Thanks for the overview of your technique! Learned quite a bit. I had never used the phonemes checker, only used the dictionary, working off similar sounding words. Looks like phoneme checker should save time.
I had the phoneme checker as my desktop for a while, but honestly, I found it quicker to think of a word with a similar sound and put it straight into S-V to find the phoneme I wanted, because I never had to leave S-V. It hasn't failed me yet as a technique. But it depends how you work. I guess a printed phoneme checker stuck to the side of the monitor would be even quicker. I think flow is important, and anything that breaks that - even leaving Synth-V (in this case) for a second, cuts the flow and gives the possibility to forget what you're doing. Especially for old people like me. 😏
Another method for “read” to sound correct and probably easier, is another word that only has one sound like: “reed”. It can work in many but not all cases. Gianni❤
You can use the Dictionary feature to determine Phonemes. Create a Dictionary and you can search words that sound like what you want and get the phoneme breakdown. Then you won't have to add notes, and remove notes. You can just use the dictionary on the right hand side menu. If there's a pronunciation you *always* want, you can add the edited word to your dictionary and it'll populate using it as a starting point.
I must admit that I don't use the Dictionary, but I don't see that working in this case, because its two different pronunciations of a single spelling. You can only put a spelling in the dictionary once, so yeah, its fine if its all you ever use, but either could be right in this case depending on context. There are a number of mechanisms that you could use. The one I might otherwise use in this particular case would be to simply mis-spell the word - red vs. reed would work in this case. I was simply using one of the mechanisms for figuring out phonemes. The reason I've never really got into using the dictionary is because it doesn't often mis-pronounce a word. In fact, I find it pretty uncanny how good it is. But I appreciate your viewpoint, and I'd also be interested in knowing if I'm missing something here about having two words spelt the same in the dictionary.
@@themidimachine Hey there, I apologize if I came off aggressive. What I was trying to say is that you do not need to add a word to the dictionary to use the phonemes feature it offers. IE: You could type "reed" into it to see the phonemes and then edit the word "read" into your piano roll. I was simply suggesting that the extra clicks of changing to the pen tool, adding a note, editing the lyrics, and *THEN* Copying the phoneme to the other word and then deleting it, could be removed by just using the "Word" field on the dictionary tab. *I* don't save words, I was simply sharing that you could. I do a lot of metal/hard rock with this application, and there are a lot of syllable broken words. Learning that "Dee/Luge/shone" is needed instead of "delusions/+/+" is tricky :) I'm also using Richy and Hayden as whisper vocals in RAP mode and adding a truck load of distortion to achieve metal vox...and it's working. I wish they'd come out with a hardcore singer, but I know it's not their target audience. Mai helps with sounding like "Baby Metal" :)
@@davidnovin200 Hi David, I didn't get an aggressive vibe. No apology needed. Sorry if I gave that impression in my response. I tend to try and say stuff with as few words as possible, and it can give the wrong impression. :D All good. I get what you're saying now. BTW - did you create your dictionary from scratch and just add words as you come across them, or did you import one from somewhere as a starting point? They tried the hardcore with Nine-Zero. Not a great choice IMO, but hey, I guess it was new ground for them. Eclipsed Sounds looks like they're open to suggestions for ideas for new voices. I saw something, I think on their Twitter account about encouraging people to e-mail them with ideas. Could be an idea. Bet that would be good! I know I'm looking for a Chaka Khan, so that's what I'll be campaigning for, but send them an e-mail, and may the best man win. 😆
@@themidimachine I just created a new one, and saved it. I only have like 5 words in it right now. Mostly I use this for demoing new songs for my singer, and words he pronounces in unique ways get added to my dictionary. Otherwise I just use the word function if I need to find a phoneme. The other thing I like to use it for is to use words that sound like WHAT I want them to. intentional mispronunciations, as it were. :) I love that this tool exists, but as a metal song writer, I don't see it replacing my singer, who sounds like Max Cavalera of Sepultura. No AI is gonna give me that tone :)
This was an interesting little watch! One thing I'd like to mention about English vocals: pretty much every AI voicebank for SynthV can sing in English, no matter the language it was initially recorded for. To make the voicebank sing in English you simply need to change the Language setting (it's almostly at the top of Voice panel), which by default is set to their primary language. Though, it's important to mention that often these voicebanks tend to sing with an accent (how strong it is I can't really say, since English isn't my first language, so if you want to hear a certain voice in English I would recomment searching up covers of English songs using the voice that interests you and just listening to a few to decide if you like the pronunciation)
Yes, accents are an issue. I'm not sure if its the same with all languages, but as English speakers we seem to want to hear singers without strong accents, especially sounding like English isn't their first language. Even Natalie gets criticized from time to time over this, but certainly many consider Weina unusable, and even Sheena is borderline, so many of those that are not specifically intended for English speaking tend to get ignored except by those looking for a very specific sound. I think at this point, we have enough voices to not need to compromise on things like this (even though we still need more new voices!). Its a bit weird that this is the case, and I have no idea why it is.
