prev1
stringlengths 1
7.9k
| prev2
stringlengths 1
7.9k
⌀ | next1
stringlengths 1
7.9k
| next2
stringlengths 1
7.9k
⌀ | is_boundary
bool 2
classes | session_id
stringlengths 7
14
|
|---|---|---|---|---|---|
Sure looks
|
but
|
Hmm.
|
Well , no.
| false |
QMSum_120
|
but
|
Hmm.
|
Well , no.
|
Hmm.
| false |
QMSum_120
|
Hmm.
|
Well , no.
|
Hmm.
|
It is
| false |
QMSum_120
|
Well , no.
|
Hmm.
|
It is
|
Yeah. Because we begin , uh , in fifteen point the fifteen point.
| false |
QMSum_120
|
Hmm.
|
It is
|
Yeah. Because we begin , uh , in fifteen point the fifteen point.
|
So , does does the periodicity of this signal say something about the the
| false |
QMSum_120
|
It is
|
Yeah. Because we begin , uh , in fifteen point the fifteen point.
|
So , does does the periodicity of this signal say something about the the
|
Fifteen p
| false |
QMSum_120
|
Yeah. Because we begin , uh , in fifteen point the fifteen point.
|
So , does does the periodicity of this signal say something about the the
|
Fifteen p
|
So it 's Yeah.
| false |
QMSum_120
|
So , does does the periodicity of this signal say something about the the
|
Fifteen p
|
So it 's Yeah.
|
Pitch.
| false |
QMSum_120
|
Fifteen p
|
So it 's Yeah.
|
Pitch.
|
It 's the pitch.
| false |
QMSum_120
|
So it 's Yeah.
|
Pitch.
|
It 's the pitch.
|
the pitch ?
| false |
QMSum_120
|
Pitch.
|
It 's the pitch.
|
the pitch ?
|
Yeah. Mm - hmm.
| false |
QMSum_120
|
It 's the pitch.
|
the pitch ?
|
Yeah. Mm - hmm.
|
Yeah.
| false |
QMSum_120
|
the pitch ?
|
Yeah. Mm - hmm.
|
Yeah.
|
OK.
| false |
QMSum_120
|
Yeah. Mm - hmm.
|
Yeah.
|
OK.
|
That 's like fundamental frequency.
| false |
QMSum_120
|
Yeah.
|
OK.
|
That 's like fundamental frequency.
|
Mm - hmm.
| false |
QMSum_120
|
OK.
|
That 's like fundamental frequency.
|
Mm - hmm.
|
So , I mean , i t t
| false |
QMSum_120
|
That 's like fundamental frequency.
|
Mm - hmm.
|
So , I mean , i t t
|
OK. I see.
| false |
QMSum_120
|
Mm - hmm.
|
So , I mean , i t t
|
OK. I see.
|
I mean , to first order what you 'd what you 're doing I mean , ignore all the details and all the ways which is that these are complete lies. Uh , the the you know , what you 're doing in feature extraction for speech recognition is you have , uh , in your head a a a a simplified production model for speech ,
| false |
QMSum_120
|
So , I mean , i t t
|
OK. I see.
|
I mean , to first order what you 'd what you 're doing I mean , ignore all the details and all the ways which is that these are complete lies. Uh , the the you know , what you 're doing in feature extraction for speech recognition is you have , uh , in your head a a a a simplified production model for speech ,
|
Mm - hmm.
| false |
QMSum_120
|
OK. I see.
|
I mean , to first order what you 'd what you 're doing I mean , ignore all the details and all the ways which is that these are complete lies. Uh , the the you know , what you 're doing in feature extraction for speech recognition is you have , uh , in your head a a a a simplified production model for speech ,
|
Mm - hmm.
|
in which you have a periodic or aperiodic source that 's driving some filters.
| false |
QMSum_120
|
I mean , to first order what you 'd what you 're doing I mean , ignore all the details and all the ways which is that these are complete lies. Uh , the the you know , what you 're doing in feature extraction for speech recognition is you have , uh , in your head a a a a simplified production model for speech ,
|
Mm - hmm.
