Retaining old text-to-speech voices alongside the new neural voices
Sep 29, 2023
My team is responsible for the creation and maintenance of a rapidly growing suite of over 200 courses. All our courses have an optional voice over as an accessibility feature. Many of these courses have to be updated and modified regularly to reflect changes to legislation and government guidelines. Therefore, some time ago we opted to use Storyline 360's test-to-speech as the quickest and most cost effective way to provide this voice over. For all its idiosyncrasies and occasional limitations, we're big fans.
We are very excited about the new neural voices, early impressions are that they are big step up. The voice we have used most often (United Kingdom (English) - Amy) is one of those that has been given a neural makeover. This, unfortunately, creates an issue for us. In order to make these small changes to our existing courses which use the old Amy voice, it is necessary to revert to an earlier version of SL. I had to do this today in order to add a single word to a single slide (in response to client feedback).
Therefore, would it be possible to retain the older versions of the voices alongside the new neural ones for a period (our global update cycle is 12 months)?
Thank you.
14 Replies
Hi Barney,
Thanks for reaching out!
There isn't a way to roll back the neural voices in the most recent version of Storyline 360. You'll need to downgrade to a previous version if you want to access the older versions of the voices.
Let me know if you have any questions!
This is such an inconvenience. Why wouldn't you just keep the non-neural TTS voices alongside the new ones for a period of time? We are also running into the same issues stated by OP, so this has become a huge pain.
Same problem here with Germany - Vicky. It is nearly impossible and very expensive to update to the new voice of Vicky. I prefer the solution of Art Fanega to provide both variants of voices. Is there any possibility for this solution in the near future?
In addition to the problem indicated by Barney, I am having problems with voices mixing (neural and classical) in my courses, even if I update the voices to the new ones, rewrite the text, or add an extra space at the end to force the update, the voice that sounds sometimes is the classic one. Even in the same paragraph both types of voices appear sometimes mixed! :(
Hi Diego!
It sounds like you hit a snag with classic and neural voices blending together in your Storyline course. Sorry to hear that but I'd be happy to assist!
If performing a repair of your Storyline installation doesn't improve the behavior, I'd be happy to take a closer look at your .story file to investigate this further. Feel free to share a copy here in the discussion or privately with our support engineers.
Looking forward to hearing from you!
This happened to me too. I've submitted a ticket to Articulate regarding the issue. I even tried opening up a new, blank project and using the same text. Same issue.
Hi Art,
Great call reaching out to our Support Engineers! I see you've connected with my teammate, Philam. You’re in excellent hands! We’ll continue the conversation in your support case to keep all information in one place. We appreciate your patience as we look into this further!
Diego, are you still experiencing this issue? It seems to have fixed itself ever since I made my original post.
Hi Art Fanega! Yes, some tts in my courses still have bugs and Articulate is investigating it, although I don't know if it will be a prioritary bug to resolve.
This also cause a huge issue with the voice Mathew. It would be a huge help for designers, and developers working on current courses with the old classic voices and giving us a chance to phase out the old one with the new one. I have over 200 courses I've been using Mathew as a voice over. This creates a lot of extra work to have to change it out,
We too have created lots of courses using Classic English UK Amy. We regularly need to update them with small changes. The new neural voices do sound a lot better, but it will be a real pain if we have to update the entire module to maintain consistency between slides, or downgrade to an earlier version of Storyline each time we need to make an amend. Could we have the Classic option still available please?
I have the same problem. I have tried all of the following to no avail:
None of these have worked and we now have to send the project to the client with audio in which the voice changes mid-stream for no apparent reason. If you were not going to make the old voices available, could you not have removed ALL of the old coding to prevent this from happening??
UPDATE: After Support Chat, I was told this is a "known bug." In my experience, that means we will not see a fix to the switching voices until they push out an update to fix it. Good luck explaining that to your clients. ~sigh~
I'm sorry for the trouble, Angela! I looked at your chat case with Mark, and it seems he's still investigating whether your situation falls under the reported bug with TTS, where it uses the wrong voice when the initials DF are present.
I encourage you to continue troubleshooting with Mark so that we can find a fix. I'm happy to assist further as needed!
Hello,
I am going through the same thing updating an older course. Is there any quick or simple way to convert the old voices to the new neural one throughout?
Thanks.