucurI’m quite interested which programs are made for speech?
@@chills_tiny_mom Voiceroid, Voicepeak, & Voicevox come to mind, and there are probably others, too. Voiceroid has some familiar characters' voices avaliable, such as Yuzuki Yukari and Tsurumaki Maki (i.e. I don't recognise any of the other two softwares' characters from anywhere else).
And like singing synths can be tweaked to produce some sort of speech, these speech synths can be tweaked to produce some sort of singing (or rapping, probably with fewer tweaks). At least, I've read about this being possible with Voiceroid; I don't know about the other two.
Quick edit: Oh and there's also CeVIO, which among other voices has IA & ONE from the Aria On The Planetes project, as well as - apparently - also Yuzuki Yukari and Tsurumaki Maki.
@@5ucur thx for the quick response 😇 damn why didn’t I think of voiceroid I knew what that is 😭 also yes I have tried to tune singing synths on vocaloid to sound like they’re talking but it gets quite hard 😅 and yeah yukari is on cevio
“When we were ten years old” i knew people in high school that would loose their shit from doing that kind of stuff with the google text to speech things lol
I read that at the same time he said that.
The E Games same wtf 🤣🤣
Same,
Kids in my middle school had to be physically stopped from making the terminal on the macs say swear words
I do it all the time at work when I'm bored lmao
About the Intellivision speech synthesis, you forgot "B-17 Bomber" aka "Beee sevuhnteeen baaaaahmurrr" lol
Yo finnaly a cinemassicre reference
Who else hoped for any sort of mention of Miku / Vocaloid?^^
I know that technically isn't speech synthesis but... well singing synthesim and also he of course mostly covered old stuff, but since the end of the video had a look in the current times it certainly would have been a neat addition.
Still an interesting video about how it all started!
he actually completely fuxed it. Those old chryslers used a TI chip, and the TI chips used true synthesis - just like any synthesizer they only had a limited number of stored voices, it wasnt recorded sound in the ROMs but the parameters for a synthesizer.
Phonograph mechanics just blows my mind. It's simple in comparison to today but how innovative we can capture sound physically.
So, that old Chrysler you showed actually uses a Texas Instrument chipset that's very closely related to the Speak n spell. GM used a similar unit as well. That record player one is from the pre-1984 Nissans. The later Nissan ones actually used digital samples to play the messages, so it was an electronic version of that little record. There were also various other 80's cars from Germany, France, and Japan that used different kinds of speech synthesis but I don't have much information on them because I have been unsuccessful in finding them.
There's the digital talking dashboard on the Audi Quattro - /watch?v=-6YV5rxzxXY
And the english version - /watch?v=P88AnM3z_wI
He also utterly fuxed everything about the TI system - it actually models a voice box and what is on the ROMs is modelling data, the equivalent of a voice on a normal synth. I am guessing he didnt know the difference between LPC speech synthesis and LPCM wavefile audio. en.wikipedia.org/wiki/Linear_predictive_coding
In France, there's been for exemple an option on some Renault 25s that provided speech for some functions.
I have a difficult time believing the 1985 Maxima used a digital voice. It didn't sound like and old chain smoker.
@@peshozmiata The Quattro I had was fitted with a small dictaphone size cassette mechanism which qued up to the message, then played it and rewound, different cassettes gave different languages.
Nice Techmoan cameo
@Timothy Honiss 15:55
Nice touch that he starts by dialing 01144 which is the callout code from the US to the UK.
Flippin' idiot :) glares at phone
Yeah, it didn't show him dialing the whole number, but I wonder if he did actually call Mat in England...or if he has actually talked to him on the phone at any time. Seems like it would cost a pretty penny.
Truly enjoyed that.
Allophones aren't really the "building blocks of speech"
Phonemes are, and allophones are variations of a single phoneme with in a language (not as in dialectal variation, but as in variation based on phonetic context)
For example, most speakers of American English say "h" quite differently in the words "huge" and "whole," although they may never notice. This means that the phoneme "h" in American English has (at least) two allophones.
And that kinda also sums up why vocaloid is so hard to get good at.
野龍 english vocaloids have allophones tho
Thank you. Was just about to say this.these phonemes are written as IPA (funny looking letters and such in a dictionary entry, for pronounciation purposes)
I had S.A.M for my Atari 800 when I was a teen in high school.
You could give it phonetic codes for further tuning what you wanted to say and even how to say it such as speech inflection.
