(205) previous ~ index ~ next

---

To: tdt-distrib@unagi.cis.upenn.edu
From: David Graff <graff@unagi.cis.upenn.edu>
Subject: Patch for TDT2 ASR data
Date: Tue, 05 Oct 1999 12:16:04 EDT

---

Folks,

I have placed a file in our "members_only" ftp directory that contains fixed
versions of 126 *.as1 files and their boundary tables -- in the previous
release, these files had one or more cases of a "<W>" tag that had no word
token; the patched version corrects this problem.

(NIST had sent me a complete new set of TDT2 as1 files over the weekend, and I
confirmed that only these 126 files showed any diffence relative to the data
they had sent me last spring. I also confirmed that the missing word problem
was completely fixed.)

Also included in this patch are new versions of the following Mandarin (and
machine-translated) as0 files:

as0/19980302_0700_0800_VOA_MAN.as0
as0/19980302_0900_1000_VOA_MAN.as0
as0_bnd/19980302_0700_0800_VOA_MAN.as0_bnd
as0_bnd/19980302_0900_1000_VOA_MAN.as0_bnd
mtas0/19980302_0700_0800_VOA_MAN.mtas0
mtas0/19980302_0900_1000_VOA_MAN.mtas0
mtas0_bnd/19980302_0700_0800_VOA_MAN.mtas0_bnd
mtas0_bnd/19980302_0900_1000_VOA_MAN.mtas0_bnd

These are all the files that had been affected by the problem that Sreenivasa
Sista reported on Sep. 28, in which there was a mix-up of file names relative
to the corresponding "tkn" and "mttkn" data.

[for ftp instructions, contact Dave Graff ]

The compressed tar file is 9447007 bytes; you should "cd" to the base
directory where you unpacked the latest TDT2 cdrom release, and unpack this
tar file there. This will replace the faulty files (assuming you have write
permission on the directories and files involved).

(I've made sure that any combination of `uncompress' or `gunzip' with
"standard" or gnu `tar' will work to unpack this file.)

Please let me know if you encounter any problems.

Dave Graff

---

(205) previous ~ index ~ next

Last updated Tue Oct 19 10:10:08 1999