Bernice Summerfield: The Poison Seas

[vivdunstan]

Currently listening to this. And this is by far the most atrocious voice modulation I've ever experienced with Big Finish. I thought it was bad with the opening monologue. But that Sea Devil was vastly more listenable to than the other Sea Devils later. As a result I'm only getting a fraction of the plot. And it's just so very disappointing. And frustrating. Because the underlying plot is intriguing. And Benny is fab as always. But even if I whack the volume up massively I can barely understand a word from the Sea Devils. And then my ears are completely blasted if I don't adjust the volume quickly enough when other people are about to speak. If I wasn't such an obsessive I'd give up on this. But am hanging on in there. Albeit expecting to miss 90% of what the Sea Devils say.

Benny has an utterly brilliant monologue near the end though. That was worth hearing if nothing else. But yup, definitely missed much of what was going on this time.

Comments

[a_cubed]

Could you pipe the audio through an auto-transcription service? Maybe it would be too muddy for those as well, or maybe you couldn't read the resulting text quickly enough. Just an idle thought. I did look to see if anyone has transcribed these audio stories but a quick search didn't turn one up for this, anyway.

[vivdunstan]

Thanks, that's a really good idea. I've just tried a 1 minute test portion in MacWhisper, which I can run on my Mac, no remote upload, and try for free before buying. It got an impressively large amount of a female whispered Sea Devil speech! Though couldn't detect some at all. But still very impressive. I will find this useful for more critical stuff too. So many thanks!

[a_cubed]

You're welcome. Just as with computer vision, the way computers process speech to text is completely different to the way humans do, so sometimes they're crap at things people are good at, but vice versa. I've used a free transcription program to create an initial set of English subtitles for some of my pre-recorded lectures aimed at Japanese students. It was definitely better than transcribing them myself from scratch (which I have done for both subtitles and recorded interviews in the past) though still needed a lot of correcting. Some things could be done with global search and replace since it picked up my accent as different words consistently, or an unusual piece of vocabulary as a more common similar sounding word. Others it just needed correcting because it had mis-interpreted. I know my students sometimes use Zoom's auto-caption for my live online lectures, to help them.

[vivdunstan]

I am extremely impressed at how well it coped with the whispery quiet Sea Devils. Though also somewhat comforted that there were still some aspects of their speech it couldn't detect anything of at all!

[a_cubed]

I read (sorry, I'd have to do a deep search to find it) an interesting analysis of how poorly modern sound mixing for television (and for movies "broadcast" on television - I am of course including streaming services on this) has been. As sound reproduction by TVs (and on computers, tablets and smartphones) has got better, the sound design has allowed the music and sound effects to be better heard. However, this is often at the expense of speech legibility. I'm pretty sure there are now a lot of people using subtitles who don't have any hearing loss, they just want to ensure they don't miss dialogue washed out by sound effects or music. Deliberately distorted speech (whether SFX or acted) can make this worse and of course although it's possible to have subtitles for audio plays, few provide them (there is a standard for lyrics to be shown during music, which could just as easily be applied to subtitles for audio plays).

On the Japanese news the other day (I watch a five minute late night bulletin from NHK most days, which we record from broadcast and $DAUGHTER and I usually watch in the evening - me for the language training and her for current affairs education) there was a brief interview with someone who had successfully sued the Japan Olympic Committee for harassment. They did not show her face and severely distorted her voice. It was so bad $DAUGHTER and I were wincing at the fingernails-on-blackboard quality. We both agreed that a voiceover with a deliberately different accent (maybe even a male speaker just to make it clear) would have been better than that horrible effect, which was very hard to decipher as well as being annoying to hear.

Wow, that turned into a long post!

[vivdunstan]

That's really interesting, thanks. I've had a lot of trouble especially with Doctor Who TV audio in recent years. It's been especially a problem since Murray Gold came back as the music composer, and the music seems _very_ loud, making the audio harder to hear. But understanding more about the mixing issues makes sense. A lot of other ardent Whovians complain of similar issues. And not all the types you'd expect to have traditional hearing loss. I know I have some extra hearing problems, from my neurological disease. But the sound mix in some specific programmes is a particular issue. I don't usually turn the subtitles on, because I'd have trouble reading those quickly enough too! Poor Martin has to just cope with the volume turned up instead :)