I was so hoping for “BEE SHEVENTEEN BOWMBER”
Well at least we got BOM SQUOD
“Flack”
SAME
ruclips.net/video/IJi-zq0rPAg/видео.html
Memories I didn't even know I had 😂😂
I indeed derived hours of entertainment from making SAM say all sorts of ridiculous things on my Commodore 64 as a kid. Its phonics rules were atrocious, so the challenge was to figure out a phonetic spelling that would make it say what you wanted it to say correctly.
And it was even a bigger challenge to let SAM speak in a other language like mine is :-) I remember how i used many tricks.
I got SAM to speak Swedish... but the text to make that happen wasn't readable in any language at all
@@thiesenf i believe that :D
Could it say soi?
Making these speech synthesizers speak in Polish was a great fun, especially that I didn't speak English back then, so figuring out how to use letters so they sound like my language felt like coding.
BEEEEE
SEVENTEEN
BALLMERR
GERSTBERSTERSS
I LOVE THIS COMPANY!!!
mattel electronics presents
BEEEEEEEEEEEE SEVENTEEEEEEN BOAAOAOAOAOAOOAMBER
@@roberte2945 it won't be easy
DUH COAD DAH COUD
FIGGER OT DEH COED
Really appreciate your videos! I grew up in the 80's-90's and didn't really realize my interest in computers in general until I was older. You're like a technology historian :) Looking forward to watching your other videos
Bruh when he booted SAM up it reminded me of those terrifying sounds that would ocasionally play on the original Xbox menu
hahaha as a mostly blind user, and using every speech synth since the apple II addon box, I miss always typing in a phonetic spelling to make it sound good. Then the Kurzweil personal reader changed all that, with some pretty great human voices for their time!
Bacon420 hey
Do you have retinitis pigmentosa?
@@gregorymalchuk272 naah it was just extreme cataracts heh.
@@Bacon420
Did you get them fixed? Cataracts aren't really a reason to be blind to the point of needing braille or aids if you're living in a first world country.
@@gregorymalchuk272 My parents in 1975 decided not to get lens transplants, so I just lived with shitty vision forever haha. Due to a few other issues, there was too much scar tissue to ever do it again. I do read braille still, which is rare these days! I had lots of magnification devices through school.
Im so happy you mentioned the VODER as I'm currently in the process of constructing a replica right now.
Hmm, I just assumed you meant *vocoder*...but after googling "voder" I see what you mean.
@@samandrew8158 or watched the video til the end
Flack... watch out for flack....
I can assure you that I'm 19 and my brother and I still lose our minds making robots curse. It never gets old!
I make Google voice say"Mary had a big fat ass."
Mattel Electronics presents: BEEEE SEAVONTEEN BAWLLMER.
?
@@superdoom1unrevealed231 AVGN reference. he did a video that included the, I believe, intellivoice
@@boa_firebrand intellivision
Mattel Electronics presents: DESTROY ALL HOOMANS
GHERSBERSTERRRRRRR
Poor Hermy-own. She always gets picked on...
Well, original Greek pronunciation is _[ER ME O NEE]._
Adopted into English it became _[HER MY KNEE]._
hermy-one.
Her-my-on-ne
My native language is not English, and I am not into movies or a fan of Potter, so I was actually as mistaken as the synthesizers - totally though her name sounded "Herr-mi-on" all those years. (so that part of the video didn't immediately make sense to me)
@@jwhite5008 It's Herr-mi-on in french.
Roger Waters’ album Radio Kaos used a.speech synthesiser on a BBC Master 128 for the character of Billy, and there was a speech synth ROM for the original BBC B featuring the voice of Kenneth Kendall, the BBC News reader. Both of these were early to mid ‘80s
Superior Software's Speech for the BBC B was true software speech synthesis in 1986
There is something I like with S.A.M’s voice that I can’t stop listening.
MORTIS
You forgot the best one
B 17 BOOOOOOOOMBBEEEERRRRRRRR
attackquack I distinctly remember it sounding like there was an L in there: B 17 BAAAAALLLLLLLLLLMRRRRRRR
howchoo
I'd say more like "Bowmer", however it'll vary from person to person how they interpret the sounds. (think yanny/laurel)
BEEEEE SEVAAAAHHHNNNNTEEEEENNNNN BOOAAAAAAOOOOMMMMMEEEERRRRRRRRRRR
You are aware he doesn’t have it, right?