|
in which you have a periodic or aperiodic source that 's driving some filters.
|
Mm - hmm.
| false |
QMSum_120
|
Mm - hmm.
|
in which you have a periodic or aperiodic source that 's driving some filters.
|
Mm - hmm.
|
Yeah. This is the the auto - correlation the R - zero energy.
| false |
QMSum_120
|
in which you have a periodic or aperiodic source that 's driving some filters.
|
Mm - hmm.
|
Yeah. This is the the auto - correlation the R - zero energy.
|
Do you have the mean do you have the mean for the auto - correlation ?
| false |
QMSum_120
|
Mm - hmm.
|
Yeah. This is the the auto - correlation the R - zero energy.
|
Do you have the mean do you have the mean for the auto - correlation ?
|
Uh , first order for speech recognition , you say " I don't care about the source ".
| false |
QMSum_120
|
Yeah. This is the the auto - correlation the R - zero energy.
|
Do you have the mean do you have the mean for the auto - correlation ?
|
Uh , first order for speech recognition , you say " I don't care about the source ".
|
For Yeah.
| false |
QMSum_120
|
Do you have the mean do you have the mean for the auto - correlation ?
|
Uh , first order for speech recognition , you say " I don't care about the source ".
|
For Yeah.
|
Well , I mean for the the energy.
| false |
QMSum_120
|
Uh , first order for speech recognition , you say " I don't care about the source ".
|
For Yeah.
|
Well , I mean for the the energy.
|
I have the mean.
| false |
QMSum_120
|
For Yeah.
|
Well , I mean for the the energy.
|
I have the mean.
|
Right ?
| false |
QMSum_120
|
Well , I mean for the the energy.
|
I have the mean.
|
Right ?
|
Right.
| false |
QMSum_120
|
I have the mean.
|
Right ?
|
Right.
|
And so you just want to find out what the filters are.
| false |
QMSum_120
|
Right ?
|
Right.
|
And so you just want to find out what the filters are.
|
Right.
| false |
QMSum_120
|
Right.
|
And so you just want to find out what the filters are.
|
Right.
|
Yeah.
| false |
QMSum_120
|
And so you just want to find out what the filters are.
|
Right.
|
Yeah.
|
The filters roughly act like a , um a , uh a an overall resonant you know , f some resonances and so forth that th that 's processing excitation.
| false |
QMSum_120
|
Right.
|
Yeah.
|
The filters roughly act like a , um a , uh a an overall resonant you know , f some resonances and so forth that th that 's processing excitation.
|
Here.
| false |
QMSum_120
|
Yeah.
|
The filters roughly act like a , um a , uh a an overall resonant you know , f some resonances and so forth that th that 's processing excitation.
|
Here.
|
They should be more close.
| false |
QMSum_120
|
The filters roughly act like a , um a , uh a an overall resonant you know , f some resonances and so forth that th that 's processing excitation.
|
Here.
|
They should be more close.
|
Ah , no. This is this ? More close. Is this ? And this.
| false |
QMSum_120
|
Here.
|
They should be more close.
|
Ah , no. This is this ? More close. Is this ? And this.
|
Mm - hmm.
| false |
QMSum_120
|
They should be more close.
|
Ah , no. This is this ? More close. Is this ? And this.
|
Mm - hmm.
|
Yeah.
| false |
QMSum_120
|
Ah , no. This is this ? More close. Is this ? And this.
|
Mm - hmm.
|
Yeah.
|
Mm - hmm.
| false |
QMSum_120
|
Mm - hmm.
|
Yeah.
|
Mm - hmm.
|
So they are this is there is less difference.
| false |
QMSum_120
|
Yeah.
|
Mm - hmm.
|
So they are this is there is less difference.
|
Mm - hmm.
| false |
QMSum_120
|
Mm - hmm.
|
So they are this is there is less difference.
|
Mm - hmm.
|
So if you look at the spectral envelope , just the very smooth properties of it , you get something closer to that.
| false |
QMSum_120
|
So they are this is there is less difference.
|
Mm - hmm.
