Speech:Spring 2012 Brandon McLaughlin Log

From Openitware
Jump to: navigation, search


Week Ending Feb 6th, 2012

Task

My tasks for the week are to install VirtualBox, OpenSUSE, and Sphinx and familiarize myself with them. Along with keeping up to date with everybody's logs. Also familiarize myself with what we went over in class on how speech recognition works.

Results

Feb, 2: Used Aaron and Brice's tutorial for installing VirtualBox and OpenSUSE, VB/SUSE/Sphinx Install Tutorial.

Feb, 3: Working on figuring out how to install Sphinx 3 and get it to work. Read everybody else's logs.

Feb, 5: Talked with Johnny on Saturday and worked together to see if we can figure out how to install Sphinx onto OpenSUSE. Worked again on Sunday with Johnny and picked up where we left off with installing Sphinx. Installed JDK, ANT, and SVN through the use of Yast. Still currently getting comfortable with the OpenSUSE software and the Linux CLI.

Feb, 6: Read everybody else's logs for the remainder of the weekend. Still can not figure out Sphinx 3, working at it as I make this post.

Plan

Try to get Sphinx installed and running on OpenSUSE so we can progress further with the project.

Concerns

Can not seem to get Sphinx to install on OpenSUSE.

Note from Aaron J.

How did you install Sphinx 3?
Aaron, I was not able to install Sphinx yet, I meant to type installing Sphinx 3. I do think that you need to install it terminally and that would mean to download the .tar file. I am not 100% sure about this since I do not work with Linux very often at all. I just read that online.


Week Ending Feb 13th, 2012

Task

Get in touch with Mike H. to set up a Skype meeting call to talk about the project and set everyone up with roles. Get in touch with James to see what he did on the project and ask questions for what we should continue doing to catch up on what needs to be done. Work with Traubadix to get a better understanding of its directories and everything else. Understanding PERL scripts.

Results

Thursday: Emailed Mike H. to get his Skype username, Talked with Johnny on plans for the weekend.

Friday: Got Mike H. Skype username and have set up a time for a meeting on Sunday to get him up to date with the project. Read a little bit on PERL scripts to get a small understanding on what it is. Read everyone else's logs.

Sunday: Skype Meeting held with Johnny and Mike H. Installed UNHM VPN along with Putty so we can navigate around the Sphinx servers and databases. As of right now I can not log into any of the servers with my UNHM credentials. Read more about PERL.

Monday: Read everyone else's logs, seems like a group is still stuck on how to install Sphinx 3. Getting ready for class tomorrow.

Plan

The next week we plan on getting a better understanding of the data set definitions, transcripts, and .wav files so that we know how to deal with them in conjunction with Sphinx. Install UNHM VPN along with Putty to get onto the servers.

Concerns

Thursday: Trying to get Mike H., Johnny, and I all on Skype together. Mike H. has not yet made a wiki so both Johnny and I are unsure of how far he is in this project.

Sunday: Can not log into any of the servers with my UNHM credentials, might be caused by someone working on the machines.


Week Ending Feb 20th, 2012

Task
  • Work on the Proposal.
  • Work on Perl scripts to neaten up the transcripts.
  • Converting the .sph files to .wav files.
Results

Tuesday:

  • Worked on the proposal.
  • Discussed roles on Skype with Johnny and Mike H. Caught Mike H. up on what we went over in class today that was extremely helpful.

Wednesday:

  • Read everyone else's logs, there was not exactly that much to read, but it is good to stay up to date with everyone else.

Thursday:

  • Held a Skype meeting with Johnny and Mike to start the proposal and get a good opening to it.
  • Got a good chunk of what we hope is a good start to the proposal.

Sunda:

  • Held a Skype meeting with Johnny and Mike H.
  • Finished what we hope a a good rough draft of the proposal for our parts, did not assign roles yet, will do hopefully tomorrow (Monday) night.
Plan
  • Skype meeting on Thursday to start the proposal.
  • Have a much better understanding of what needs to be done
  • Skype meeting on Sunday to finish up the proposal.
Concerns

Week Ending Feb 27th, 2012

Task
  • Work on the Proposal.
Results

Tuesday:

  • Discussed the proposal with Mike H. on Skype.
  • Mike H will be making the directories that Professor Jonas asked for our group to make.
  • I will be doing the data part of the project, until Mike is ready to join back up with me.

Thursday:

  • Discussed the final roles with Mike H. on Skype.
  • Talked with Johnny about his role for the Speech Project on Skype.
  • Read everybody else's logs.

