speaker stringclasses 50
values | text stringlengths 1 431 | basic_da stringclasses 5
values | general_da stringclasses 12
values | full_da stringclasses 52
values |
|---|---|---|---|---|
fe016 | and | F | fh | fh |
me011 | oh well. | S | s | bk |
fe016 | um | F | fh | fh |
fe016 | i i guess my question is more uh what what do you do with say a forced alignment? | Q | qw | qw |
fe016 | i mean you've got all these phone labels. | S | s | df |
fe016 | and what do you do if you just conceptually if you get um transcriptions where the words are staying but the time boundaries are changing because you've got a new recognition output? | Q | qw | qw |
fe016 | or sort of what's the um sequence of going from the waveforms that stay the same the transcripts that may or may not change and then the utterance which where the time boundaries that may or may not change? | Q | qw | qw |
mn017 | oh that's | D | s | bk |
mn017 | that's actually very nicely handled here. | S | s | s |
fe016 | um | F | fh | fh |
mn017 | because you could you could all you'd have to change is the um time stamps in the timeline without without uh changing the i.d.'s. | S | s | df |
fe016 | and you'd be able to propagate all of the the information? | Q | qy | bu |
me011 | right. | S | s | bk |
me011 | that's the that's why you do that extra level of indirection. | S | s | s |
me011 | so that you can just change the timeline. | S | s | df |
mn017 | except the timeline is going to be huge. | S | s | s |
mn017 | if you say | D | s | s |
me011 | yes. | S | s | aa |
fe016 | yeah. | S | s | aa |
fe016 | yeah especially at the phone level. | S | s | na |
mn017 | suppose you have a phone level alignment. | S | s | e |
fe016 | the we we have phone level backtraces. | S | s | s |
mn017 | you'd have you'd have | D | % | % |
me011 | yeah this | D | fg | fg |
me011 | i don't think i would do this for phone level. | S | s | ng |
fe016 | um | F | fh | fh |
me011 | i think for phone level you want to use some sort of binary representation. | S | s | s |
me011 | because it'll be too dense otherwise. | S | s | df |
fe016 | okay. | F | fg | fg |
fe016 | so if you were doing that and you had this sort of companion uh thing that gets called up for phone level uh what would that look like? | Q | fh | fh |
fe016 | how would you | D | qw | qw |
me011 | i would use just an existing an existing way of doing it. | S | s | s |
mn017 | huh. | B | b | b |
mn017 | but but why not use it for phone level? | Q | qw | qw |
mn017 | it's just a matter of it's just a matter of it being bigger. | S | s | df |
mn017 | but if you have | D | s | s |
mn017 | you know barring memory limitations uh i mean this is still the | D | s | s |
me011 | it's parsing limitations. | S | s | nd |
me011 | i don't want to have this text file that you have to read in the whole thing to do something very simple for. | S | s | s |
mn017 | oh no. | S | s | ar |
mn017 | you would use it only for purposes where you actually want the phone level information i'd imagine. | S | s | nd |
fe016 | so you could have some file that configures how much information you want in your in your x.m.l. or something. | S | s | s |
me011 | right. | S | s | bk |
me011 | i mean you'd | D | % | % |
fe016 | um | F | fh | fh |
fe016 | because it does get very with | D | s | df |
me011 | i i am imagining you'd have multiple versions of this depending on the information that you want. | S | s | s |
mn017 | you | D | % | % |
fe016 | right. | S | s | aa |
me011 | um | F | fh | fh |
me011 | i'm just | D | s | s |
me011 | what i'm wondering is whether | D | s | s |
me011 | i think for word level this would be okay. | S | s | s |
fe016 | yeah. | S | s | aa |
fe016 | yeah. | S | s | aa |
me011 | for word level it's all right. | S | s | s |
fe016 | definitely. | S | s | na |
mn017 | uhhuh. | B | b | b |
me011 | for lower than word level you're talking about so much data that i just i don't know. | S | s | no |
me011 | i don't know if that | D | s | s |
fe016 | i mean we actually have | D | s | s |
fe016 | so one thing that don is doing is we're we're running for every frame you get a pitch value. | S | s | s |
me018 | lattices are big too. | S | s | s |
fe016 | and not only one pitch value but different kinds of pitch values. | S | s | e |
fe016 | depending on | D | s | e |
me011 | yeah i mean for something like that i would use p. file. | S | fg | fg |
me011 | or or any frame level stuff i would use p. file. | S | s | s |
fe016 | meaning? | Q | qw | d |
me011 | uh that's a well or something like it. | S | s | s |
me011 | it's uh icsi has a format for frame level representation of features. | S | s | s |
fe016 | okay. | S | s | bk |
me011 | um | F | fh | fh |
fe016 | that you could call that you would tie into this representation with like an i.d. | S | s | bu |
me011 | right. | S | s | aa |
me011 | right. | S | s | aa |
me011 | or or there's a there's a particular way in x.m.l. to refer to external resources. | S | s | s |
fe016 | and | F | fh | fh |
fe016 | okay. | S | s | bk |
me011 | so you would say refer to this external file. | S | s | s |
me011 | um | F | fh | fh |
me011 | so that external file wouldn't be in | D | s | s |
fe016 | so that might that might work. | S | s | na |
me018 | but what what's the advantage of doing that versus just putting it into this format? | Q | qw | rt |
me011 | more compact. | S | s | s |
me011 | which i think is is better. | S | s | ba |
me018 | uhhuh. | B | b | b |
me011 | i mean if you did it at this | D | s | s |
fe016 | i mean these are long meetings. | S | s | s |
fe016 | and with for every frame | D | s | s |
me011 | you don't want to do it with that | D | s | s |
fe016 | um | F | fh | fh |
me011 | anything at frame level you had better encode binary. | S | s | s |
me011 | or it's going to be really painful. | S | s | df |
mn017 | or you just | D | s | cs |
mn017 | i mean i like text formats. | S | s | cs |
mn017 | um | F | fh | fh |
mn017 | you can always uh g. zip them. | S | s | cs |
mn017 | and um you know decompress them on the fly if if space is really a concern. | S | fh | fh |
me018 | yeah i was | D | fg | fg |
me018 | i was thinking the advantage is that we can share this with other people. | S | s | s |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.