BEYYYYY SEVENTEEEN BAWWWWMMBER
David, You actually didn't explain how the speech synthesis works. You did explain how the fakes work. Looking forward for part two where you go deeper with how the allophones sound and how they are joined together!
the thing is he was completely wrong all round - those so called fakes are true speech synths - what they have stoired are synthesis parameters not recordings - a quick google would have told him that
Mijc Osis He showed the insides of the fakex. They were sound snippet playback machines with fancy marketing.
@@johnfrancisdoe1563 and that isnt at all how the TI based stuff works, including the Chrysler cars. They use LPC speech synthesis, which 8bitguy has sadly confused with LPCM audio, they have nothing in common beyond the word linear.
The speak and spell uses a TI TMC0280 speech synthesiser which is fed parameters for a basic model of voice box which it then models, it doesnt just play back recorded sound en.wikipedia.org/wiki/Linear_predictive_coding
They are very much real synthesizers and what he claimed was like saying a synthesizer isnt a real synthesizer because it uses stored voices and midi. He utterly fuxed the video WRT TI based stuff, and with regards what is and isnt speech synthesis.
@@johnfrancisdoe1563 Also what he claimed with the shitty commodore magic voice was correct - but it is STILL a synthesizer as much as wave table synthesizers and samplers are synthesizers (he certainly wouldn't argue a Fairlight CMI or an EMU on an AWE32 isnt a synthesizer).
@@johnfrancisdoe1563 oh yeah - and that car thing was form a Nissan, Chrysler DEFINITELY used TI synths
5:48 "EA Sports, It's in the game", I was waiting for it.
Capital E, capital A, capital _SPORTS_
Same
Put it in exactly like this and it says it perfect. E Ay Sports, It's in the game
@@coondogtheman with the way the guy says it it's more like "sin the game"
I had an Amiga 2500 as a kid and the word processing program had a speech synthesizer to play back what you had typed. It actually provided me with so much entertainment as a kid.
I love the melody of the speech synthesizer's prank call: "Hello Techmooooan ... total craaaaap!" It really sounded like it meant it xD
the pronunciation quirks around the 12 minute mark are why if you see quotes from Stephen Hawking that aren't corrected for spelling, some words are intentionally spelled incorrectly so the voice box he had would sound correct saying them.
Really? Cool!
@15:32 "When we were 10 year old kids...the favorite thing we liked to do with them was to make them say dirty words." Yeah, when we were 10 we liked that...and then we stopped, and haven't liked it since. No way it's our favorite thing to do at 36. No, just when we were 10... right guys?
Actually I'm 37
I still make mine say dirty words
I was 12
Today I let Google translator say dirty words (sometimes). I'm 41 years old.
My brother and I called a radio station once with Sbatso, spelled wrong it was the speech synthesis for the Sound Blaster sound card, and we asked the DJ to play a song. The guy said, "I'll see what I can do man." :) If he only knew what he just talked to. :)
Today I learned: everything is the equivalent to a see and say
There's a ancient DOS program called _TRAN.EXE_ that can do true speech-synthesis. It might be interesting to disassemble it to figure out how it works; it's only 47KB.
There was another nice TTS program for DOS called Dr. Sbaitso. However it was more a TTS Chatbot.
@@KRAFTWERK2K6 The same pack included SBTALKER which was a straight TTS program without the Eliza. And the speech synthesizer it used (judging from the startup message printed by the driver) was First Byte Software's "SmoothTalker" engine.
it also had that Parrot that supposedly reacted to what you said through a microphone. Never got it to react to anything except pressing the Space Bar ("Ouch!").
There was also a Text-To-Speech engine called Monologuw, that had a distinct, robotic characteristic, but sounded quite good for the early 90's.
David, I am a huge fan of your work. I dropped everything the second i was notified of your video. But i dont think the word you want is "allophones". Allophones are not the basic sound unit. Phonemes are the basic sound unit. Allophones are sounds in any given language that natives see as the same sound, but they might be distinct in another language. For example, the p in pin is not the same as the p in spin, even though we would all think it is. the p in spin is said with a puff of air, and apparently that makes a difference in other languages, so we would say they are allophones in english.
That's not really Dave's fault: he was just quoting the manual that came with the device. Still, what we're dealing with here is a bit lower-level than phonology, is actually _formants_.
@@talideon cool! Can you elaborate on formants and how they are being used?
As a Canadian, I was very confused since to us, it means someone who speaks neither English nor French.