Friday:

  • Worked on the Proposal.
  • Will try to contact Mike H. on Skype on Sunday to get an update on how he is doing.
  • Will continue working on the Proposal on Sunday.
  • Also, a little late but I did sign up on the Google Group this past Monday night, just forgot to post that I did.

Sunday:

  • Held Skype meeting with Johnny and Mike H.
  • Finished the proposal with time lines.
  • Noticed that not all the sections in the proposal are finished.
  • Read everyone else's logs.
Plan
  • Write the 4-5 paragraphs on the proposal.
  • Skype meeting with Mike H. at 9:30pm on Thursday.
  • Start messing around with the scripts on Caesar to get the .sph files into .wav files.
Concerns
  • The Proposal only has one more day to be worked on and it looks like everyone is waiting till the last minute to finish their sections... Not exactly the best idea.

Week Ending March 5th, 2012

Task

Tuesday:

  • Research SOX command
  • Look over the Proposal
Results

Tuesday:

  • Skype meeting with Johnny
  • Did some research on the SOX command
  • Looked over the Proposal

Thursday:

  • Skype meeting with both Johnny and Mike.
  • Talked with Mike about what directories need to be made and what files need to be transferred over into those new directories.
  • Start researching how SOX will work with PERL scripts to start working on the transcripts and wav files.

Sunday:

  • Skype meeting with Johnny and waiting for Mike to get on Skype.
  • Waiting to catch up with Mike on what his outcomes are for this week.
  • Caught up with Mike on Skype on everything that he has been working on.
  • Will start messing around a lot with the project tomorrow (Monday).
  • Read everyone's logs.

Monday:

  • Skype meeting again with Mike.
  • Laid out our strategy for the process of how we will both be able to help each other out as we progress.
  • Read everyone's logs once again.
Plan

Plan Skype meetings with both Johnny and Mike to stay up-to-date with what they are both working on.

Research the SOX command for writing scripts.

Start writing scripts on Monday.

Thursday Skype meeting with Johnny and Mike.

Friday/Saturday Skype meeting with both Johnny and Mike.

Start working real hard on project Monday.

Skype meeting with mike Monday night at 9:30pm.

Concerns

None at the moment

Week Ending March 12th, 2012

Task
  • Converting the sphere files to wave files and move them to the correct folder in the /mnt/main/corpus/dist/media/data folder.
  • Cleaning the transcript files.
  • Document how and why Mike and I created the directories for the /mnt/main/corpus directories.
Results

Wednesday:

  • Researched how to write the PERL script for cleaning the transcripts.
  • Held Skype meeting with Mike to plan the week ahead.

Sunday:

  • Skype meeting with Mike.
  • Read everyone else's logs.
  • Waiting for Mike to finish the document so I can review it.
Plan
  • Create a script to clean the transcripts.
  • Create a script to convert the sphere files to wave files.
    • Scrips will be made to let us do more than one task at a time.
  • Skype meeting with Mike on Saturday/Sunday to catch up with each other.
Concerns
  • Writing a script to clean the transcripts.
  • Writing a script to convert the sphere files to wave files.

Week Ending March 19th, 2012

Task
  • Catch up with Mike.
  • Look over Mike's documentation for creating the directories with the ownership's.
Results

Thursday:

  • Read everyone else's logs.

Friday:

  • Skype meeting with Mike.
  • Worked on the documentation for creating the directories.

Sunday:

  • Read everyone's logs.

Monday:

  • Skype meeting with Mike.
  • Discussed what needs to be done for the remainder of the schedule.
Plan
  • Catch up with Mike as early as possible.
  • Work on the Documentation.
  • Skype meeting Sunday/Monday night.
Concerns
  • Can't remember that command that was used to reroute the directory /mnt/main/corpus/dist/Switchboard from the old /media/data/Switchboard.

Week Ending March 26th, 2012

Task
  • Converting the sphere files to wave files and move them to the correct folder in the /mnt/main/corpus/dist/media/data folder.
  • Cleaning the transcript files.
  • Read logs.
Results
  • Got a Perl script to function correctly for converting the files.
#!/bin/bash
cd /mnt/main/home/sp12/baf76/Switchboard/disk2/swb1
for file in *.sph ; do
    sox "$file" "$(basename "$file" .sph).wav"
done
  • Got a command to almost work correctly for moving the converted files to the correct folder.
find /mnt/main/home/sp12/baf76/disk2/swb1 -iname "*.wav" -exec mv ' ' /mnt/main/home/sp12/baf76/disk2_wav/ \;

Thursday:

  • Read Everyone's logs.

