Task 3: Voice to Text

This story was transcribed from spoken word to text via an application called Speechnotes (https://speechnotes.co/). The story is followed by an analysis of the voice to text process and the text itself.

My story:

My daughter was born on July 4th 2016 it was a Monday and I remember was a Monday because I remember thinking that she wanted to give us the long weekend and then she would be born anyways it was the Canada Day long weekend but as it stands she was born on the American independence day so what happened was early that morning my wife woke up at around 7 I am and she said that she had felt a little different and we we had we’re obviously we’re anticipating the birth of our child is getting pretty close so but the dishes are midwives had told us to I told her to take a shower or have a warm bath and then go back to sleep because you want to delay going to the hospital for as long as you can so so she did she took a shower and then she hopped back internet and she woke up a couple hours later feeling maybe it was Labor so I called are midwives and they told us that she told us that she was already at the hospital but then when she was finished there she would come over and attend to my wife and I thought okay that’s that’s totally fine we’ll just wait but as the time passed labour. Simply more intense I remember there being stupid Wives coming to attend and my wife is worried refused to get into the car and go to the hospital and The Midwives were pretty intent on having the birth of our first child in our bedroom on a bunch of help so that’s what I thought would happen and so she was in labour for many hours until there was a shift change in the midwives and another Midwife came to attend so we had three midwives in our bedroom with myself and my wife trying to breathe through the labour and at a certain point I can’t even remember what time the two initial midwives left and went home and we were left with shift change midwife and so she said that we would have to that my wife would have to push for maybe about another half hour and then we will have to call an ambulance to come and assist with the birth or at least assist us to getting to the hospital and so my wife did try to push for a couple hours for not a couple hours pick another half hour and when my daughter was still not born we called the ambulance to come give us a lift 2 the hospital I remember the Midwife sitting next to my wife trying to get her to continue to breathe calmly and not push because we don’t want have the birth in the ambulance we would rather have it at the hospital so what I do remember is that we checked in at the hospital at 7:40 yeah and my daughter was born at 8:20 yeah

How does the text deviate from conventions of written English?

The text deviates from written English in many area, such as punctuation, grammar and spelling, though the gist of the story remains intact. There is no punctuation to indicate the end of a sentence; no periods, commas or exclamation marks. There are; however, apostrophes for where I used a contraction. I did make natural stops and pauses in my story telling but the technology did not consider that for using sentence ending punctuation. And there were instructions to say the type of punctuation needed to end the sentence, but it felt more natural to not to speak in that manner.

What is “wrong” in the text? What is “right”?

There are several errors in the text including many repeated words and words that were ‘misheard’ by the application that were transcribed incorrectly in the text. For example, I said ‘hopped back into bed,’ and the app heard ‘hopped back internet’ and ‘midwives’ as ‘stupid Wives,’ (which I would never call my wife or midwife). There are several of these errors in the text. The story appears in text form as one run on sentence. As I mentioned above, there is no punctuation to indicate beginning and ending of sentences, nor is there any punctuation for dates. Some words in the text had capitalization errors, such as capitalizing Labour, Midwives and Wives. The app also did not transcribe time into text well, ‘7:47’ was transcribed as 7:40 and ‘8:28’ was transcribed as 8:20.

The voice application got the story ‘right.’ The majority of the words are correct, if you were to read the text you would understand the story I was trying to convey. The text did not deviate much from the story.

What are the most common “mistakes” in the text and why do you consider them “mistakes”?

The most common mistakes were a lack of punctuation, doubling some words and incorrect transcription of words and phrases. I would consider these as mistakes as compared to conventional written text; however, I also understand that written text also goes through a process of editing.

What if you had “scripted” the story? What difference might that have made?

I think if I had scripted the story, the app would have an easier time transcribing my story to text. Scripting the story and allowing myself to read it smoothly, without starts and stops that affected the transcription of the text. As a result, the text would be tighter, in terms of less errors and duplication of words. I think the written story would have better in terms of readability, there would be more flow when reading with less stops and starts. I found myself fumbling to find the best word during my story telling, that that might have created many of the errors that appeared in the text. In addition, scripting the story would also allow myself to think about the story, plot it out and choose words ahead of time all contributing to the fluidity of the text in the end product.

In what ways does oral storytelling differ from written storytelling?

In general, an oral tradition of story telling is more fluid and dynamic. I think stories told orally are never told exactly the same way twice, and I think oral traditions are intend it to be that way. In a written story, only different readers get to interpret the information differently, as the text is static and the readers are the variables. Furthermore, because a story is never told the same way twice, the storyteller can have a different interpretation and telling of the story each time. For example, by adding new elements, delaying or deleting certain parts, and changing the order of events, emphasizing and de-emphasizing portions of the story, the meaning of the story can vary each time it’s told. I also think there is an element of collaboration in oral storytelling, in that, if the listener can ask questions or comments during the story. A conversation during the story can affect the direction and details of the story. The opportunity to have interaction can change both the intention and the story itself. A written story is quite static, where only the reader reads the text and interprets it for themselves.

4 thoughts on “Task 3: Voice to Text

  1. I had the same experience as you using Speechnotes! I had natural pauses and stops too but the software did not even attempt to put in basic commas and periods for me.

    I appreciate your comment about oral story having an element of collaboration. “Conversation” was certainly something I overlooked as I reflected on oral vs written storytelling. It did not cross my mind at all that there could be conversations during the telling of a story; I interpreted oral storytelling as a monologue, but there is no reason why it couldn’t be a dialogue! This makes me wonder though…whether the dialogue between the storyteller and the listener(s) should be consider part of the story, since it certainly would be part of that specific storytelling experience. Obviously, subsequent retelling of the story may or may not include dialogues, depending on whether the listener(s) engage with the story the same way.

    • Hi Christopher, thanks for reading and leaving your comment. Interestingly, I’ve always thought of storytelling almost like a facilitated conversation. I think that the dialogue between the teller and the listeners would be part of the entire storytelling experience. I also think that being able to have that conversation, personalizes the story and creates more meaning for the listener. Thanks, Shawn

  2. Hey Shawn,
    I also referred to some mistakes as errors of ‘mishearing’ – I think that is such a good way to encapsulate the origins of the mistakes. Why do you think this happens? Is the program not nuanced enough, or is it due to the building blocks given to the program (i.e. the soundbites it was taught with)?
    I like your comment on collaboration in storytelling too. I think that is a vital part of most communication – one of the strangest parts of this task for me was just talking to myself for 5 mins!
    🙂

    • Hi Jamie,
      Thanks for the comment. I think you’re probably right about how the program is ‘trained’ to hear and the nuances of individuals speaking styles. I know that I pause a lot. I mostly developed that from giving many presentations and I used to say ‘ummm’ as a pause to think and essentially replaced it with a pause. I found that many of the ‘mishearings’ came after a pause. Maybe it’s the program trying to make grammatical at the same time doing the voice to text. I don’t think an app can reconcile doing two things at the same time, while a person can make these accommodations to make sense of language.
      I also found it weird to talk about a story to myself (though to a computer). I had to do it in the bathroom with the fan on to drown out my voice so I don’t wake up the kids.
      Shawn

Leave a Reply

Your email address will not be published. Required fields are marked *