@@talideon Isn't it phones? Those are the unit of phonetics.
If I say spin and pin, it is with pin you get a puff of air, not spin. It adds an H to the sound.
"Techmoan, do you have 10 pound balls"... "how do you walk..."
Long Distance FTW!
I wonder how much it costs to crank yank Britain from the United States?
I just want to say that years ago my grandmother passed away in the University of Michigan's hospital. One small thing that made the awful weekend more bearable for my wife and I was that the elevators spoke to you. "Floor two." The voice was EXACTLY that of Mr. Spell from Toy Story. Dead perfect.
Edit: OMIGOSH I'm so so happy you featured the Tandy speech cartridge. Me and my sister played with that thing for HOURS when we were kids!! One thing we did notice though is that after a while the computer would start to overheat, and as a result the speech output would be different letters than what were typed. After a while we started to notice what letters were connected to other letters. It was weird.
14:59 when you only get 17 laps on the fitness gram Pace Test
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
What's the game on 15:00? I have been looking for it for years but don't know what system or the name. Thanks
Great video. I remember using Dr. Sbaitso on my Soundblaster back in the early 90's.
I originally scripted in some examples of Dr. Sbaitso in this video, but eventually decided to cut them out due to the size of the video.
One of the greatest voice synthesizer psychiatrists ever :)
"Hello, my name is Dr. SBAITSO. I am here to help you. Please tell me your problems." I loved this thing, mainly to play with different combinations of words and letters to make it pronounce things properly or say goofy things. But the simple "AI" interactions and conversation was pretty fun for that era.
Sorry, but the TI stuff was true speech synthesis. It did NOT play back recordings of words. It played series of speech sound commands. It just doesn’t have a “plain English”-to-speech-sound-commands converter. So the extra ROMs simply added these mappings. (And provided an extra revenue source.)
It’s very similar to the VODER. One can think of the mappings of speech sound commands as being “sheet music” for the VODER player.
sorry, what is the "TI stuff" that you speak of?
@@Gabrol Texas Instruments - things like the Speak and Spell.
aren't the "speech sound commands" pre recorded by a person, like many TTS nowadays?
@@Gabrol No, they where more like stringing together phonemes. You could technically have built a text to speech synthesizer with one, but the technology at the time was primarily restricted by the many ways given letters are pronounced in different words such that the look up tables would have been huge The processing power available was too low and data base would have been too large to be practical in consumer applications. Therefore given words were analyzed to determine the proper sequence of the phonetic parts of a word. and those pieces were strung together in code. The device also could determine on its own how to blend the phonetics together so the words were more fluid and natural. This was something I think they called 'linear predictive encoding' which allowed it to sound more natural than starting with recordings of people pronouncing a bunch of phonemes and then trying to string them together. The TI chip did not depend at all on any prerecorded sounds or words. You could technically have taken the phonetic representation of a word from a dictionary and used that to code a fairly accurate sounding word (though there was a little more to it than simple common phonetic definitions).
I actually worked at TI with the people who had developed the Speak-and-Spell. The synthesis was done on a special speech chip, but the control of the system was a simple 4-bit TMS1000 microprocessor.
@@tsites1 so the phonemes are synthesized, like the voder?
Finally my search ended! I was wondering for years since I heard “ its my life - dr. Albun” intro. It had that robot sound. Today I came to know its the SAM! Thank you! You always bring something new from the past and make everyone feel younger. 🙏🏻❤️
I'm a bit shocked you left out the TI-99/4a speech synthesizer, it was really good for the early 80's, It also had full text to speech with the Terminal Emulator II cartridge.
The TI-99/4a in extended basic had a set list of words. Some games, like Parsec, for example, had different voices that were recorded, but really really well for the time frame.
Very good video explaining the progression of synthesized speech.
It was a terrible vid in that everything about the TI was wrong. The same system was used on the 99/4a - it actually models the voice box and what is fed to it (either form the limited vocab on ROM or from a set of phonemes on the 99/4a or the BBC micro or many other computers) are modelling parameters for that virtual voice box, not recordings at all. It is more true synthesis than any other system you could find.
It seems like he has mixed up LPC speech synthesis and LPCM audio recording.
Doesn’t Sam sound like the creep voices that play if you waited long enough on the original Xbox home screen
Wait what?
It sounds like the voices from Faith, which is supposed to look like it was made for C64.
