Output format of posterior htk file

[Maria Cabello]

Hello,
I've processed my posterior .rec file with the following command;

./HList -r $rec_file

I obtain a matrix with 3 states per fonetic unit and per frame. Please, could you help to know if matrix format is:
frame1        unit1(state1) unit1(state2) unit1(state3) unit2(state1) unit2(state2) unit2(state3) ...

or:
frame1        unit1(state1) unit2(state1) unit1(state2) unit2(state2) unit1(state3) unit2(state3)...

So, each row is it shorted by fonetic unit or state?

Thanks in advance.

[Nay San]

Hi, I also have the exact same question. I am looking to sum up the posteriors to 46 classes from the 138 in the file output by the CZ recogniser.

Hope to hear back!

[Petr Schwarz]

Hi, it is sorted by phonetic unit.

unit1_state1 unit1_state2 unit1_state3 unit2_state1 ...

The final 3 outputs are for parts of audio with no phonetic labeling or silence (marked as ‘other’).

 

Petr

--

---
You received this message because you are subscribed to the Google Groups "phnrec" group.
To unsubscribe from this group and stop receiving emails from it, send an email to phnrec+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/phnrec/0d8c143b-a553-44be-bd15-482e8d3036cen%40googlegroups.com.