|
So if you look at the spectral envelope , just the very smooth properties of it , you get something closer to that.
|
This is less it 's less robust.
| false |
QMSum_120
|
Mm - hmm.
|
So if you look at the spectral envelope , just the very smooth properties of it , you get something closer to that.
|
This is less it 's less robust.
|
Less robust. Yeah.
| false |
QMSum_120
|
So if you look at the spectral envelope , just the very smooth properties of it , you get something closer to that.
|
This is less it 's less robust.
|
Less robust. Yeah.
|
Oh , yeah.
| false |
QMSum_120
|
This is less it 's less robust.
|
Less robust. Yeah.
|
Oh , yeah.
|
And the notion is if you have the full spectrum , with all the little nitty - gritty details , that that has the effect of both ,
| false |
QMSum_120
|
Less robust. Yeah.
|
Oh , yeah.
|
And the notion is if you have the full spectrum , with all the little nitty - gritty details , that that has the effect of both ,
|
Yeah.
| false |
QMSum_120
|
Oh , yeah.
|
And the notion is if you have the full spectrum , with all the little nitty - gritty details , that that has the effect of both ,
|
Yeah.
|
and it would be a multiplication in in frequency domain
| false |
QMSum_120
|
And the notion is if you have the full spectrum , with all the little nitty - gritty details , that that has the effect of both ,
|
Yeah.
|
and it would be a multiplication in in frequency domain
|
Mm - hmm.
| false |
QMSum_120
|
Yeah.
|
and it would be a multiplication in in frequency domain
|
Mm - hmm.
|
so that would be like an addition in log power spectrum domain.
| false |
QMSum_120
|
and it would be a multiplication in in frequency domain
|
Mm - hmm.
|
so that would be like an addition in log power spectrum domain.
|
Mm - hmm. Mm - hmm.
| false |
QMSum_120
|
Mm - hmm.
|
so that would be like an addition in log power spectrum domain.
|
Mm - hmm. Mm - hmm.
|
And so this is saying , well , if you really do have that sort of vocal tract envelope , and you subtract that off , what you get is the excitation. And I call that lies because you don't really have that , you just have some kind of signal - processing trickery to get something that 's kind of smooth. It 's not really what 's happening in the vocal tract
| false |
QMSum_120
|
so that would be like an addition in log power spectrum domain.
|
Mm - hmm. Mm - hmm.
|
And so this is saying , well , if you really do have that sort of vocal tract envelope , and you subtract that off , what you get is the excitation. And I call that lies because you don't really have that , you just have some kind of signal - processing trickery to get something that 's kind of smooth. It 's not really what 's happening in the vocal tract
|
Yeah.
| false |
QMSum_120
|
Mm - hmm. Mm - hmm.
|
And so this is saying , well , if you really do have that sort of vocal tract envelope , and you subtract that off , what you get is the excitation. And I call that lies because you don't really have that , you just have some kind of signal - processing trickery to get something that 's kind of smooth. It 's not really what 's happening in the vocal tract
|
Yeah.
|
so you 're not really getting the vocal excitation.
| false |
QMSum_120
|
And so this is saying , well , if you really do have that sort of vocal tract envelope , and you subtract that off , what you get is the excitation. And I call that lies because you don't really have that , you just have some kind of signal - processing trickery to get something that 's kind of smooth. It 's not really what 's happening in the vocal tract
|
Yeah.
|
so you 're not really getting the vocal excitation.
|
Right.
| false |
QMSum_120
|
Yeah.
|
so you 're not really getting the vocal excitation.
|
Right.
|
That 's why I was going to the why I was referring to it in a more a more , uh , uh , conservative way , when I was saying " well , it 's yeah , it 's the excitation ". But it 's not really the excitation. It 's whatever it is that 's different between
| false |
QMSum_120
|
so you 're not really getting the vocal excitation.
|
Right.
|
That 's why I was going to the why I was referring to it in a more a more , uh , uh , conservative way , when I was saying " well , it 's yeah , it 's the excitation ". But it 's not really the excitation. It 's whatever it is that 's different between
|
Oh. This moved in the
| false |
QMSum_120
|
Right.