@@greysonmiller9407 That's exactly what it is. SAM was used for all the voices in Faith.
69 likes
Click on this
ihasabucket.com
My brother had Sam and I made it rap Run DMC lyrics. We too made it cuss.
Look, just about every teenager in those days would type in "F*CK" as their first word into SAM... I know I did... lol
Papi Uuhmelmehahay this is an amazing comment! Thanks for posting!
I remember finding an old Apple II in the store room at a mainframe manufacturer I was working for in about 1988ish. It had a speech synthesizer and totally blew me away even then. To be honest I don't feel we've come that far since the 80's
what speech synthesizer do you use
Glad you addressed the childhood hilarity of rude words spoken on computer synthesisers 🤣
I love old speech synthesizers, the worse they sound the more I like them.
Thanks for the enlightenment, Dave!
Techmoan's call was so cool hahaha... Thank you for great video as always. !!!
13:48
So tat's from where SAM's voice came from! I'm honestly impressed by the fact Chipspeech has decided to revive these old legendary voices!
Yep.
Chipspeech may be small in community, but it’s certainly dedicated to preservation.
Thanks for the walk down memory lane. It brought back memories of being in the Explorer Scouts. The computer oriented version of Boy Scouts hosted by AT&T. The scout leader took us to the lab where they were working on a speech synthesizer. To all our amazement the leader made it say "I am a shit head".
HELLO
DOCTOR
NAME
CONTINUE
YESTERDAY
TOMORROW
This is the correct test to use for a voice synthesizer. ;)
Cheers
Open the pod bay doors HAL.
@@X-Gen-001 I'm sorry, Dave. ;)
Cheers
@@Semeyaza lol 👍
What about "the quick red fox jumped over the lazy brown dogs" since it uses every letter in the alphabet?
@@danielsjohnson You know, in polish language the sentence "zażółć gęsią jaźń" is used to test. It's not really related to your comment. The sentence simply looks weird.
OK, speech synthesis nerd here, so...brace yourself.
Pretty sure there's a vocoder in the "War Games" voice chain. Something about the very high end of it and how it doesn't get sibilance quite right.
I think the talking car in the first video (a Chrysler, I believe?) was a Texas Instruments synth, similar to the speak and spell. In fact Plogue posted some videos of them emulating it (or something similar) as part of their research for their plug-in "Chipspeech". However, many other talking cars did use the plastic record mechanism.
You're right in that the Speak and Spell, Magic Voice etc work from a predetermined word bank that was sourced from recordings of human voices. However, there was no way to hold raw audio data of intelligible speech of any length with the memory constraints at the time, so the voice recordings were analyzed by a computer program that translated them into a series of "keyframes" that were used to control a speech synthesizer chip. This is called "synthesis by analysis", and the audio encoding for C64 Ghostbusters worked a similar way for a similar reason. This is why there's a garbled, almost auto-tuned pitch "stepping" as the voice sweeps downward.
Fun fact, the chip used in the Currah cartridge is extremely similar to the one in the Intellivoice (both in the General Instruments SP0256 family), except its internal ROM has allophones instead of synthesis-by-analysis data. So really, it's doing the same sort of "stringing together" separate recordings, although on a more granular scale. (This is also why Voice 0 speaks more slowly than Voice 1-in reality they're the same data playing at different speeds! In fact the CoCo synth uses the EXACT same chip, playing at yet ANOTHER speed-although it does appear that the text-to-allophone software routine is significantly superior.)
I hope I'm not coming off TOO pedantic here, I'm SUPER happy to see you making a video about speech synthesis because it's an incredibly fascinating topic (at least to me). :)
Yep, the Chrysler system was a Texas Instruments CM63002ANS, I have one sitting four feet behind me ;-)
IMO you arent coming off near pedantic enough when he actually plays 8 bit synths and calls himself a synth specialist etc. He is completely arse backward about what speech synthesis is, and those TIs are true speech synthesis because that what they have stored are the parameters for wave generation, not recorded waveforms. What he said is the equivalent of saying a synthesizer isnt a real synthesizer unless it can play every tune ever made from the name, and that if it needs midi or a keyboard it isnt a real synthesizer.
Wow! I love that you gave us all this information seriously I'm fascinated by this. I've been blind all my life so I use text to speech everyday, and just hearing about how all this works is so fascinating it's amazing. Thank you so much for telling us about all this, I love it it's awesome.