Saturday:

  • Checked the transcripts out.
  • Looked at the Perl script a little to figure out how it works and what it does.
  • Will finish up the transcripts Sunday and Monday with Mike.
  • Caught up with reading everyone else's logs.

Sunday:

  • Talked with Mike to split up the work.
  • Mike is doing the transcriptions and I am doing the converting.
  • Skype meeting tomorrow night to see how both of us are doing on our progression.

Monday:

  • Had a meeting with Mike.
  • Been working on converting the sph files to wav files.
  • Got a Perl script to work, just need to finish the command to move the converted files to the correct directory.
  • Mike said he was making good progress on cleaning up the transcripts.
Plan
  • Talk with Mike to stay on track.
  • Plan on converting the files this weekend.
  • Plan on Cleaning up the transcripts this weekend.
  • Split up the work between Mike and I.
Concerns
  • Getting the command to move the converted wav files to the correct directory to function correctly.

Week Ending April 2nd, 2012

Task
  • Converting the utterances in the sphere files to wave files and move them to the correct folder in the /mnt/main/corpus/dist/media/data folder.
  • Cleaning the transcript files.
  • Read logs.
Results

Thursday:

  • Looked at what Brian had made for a script to grab utterances out of the sphere files.
  • The script is a little confusing, but I hope to get a better understanding of it tomorrow.
  • Found a better script for cleaning up the transcripts since it seems to me that Mike did not really look to hard for something that was already created, so he does not have to start from scratch.

Friday:

  • Got caught up with Mike.
  • Skype meeting Sunday.
  • Read logs.

Saturday/Sunday Morning:

  • Looked at Brian's script from last spring and tried to figure out why it will not run.
  • Messed around with the script and I am completely lost with why it does not work.
  • I do understand how the script works for the most part, but I do not know why the script will not work when I give it a filename.

Sunday:

  • Got the a sox command to work for creating wav files out of time lengths for utterances.
sox sw02001.sph 02001.wav trim 64 7
  • Now working on trying to see where this sox command can fit into Brian's perl script so it will work.
  • If that seems to hard to do for now I will try and create a script that does just that one conversion rather than everything at once.

Monday:

  • Still messing around with a couple of Brians scripts.
  • Still lost on how to go about making his scripts work.
  • Talked with Mike and he too seems very confused on this as well.
  • When trying to run the converter.pl script I know it is not complete, but the error messages do not make sense to me. They just tell me that there is a syntax error on a certain line and that is all.
  • Was hard to catch up since Mike did not contact me at all this weekend till today, so it made it hard to even progress on this problem since I need to know what he is working on to be able to follow up.
  • Will have to see what tomorrow in class brings for any sort of help, will talk to James also.
Plan
  • Talk with Mike to stay on track.
  • Convert the files this weekend.
  • Clean up the transcripts.
Concerns
  • Both seem a little skeptical about how we should go about with this problem of grabbing utterances out of one .sph file.
  • I am not as skeptical as I was earlier in the week, but I am still a little un-comfortable with this part of the project.
  • As of now trying to get the sox command into a perl script and fully function.

Week Ending April 9th, 2012

Task
  • Finish the script for cleaning and converting the .sph files.
Results

Wednesday:

  • Finished creating the symbolic links from the 23 disks into the flat folder using:
ln -s ../disk1/swb1/*.sph -t .
  • Found out that disk8 has all uppercase letters for the directories and files in the disk8 directory. Just a note in case that is a problem for any future purpose. I had to change the swb1 from lowercase to uppercase to get around that problem.

Thursday:

  • Updated the Information page with the symbolic link command that was used for making the flat directory with all the .sph files from all the disks.
  • Worked on the script for converting and cleaning the .sph files to wav files.
  • Made some good progress, will pick it back up tomorrow to hopefully get it finished and if not make a lot more progression on it.

Friday:

  • Worked on the script, still progressing further.
  • Using Brian's scripts as a guide to learn how it should somewhat resemble.
  • Still getting errors.

Saturday:

  • Worked on script.
  • kept getting errors and thought I was getting somewhere.
  • Sent script to Professor Jonas to review and he replied with valuable feedback.

Sunday:

  • Working on script again using Professor Jonas's feedback.
  • Looked at genTrans.pl and got it to work. Still unsure that if this file works and does what is needed to do, why i was tasked with creating a new script that does this...
  • Will tweak the transcript tomorrow to make it delete the "[LAUGHTER]" since that is not a real word to say in a sentence.