|
That 's why I was going to the why I was referring to it in a more a more , uh , uh , conservative way , when I was saying " well , it 's yeah , it 's the excitation ". But it 's not really the excitation. It 's whatever it is that 's different between
|
Oh. This moved in the
|
So so , stand standing back from that , you sort of say there 's this very detailed representation.
| false |
QMSum_120
|
That 's why I was going to the why I was referring to it in a more a more , uh , uh , conservative way , when I was saying " well , it 's yeah , it 's the excitation ". But it 's not really the excitation. It 's whatever it is that 's different between
|
Oh. This moved in the
|
So so , stand standing back from that , you sort of say there 's this very detailed representation.
|
Yeah.
| false |
QMSum_120
|
Oh. This moved in the
|
So so , stand standing back from that , you sort of say there 's this very detailed representation.
|
Yeah.
|
You go to a smooth representation.
| false |
QMSum_120
|
So so , stand standing back from that , you sort of say there 's this very detailed representation.
|
Yeah.
|
You go to a smooth representation.
|
Mm - hmm.
| false |
QMSum_120
|
Yeah.
|
You go to a smooth representation.
|
Mm - hmm.
|
You go to a smooth representation cuz this typically generalizes better.
| false |
QMSum_120
|
You go to a smooth representation.
|
Mm - hmm.
|
You go to a smooth representation cuz this typically generalizes better.
|
Mm - hmm.
| false |
QMSum_120
|
Mm - hmm.
|
You go to a smooth representation cuz this typically generalizes better.
|
Mm - hmm.
|
Um , but whenever you smooth you lose something , so the question is have you lost something you can you use ?
| false |
QMSum_120
|
You go to a smooth representation cuz this typically generalizes better.
|
Mm - hmm.
|
Um , but whenever you smooth you lose something , so the question is have you lost something you can you use ?
|
Right.
| false |
QMSum_120
|
Mm - hmm.
|
Um , but whenever you smooth you lose something , so the question is have you lost something you can you use ?
|
Right.
|
Um , probably you wouldn't want to go to the extreme of just ta saying " OK , our feature set will be the FFT " , cuz we really think we do gain something in robustness from going to something smoother , but maybe there 's something that we missed.
| false |
QMSum_120
|
Um , but whenever you smooth you lose something , so the question is have you lost something you can you use ?
|
Right.
|
Um , probably you wouldn't want to go to the extreme of just ta saying " OK , our feature set will be the FFT " , cuz we really think we do gain something in robustness from going to something smoother , but maybe there 's something that we missed.
|
Mm - hmm.
| false |
QMSum_120
|
Right.
|
Um , probably you wouldn't want to go to the extreme of just ta saying " OK , our feature set will be the FFT " , cuz we really think we do gain something in robustness from going to something smoother , but maybe there 's something that we missed.
|
Mm - hmm.
|
So what is it ?
| false |
QMSum_120
|
Um , probably you wouldn't want to go to the extreme of just ta saying " OK , our feature set will be the FFT " , cuz we really think we do gain something in robustness from going to something smoother , but maybe there 's something that we missed.
|
Mm - hmm.
|
So what is it ?
|
Yeah.
| false |
QMSum_120
|
Mm - hmm.
|
So what is it ?
|
Yeah.
|
And then you go back to the intuition that , well , you don't really get the excitation , but you get something related to it.
| false |
QMSum_120
|
So what is it ?
|
Yeah.
|
And then you go back to the intuition that , well , you don't really get the excitation , but you get something related to it.
|
Mm - hmm.
| false |
QMSum_120
|
Yeah.
|
And then you go back to the intuition that , well , you don't really get the excitation , but you get something related to it.
|
Mm - hmm.
|
And it and as you can see from those pictures , you do get something that shows some periodicity , uh , in frequency ,
| false |
QMSum_120
|
And then you go back to the intuition that , well , you don't really get the excitation , but you get something related to it.
|
Mm - hmm.
|
And it and as you can see from those pictures , you do get something that shows some periodicity , uh , in frequency ,
|
Mm - hmm.