6:32 These speaking clocks are a bit past my time, but I still get nostalgia from them for a different reason. At the beginning of Half Life 1, there's an automated voice in the intro train ride that reads the time aloud in a very similar way. I have a lot of fond memories of that game and it's remake, Black Mesa, so I still get a nostalgia hit, just from a different source :p.
The Navy originally devised computationally efficient rules for automatically converting plain English text to a sequence of phonetic commands for a speech synthesizer.
Nice. Source? I’d like to read more on that
@@TheDanielLivingston Amiga's translator.library also consisted of such rules.
I watched in anticipation, hoping he would show S.A.M. I loved this program.
Microsoft Sam as a kid making him say RRRRRRRRRRR or FFFFFFFFFF and naughty words, ah the simple pleasures.
Speech synthesis is the one thing which gives me life. I use it everyday since I'm a blind. It enables me to work. It helps me to entertain myself. It helps me to communicate so in a nutshell it gives me life. I am using speech synthesise in various forms for the last 15 or so years. It is really fascinating to see how it evolved and I'm glad for what I'm having right now
There was also a DOS program called tran.exe that was a true voice synthesizer for the PC speaker, and then there was Dr. SBAITSO (Sound Blaster Acting Intelligent Text to Speech Operator), a mock psychologist program with a text parser and true voice synthesizer that came with our Sound Blaster card.
I remember Dr SBAITSO, I loved playing around with that. Good old Sound Blaster lol
Dr Sbaitso was really just an audio demo program built ontop of ELIZA.
ELIZA was already able to fool and convince people (or rather, let them fool and convince themselves) that it was "real" and intelligent long before Sound Blaster added a voice.
13:48
*MORTIS*
OUT DEMON
YESSSSSS
I watch jacksepticeye play it
𝘼 𝙜𝙪𝙣 𝙬𝙞𝙩𝙝 𝙤𝙣𝙚 𝙗𝙪𝙡𝙡𝙚𝙩.
I HEARD A DOOR OPEN UPSTAIRS.
OMG i have that vox clock, it has been in constant use since i bought it form Tandy in 1992
That see and say bit was actually super fascinating. The record design is a real piece of ingenuity.
" BEEEEE SEVUNTEEEEN BAWMERRR" -intellivoice
YES haha I was looking for this comment
Beat me to it
METTEL ELEKTRONIKS PREZENTS
BAWMB SKWAD
That’s been stuck in my head for years... takes me right back to the AVGN days! 😂
Came here for this. Wasn't disappointed
I swear that car advised the driver, "Don't forget your jeans".
Which is pretty good advice I guess.
It said “Don’t forget your keys.”
Yes, that's what I heard. After thinking about it I realised it probably meant keys.
Close @jaxxstraw , but what you really heard was, "Don't forget your genes". It was a family car...
Video should be "history of speech synthesizers" :P
Unless I missed the "how they work" part
Agree
You missed it
it is more history and less how it works
Cool.
I have a Speak & Spell knocking around.
I also still have the family Amiga 500.
Hours was spent using the text to speech program found within Workbench.
Nice video as always. However, it doesn't really explain *how* they actually work. This is more like "how have they evolved over the years" type of video.
JellyGamer Sorry man, but you make it too obvious: [Insecurity Intensifies]
Charles Campuz what do you mean?
You He took offense to the OP’s opinion.
@@charlescampuz5812 Not really, he just stated opinion of his own. I hear that is allowed now.
Seippolf Yes you can have opinions, but don’t tell me you don’t see where it starts to derail and show him very subtlety insulting Paul midway in his comment.
When I was a teenager (in the 90s), I found a spot in the circuit panel of the Speak 'N Spell, where if you jammed a flathead screwdriver between two of the solder points of two different chips, it would cause most of the keys to spit out garbled speech, and one key to say "relief." I really thought I was on to something there...
And I thought I was being smart when I tried to touch the audio jack of a speaker to my finger to create a triangle wave.
Pitching Sam's voice up makes him sound like bonzi buddy
There was a practical use for speech synthesis back during this period of computing. Anyone who was vision impaired could not use a monitor with any practical expectations. But the speech synthesis modules could speak to them and tell them what they were missing. A good friend of mine had a series of speech synthesis devices from the C64 on up through the development of the PC. The more modern ones had a headphone jack so that he could do personal business over his computer without the fear of someone looking over his shoulder. Sadly he passed away a few years ago from Cancer but his legacy as a sightless genius lives on.