Task 3: Voice to Text

This is me in South Korea when I travelled there solo for the first time in summer 2016 after graduating under grad. The following story is an unscripted and unedited story from back then told using voice to text via the Notes app on my Iphone.

Back in 2016 I went to travel by myself for the first Time to South Korea I decided to go travelling you know as a graduation gift and I wanted to experience solo travel and also visit family of course so this is a story that took place during my solo travel and I was in Seoul South Korea which is a capital in a particular area called Holiday. Holiday is like a pretty touristy but hip neighbourhood and I was just standing around trying to figure out what I should do for that day in that area and these two girls that looked like their university students came out to me asking me for a directions for some store and could release since I’m not from there I I said I did not know where that was because I’m not from here. So that initiated a conversation with them they were asking me all like where are you from and I said I’m from Canada RNs conversation went on for a while. They were on then to start introducing me to some kind of a “on” cultural experience that I could go to as a visitor of Korea and I said oh that sounds interesting like where is this place and is it free and they said oh it is it’s not too far from here and it’s free for sure and I said good because I don’t have any money on me so if you do ask me for money after this I don’t have anything to offer.  I decided to follow them which was very naïve of me at the time now that I think back on it and it was actually in a pretty remote area that wasn’t close to anything really and so we got to take the subway and we got to this random building that just didn’t look like much. Is it look like a culture centre which is what I was expecting it was just a rundown building and so that got my alarm sirens were rolling in my head and I stopped in front of the building asking them like is this a safe thing that I’m going to and they’re like yeah it’s of course it’s fine. So again naïvely I went inside the building and yeah they were doing some kind of cultural experience but it was more like a weird shamanistic ritual that they were doing which in all say that in Korea there is a history of shaman ism but now it’s sort of died quite a bit. They they let me watch Watch the ritual and I yeah I did watch and then afterwards they offered me food and I did not take any because I don’t know if felt weird and then they asked me for money which is what I asked him initially I asked where that was I have to pay anything and they said no. So yeah I did end up giving them one dollar and then I left and another fishy thing was that they didn’t allow me to take any photos in that building so that definitely let you know that they were doing something very fishy. I use the story quite a bit to tell my students as a story time when we have a bit ofDowntime at least I came out of that situation with all of my organs in tact and they didn’t rob me of anything.

 

I personally haven’t actually used speech to text before this task. I have been on the receiving end of someone using it all the time to contact me or complete assignments using it. They always came with spelling and grammar errors, which goes to show how unreliable they can be even if you’re a frequent user of it. With my first attempt at using this feature on my phone I found that I unconsciously tried to enunciate more when speaking. If I was truly talking as I normally would, there would probably be a lot more mistakes. One of the first notable mistakes was the switch to the word “Holiday” for a Korean neighbourhood called HongDae. I think that if I used more foreign words, the Voice-to-text would have misinterpreted all of them. Another notable thing I noticed was where I said “Quote on quote” to emphasize the words cultural experience. But the speech-to-text wrote it as “on”. Many English speakers listening to that part would understand that the quotations should be around “cultural experience”. I didn’t realize that saying the word quote would actually make the punctuation be written.

One major thing that the Voice-to-text did not pick up was the punctuations. There are many run-on sentences because it didn’t know how to put in periods or commas or quotations. I thought it would be able to at least put in periods or commas where I paused while talking, but that barely happened. I think that these are mistakes because it can cause discomfort for the reader when punctuation is missing. Without it, you don’t know how something should be read and wouldn’t know how to clearly interpret the message. There were also a few times when the recording would randomly stop in the middle. So when I pressed record again and started where I left off that’s where there were it would write down the same words twice.

I think using this feature can help when making a first draft of something where you are mainly brainstorming ideas. It’s an easy way to put down as much information as you can in written form. You would still be able to get the general idea or gist of what was transcribed. Reading through my own story, I can still get a sense of what is happening. However, whenever I read over the mistakes, it takes more time for me to read that part of the story since I’m editing it in my head and thinking about what was meant to be there.

A lot of the emotions from my oral telling of my story are definitely lost when it’s written in text. The sarcasm, tone, pausing for dramatic effect, and most of all the reactions of the listener are lost. The listener can affect how the story is told and thus each time the story is told it’s a little different. Walter Ong (1984) talks about how oral telling’s of stories came before writing was invented and how they are completely different experiences for humans. Talking is different in that it’s not as structured but writing has more structure and thought put into making things “make sense” and has caused our talking to change to be more like how we write. Knowing that what I was saying was being written down made the way I talk change during the recording. Written text to me feels more permanent and thus feels like they are more important in a sense. Oral stories are more fleeting and less formal – thus making mistakes while talking feel less daunting since you can correct your mistakes right away. There is more flexibility and leniency when it comes to speaking in comparison to written words.

If I had scripted the story, I would have added in punctuation as I was speaking by physically saying “period, comma, quote” to add them in where they were needed. I would have spoken while enunciating more to make sure the Voice-to-text was able to pick up most of the story without any mistakes. I would be able to add more details to the story that would make it more interesting. As I was telling the story I realized at the end that I had left out key points to the story that gave it more flavour and excitement for the listener(s). Having a time limit of 5 minutes to tell my story did cause me the rush through the story to get to the end.

Even though there are many different apps for this purpose, I used the Notes App on my Iphone mainly due to the convenience. I think next time, I will try to use a different speech-to-text software that is “smart” enough to put in punctuations depending on how long my pauses are when speaking. This was an interesting experience that let me see and understand how the written form of speech can lose a lot of meaning in translation. Modern technology cannot fully interpret the human emotional nuances that are included in oral speech (at least not yet).

 

References:

Ong, W. (1982). Orality and literacy: The technologizing of the word. London and New York: Taylor and Francis Group.

And Ong’s Youtube video.

2 thoughts on “Task 3: Voice to Text

  1. I use voice to text all the time for making closed caption .txt files for my youtube videos. It is never straightforward and always requires editing. I often think of doctors creating notes for themselves using voice recorders and wonder if the technology is sophisticated enough yet to handle all the medical terms. I know historically they used to use a voice recorder then transcribe, I wonder if they are able to go straight to voice to text yet? If you could train your voice to text by correcting it then the Korean place names you used, or any other foreign language words could be added to your voice-to-text dictionary. I suppose that is where AI comes into play.

    • You are the second person I know who uses voice to text regularly! I guess it makes sense for youtube video closed captions. Youtube’s auto closed captions for English is quite good in my opinion, but it’s pretty terrible in Korean. I never thought of how AI can be used so voice-to-text can learn specific words – even foreign ones! That would be quite interesting to use and see how accurate it would get as you teach it to capture your nuanced way of speaking.

Leave a Reply

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