Monday:

  • Asking Johnny Mom for some help on how to get the words in between the brackets deleted.
  • We are making some progression, but it does not seem like it is working the way that we need it to.
  • At 11:21PM I figured out how to get rid of the words in the brackets.
$message =~ s/\[.*\]//g;
  • The sed command that was being used before was
$message =~ s/\[noise]//g;
  • Now the genTrans.pl file works as needed, there might be a few kinks that need to be worked out though, but it works to what my task was.
  • The file also needs to be changed the path as to where it should look for the .sph files. I wanted to check with Professor Jonas first to know exactly where that should be.
Plan
  • Finish this script as soon as possible.
Concerns

Tbd

Week Ending April 16th, 2012

Task
  • Group meetings
  • Work on getting the train to work on Miraculix
Results

Wednesday:

  • Updated the info page.
  • Still need to finish updating the info page with talking about the genTrans.pl perl script.

Thursday:

  • Planned to have a group meeting with Aaron G. and Ted Friday at 10am.
  • Still can not seem to get a response from Ted to see if he will make the meeting or not.

Sunday:

  • Held a Skype meeting with Aaron Green at 3pm today.
  • Found out that the sed command that I thought I fixed in the genTrans.pl script actually does not work.
  • It just deletes everything was one bracket until the last bracket that it finds in that line.
  • Not really getting too far with the mini train this week.
  • Can not seem to get in touch with Ted, he does not return any emails nor does he go on Skype.

Monday:

  • Worked on the info page.
  • Made a link to the GenTrans perl script.
  • Talked a little about what the script does and what my errors were with it so that they are not repeated.
Plan
  • Get a train to work and function correctly
Concerns

Tbd

Week Ending April 23rd, 2012

Task

Have group meetings and stay on track.

Results

Wednesday:

  • Skype meeting with team to go over what needs to be done.

Friday:

  • Skype meeting with team again.
  • Moved the correct files over.

Saturday:

  • Worked with Johnny to see if we could get much progress done on the mini train.
  • Started following the speech install guidelines.
  • Contacted teammates to catch up on everything they are doing.

Sunday:

  • Got the team log up-to-date.
  • Working on following the install guide since we all realized we had to follow these steps to even begin the mini train.
  • Asked my teammates to at least start the install on Miraculix and apparently nobody can follow the steps one-by-one. Will have to do it by myself tomorrow. Along with try and get the mini train to run.

Monday:

  • Worked on Miraculix
  • Tried following the install guide to install the files, but Miraculix has something wrong with root permissions it seems.
  • Ran everything fine while working with Johnny this past week, and tried to redo the exact same thing onto Miraculix, but every command that is run gets a permission denied statement.
  • Now working with Johnny and Aaron Green to try and see if we can get things to work correctly on Johnny's machine.
Plan
  • Get a mini train to run correctly.
Concerns
  • Getting the mini train to run correctly.

Week Ending May 1st, 2012

Task
  • Work with the modeling group to get the mini train to run successfully.
Results

Friday:

  • Checked out log noticed that their was an error when running the RunAll.pl script.
  • Researching the error to try and find a solution to the problem.

Saturday:

  • Read everyone's logs, everyone seems to be on track with their things with a few exceptions.
  • Plan on working on the mini train with various modeling group members Sunday and Monday.

Sunday:

  • Read the group report, everything is looking good. Aaron Green is doing an excellent job with keeping everyone up-to-date with the report.
  • Worked with Johnny trying to figure more out on the RunAll.pl script. So far nothing.

Monday:

  • Read the group logs and the report.
  • Still having much trouble with the RunAll.pl script and we can not seem to figure out how to fix this.


Plan
  • Work over the weekend with various group members on this mini train problem.
Concerns

Tbd

Week Ending May 8th, 2012

Task
  • Finish up the Report
  • Work on the Modeling Group problems to get the mini train to run.
Results

Friday:

  • Tried researching the problems/errors that everyone was getting when running the scripts.

Saturday:

  • Asked Mike H on Tuesday if he could write his portion of the report before the weekend so I could verify it all and then write mine. He still has yet to write anything.
  • Going to start my portion of the report.
  • Read everyone's logs.

Sunday:

  • Emailed Mike H. about the report to add more substance to it.
  • Worked on the report and did some editing to it.
  • Still having trouble finding soloutions to the problems on the verify_all.pl script.
Plan
  • As stated in Tasks
Concerns
  • Mini Train problems back to back.

Week Ending May 15th, 2012

Task
  • Finalize the Report
Results

Saturday:

  • Work on Report
  • Communicate with other members to see how the decode progress is coming along.
Plan
  • Decode
  • Finish Report
Concerns
  • Getting a full decode to run