Decoding the Difference: Present vs Past Tense in Speech Recognition

  • Thread starter Thread starter inotyce
  • Start date Start date
Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
18 replies · 3K views
inotyce
Messages
43
Reaction score
1
We can distinguish between present "read" and past "read" in their pronunciation. I wonder how a computer knows the difference. :biggrin:
 
Physics news on Phys.org
inotyce said:
We can distinguish between present "read" and past "read" in their pronunciation. I wonder how a computer knows the difference. :biggrin:
Here's a sentence:

"Yesterday I read 4 emails."

You will not hear "read" pronounced as you read that sentence, but you'll know the difference from the context.
 
SteamKing said:
I read 4 emails.

I read 4 emails.

Both of these are past tense. For present tense, English uses the present progressive:

"I am reading 4 emails."
 
Ben Niehoff said:
Both of these are past tense. For present tense, English uses the present progressive:

Well, how about this:

What if I read all your emails? ("reed" meaning I am going to do it later today)

What if I read all your emails? ("red" meaning I did it yesterday)

Now we have future/past, but the same confusion.
 
Last edited:
Ben Niehoff said:
Both of these are past tense. For present tense, English uses the present progressive:

"I am reading 4 emails."

Not always. English has multiple forms for each tense:
1. There is the simple form: I read 4 emails. ['read' here is first person singular present tense.]
2. There is the progressive form: I am reading 4 emails. ['am reading' is still first person singular present tense.]
3. There is the emphatic form: I do read 4 emails. ['do read' likewise is first person singular present tense.]

Similar constructions are found for the past and future tenses.

For the first person singular form of the verb 'to read', the simple present and the simple past tenses have the same written form, although they are pronounced differently.
 
phinds said:
Well, how about this:

What if I read all your emails? ("reed" meaning I am going to do it later today)

What if I read all your emails? ("red" meaning I did it yesterday)

Now we have future/past, but the same confusion.

It's only confusing because it's incomplete, the tense of the homograph is DEPENDENT on it's context/pronunciation - you have provided neither with the isolated statement "What if I read all your emails?"
 
phinds said:
Well, how about this:

What if I read all your emails? ("reed" meaning I am going to do it later today)

What if I read all your emails? ("red" meaning I did it yesterday)

Now we have future/past, but the same confusion.

The first example is not future tense.
In English, the future tense would be expressed by: "What if I shall read all your emails?"

You can infer intent about when the reading takes place, but you can't take the sentence as is and parse the verb 'read' as future tense.
 
SteamKing said:
The first example is not future tense.
In English, the future tense would be expressed by: "What if I shall read all your emails?"

You can infer intent about when the reading takes place, but you can't take the sentence as is and parse the verb 'read' as future tense.

Have you seen the books I read?
- What about the books you read?

"read" can easily be either reed/red in both cases
 
phinds said:
Well, how about this:

What if I read all your emails? ("reed" meaning I am going to do it later today)

What if I read all your emails? ("red" meaning I did it yesterday)

Now we have future/past, but the same confusion.
It would be very unusual for this sentence to exist in isolation, without a context. Generally we, or a computer, would know the tense meant from the conversational context.
 
zoobyshoe said:
It would be very unusual for this sentence to exist in isolation, without a context. Generally we, or a computer, would know the tense meant from the conversational context.

Damn, there you go getting all reasonable on me :smile:

I DO think that it's possible the context could be confusing enough that it would be tough for a computer (it COULD even be tough for a person).
 
So glad I didn't have to learn english as a second language :D
 
jim, spell check doesn't know the context of the word, it just checks to see if it's a real word. Which 'lead' happens to be.

You can do natural language processing to try and parse out tenses but it's a lot of work and examples like this are well known to be difficult to deal with
 
phinds said:
Damn, there you go getting all reasonable on me :smile:

I DO think that it's possible the context could be confusing enough that it would be tough for a computer (it COULD even be tough for a person).
Yeah, I certainly think it's possible to confuse a computer.

I was at a coffeehouse a couple weeks ago and a guy was checking out his voice recognition ap on his phone: he had an ap that would convert what he spoke into the phone into a text. He was demonstrating to a friend the sort of mistake the software made when trying to deal with casual speech. It did make mistakes, but not as many as you'd think. Probably because you have the option to correct it, so it learns greater and greater accuracy vis a vis your particular speech.
 
D H said:
I read you loud and clear, and I disagree. I never would say "I am reading you loud and clear."

I would never say neither (I only say neither).

I would say "I read you five by five." Or "five by two" if the volume was fine but the voice quality nearly inaudible. Or "two by five" if the volume were so low I could barely hear your voice. Or say nothing if I read you "zero by five" (I wouldn't know you even asked).