| false |
QMSum_120
|
Mm - hmm.
|
And it and as you can see from those pictures , you do get something that shows some periodicity , uh , in frequency ,
|
Mm - hmm.
|
you know , and and and also in time.
| false |
QMSum_120
|
And it and as you can see from those pictures , you do get something that shows some periodicity , uh , in frequency ,
|
Mm - hmm.
|
you know , and and and also in time.
|
Hmm.
| false |
QMSum_120
|
Mm - hmm.
|
you know , and and and also in time.
|
Hmm.
|
So
| false |
QMSum_120
|
you know , and and and also in time.
|
Hmm.
|
So
|
That 's that 's really neat.
| false |
QMSum_120
|
Hmm.
|
So
|
That 's that 's really neat.
|
so ,
| false |
QMSum_120
|
So
|
That 's that 's really neat.
|
so ,
|
So you don't have one for unvoiced picture ?
| false |
QMSum_120
|
That 's that 's really neat.
|
so ,
|
So you don't have one for unvoiced picture ?
|
Uh , not here.
| false |
QMSum_120
|
so ,
|
So you don't have one for unvoiced picture ?
|
Uh , not here.
|
Oh.
| false |
QMSum_120
|
So you don't have one for unvoiced picture ?
|
Uh , not here.
|
Oh.
|
No , I have s
| false |
QMSum_120
|
Uh , not here.
|
Oh.
|
No , I have s
|
Mm - hmm.
| false |
QMSum_120
|
Oh.
|
No , I have s
|
Mm - hmm.
|
Yeah.
| false |
QMSum_120
|
No , I have s
|
Mm - hmm.
|
Yeah.
|
But not here.
| false |
QMSum_120
|
Mm - hmm.
|
Yeah.
|
But not here.
|
But presumably you 'll see something that won't have this kind of , uh , uh , uh , regularity in frequency , uh , in the
| false |
QMSum_120
|
Yeah.
|
But not here.
|
But presumably you 'll see something that won't have this kind of , uh , uh , uh , regularity in frequency , uh , in the
|
But Yeah. Well.
| false |
QMSum_120
|
But not here.
|
But presumably you 'll see something that won't have this kind of , uh , uh , uh , regularity in frequency , uh , in the
|
But Yeah. Well.
|
Not here.
| false |
QMSum_120
|
But presumably you 'll see something that won't have this kind of , uh , uh , uh , regularity in frequency , uh , in the
|
But Yeah. Well.
|
Not here.
|
I would li I would like to see those pictures.
| false |
QMSum_120
|
But Yeah. Well.
|
Not here.
|
I would li I would like to see those pictures.
|
Well , so.
| false |
QMSum_120
|
Not here.
|
I would li I would like to see those pictures.
|
Well , so.
|
Yeah.
| false |
QMSum_120
|
I would li I would like to see those pictures.
|
Well , so.
|
Yeah.
|
I can't see you now.
| false |
QMSum_120
|
Well , so.
|
Yeah.
|
I can't see you now.
|
Yeah.
| false |
QMSum_120
|
Yeah.
|
I can't see you now.
|
Yeah.
|
Yeah.
| false |
QMSum_120
|
I can't see you now.
|
Yeah.
|
Yeah.
|
Yeah.
| false |
QMSum_120
|
Yeah.
|
Yeah.
|
Yeah.
|
Mm - hmm.
| false |
QMSum_120
|
Yeah.
|
Yeah.
|
Mm - hmm.
|
I don't have.
| false |
QMSum_120
|
Yeah.
|
Mm - hmm.
|
I don't have.
|
And so you said this is pretty doing this kind of thing is pretty robust to noise ?
| false |
QMSum_120
|
Mm - hmm.
|
I don't have.
|
And so you said this is pretty doing this kind of thing is pretty robust to noise ?
|
It seems , yeah. Um ,
| false |
QMSum_120
|
I don't have.
|
And so you said this is pretty doing this kind of thing is pretty robust to noise ?
|
It seems , yeah. Um ,
|
Huh.
| false |
QMSum_120
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.