I don’t think using SAM as an outside model within the speech module was being suggested - i think @samu was just mentioning it as an alternative he liked.
The question i made the thread to ask was whether the speech module, using Apple’s built in voices, could be safely/legally used in commercial projects?
"Apple’s built-in text-to-speech (“System Voices” used via AVSpeechSynthesizer/NSSpeechSynthesizer) are licensed for personal, non-commercial use only. Apple’s macOS Software License Agreement explicitly forbids recording, publishing, or redistributing System Voices “in a … commercial context,” which covers using them in a commercial soundtrack. "
Didn't think about that. Will be included in module's help. Sorry.
But... when you process it by vocoders, shifters, etc... the issue probably vanishes
I have no idea about the legal part, but I wouldn’t worry too much. It would be like speak’n’spell suing daft punk for sampling its output?
and in any case, there’s really no reason to care about copyright unless you make a major hit. They ain’t gonna sue unless there’s big money involved, and by then you would be benefiting from exposure, even if you had to pay back apple for any cent you made
regarding loading other models, I didn’t mean SAM, I understood that, I meant only apple can include new models, if they choose so, there’s nothing a developer can do in that respect
but take my opinion with a grain of salt, this is just my common sense reasoning.
Thanks for looking into it and clarifying @giku, appreciate it. Shame, but Apple being Apple, not a surprise!
Yes, I did also wonder whether using it with the vocoder and/or heavily processing it so it’s not the raw sounds being used would change things if it were prohibited.
Maybe there are free open source speech engines which could be implemented behind the scenes instead of Apple’s?
Yes, I absolutely take your point regarding the unlikelihood of actually getting sued by Apple unless you have a major hit. I just wondered what the situation was and, now we do, I think it’s useful to know. Cheers.
Let us know when you get a cease and desist letter from apple. I love a happy ending 😉
edit: just kidding, I guess what I mean is don’t let your creativity be hindered by fear that big corporate will come after you. Wait, let me rephrase that: %#$* them
Comments
Interesting question and I have no clue other than that it may be a bit in the grey zone for other things than personal amusement...
...I'm too old-school and prefer the sound of the classic SAM when needed :)
(You can generate a phrase and download it as 8-but 22khz *.wav files).
Yeah, I always forget about SAM; amazing tool.
Since it uses apple’s API I very much doubt you can use outside models. I don’t think even @giku has any control over that
I don’t think using SAM as an outside model within the speech module was being suggested - i think @samu was just mentioning it as an alternative he liked.
The question i made the thread to ask was whether the speech module, using Apple’s built in voices, could be safely/legally used in commercial projects?
Nobody know the answer? @giku?
"Apple’s built-in text-to-speech (“System Voices” used via AVSpeechSynthesizer/NSSpeechSynthesizer) are licensed for personal, non-commercial use only. Apple’s macOS Software License Agreement explicitly forbids recording, publishing, or redistributing System Voices “in a … commercial context,” which covers using them in a commercial soundtrack. "
Didn't think about that. Will be included in module's help. Sorry.
But... when you process it by vocoders, shifters, etc... the issue probably vanishes
I have no idea about the legal part, but I wouldn’t worry too much. It would be like speak’n’spell suing daft punk for sampling its output?
and in any case, there’s really no reason to care about copyright unless you make a major hit. They ain’t gonna sue unless there’s big money involved, and by then you would be benefiting from exposure, even if you had to pay back apple for any cent you made
regarding loading other models, I didn’t mean SAM, I understood that, I meant only apple can include new models, if they choose so, there’s nothing a developer can do in that respect
but take my opinion with a grain of salt, this is just my common sense reasoning.
Thanks for looking into it and clarifying @giku, appreciate it. Shame, but Apple being Apple, not a surprise!
Yes, I did also wonder whether using it with the vocoder and/or heavily processing it so it’s not the raw sounds being used would change things if it were prohibited.
Maybe there are free open source speech engines which could be implemented behind the scenes instead of Apple’s?
Yes, I absolutely take your point regarding the unlikelihood of actually getting sued by Apple unless you have a major hit. I just wondered what the situation was and, now we do, I think it’s useful to know. Cheers.
Let us know when you get a cease and desist letter from apple. I love a happy ending 😉
edit: just kidding, I guess what I mean is don’t let your creativity be hindered by fear that big corporate will come after you. Wait, let me rephrase that: %#$* them
Aphex Twin was using them in the 90s so I think it’s fine