the dae-ly breeze

Wow - still feels surreal, but its done!!! I had my thesis defense this afternoon and i passed! 3 years of work is done. Il write more as soon as it sinks in … but im in cloud nine hehe feels sooo great! Finally - its finished - and i didnt quit - i kept pushing til its done.

I will post my acknowledgment next time

The Department of Electrical & Electronics Engineering
is pleased to announce the
Oral Defense of the Master’s Thesis of

Ms. Amadea Paula Q. Unisa

Entitled: “Real-time Implementation of Low Bit Rate Wideband Speech Coders on ADSP-21065L”

for the degree of MS in Electrical Engineering

Date:   October 14, 2008
Time:  4 pm
Venue:  PLDT Multimedia Lecture Hall

Panel Members:
Thesis Adviser:  Rowena Cristina L. Guevara, Ph.D.
Chairman: Joel Joseph S. Marciano, Jr., Ph.D.
Members: Luis G. Sison, Ph.D.
Franz A. de Leon, MS EE
Emerson C. Tan, MS EE

This defense is open to the public.

ABSTRACT

In this thesis, real-time Wideband Sinusoidal (WS) and Wideband Linear Prediction (WLP) speech coders were implemented using the ADSP-21065L EZ-Kit LiteTM Evaluation Board. These speech coders are based on the parametric speech models which are extensions of existing sinusoidal and LP models for narrowband speech. The parameters extracted from the input speech are quantized using vector and scalar quantization methods.

A real-time implementation was achieved by programming all the subroutines of the speech coders in assembly language. The average complexities of the WS analysis and synthesis blocks are 1.803 MIPS and 2.95 MIPS.  The WS speech coder requires 0.785 kwords of program memory and 5.446 kwords of data memory. For the WLP speech coder, the analysis and synthesis blocks obtained average cycle counts of 28.179 MIPS and 11.249 MIPS. The WLP speech coder occupies 5.98 kwords of program memory and 9.355 kwords of data memory.

Two WS configurations and all of the WLP configurations achieved Mean Opinion Score (MOS) of greater than 3, indicating that the synthesized speech has noticeable but acceptable impairments. All the WS and WLP configurations achieve Diagnotic Rhyme Test (DRT) scores of greater than 80, thus the listener could discriminate between the rhyming pairs of consonant-vowel-consonant (C-V-C) words.

*****

Hope to see you there

- dae

September 10, 2008 marks the most amazing day in my life’s Speech Signal Processing history! And im not exaggerating… Fed and myself attended the talk of Dr. Alex Waibel, a professor from CMU (Pittsburgh) and University of Karlsruhe (Germany) - who’s a genius in speech signal processing!

Fed and I couldnt stop gushing about Dr. Waibel’s demos! They are just WOW! Seriously, im running out of words to describe them! I think the reason Fed and I are overwhelmed this much is because what he presented are something we’ve been trying to do in the UP DSP lab (and yes, we’re inching along, but hey, we’ll get there in time!) But to see them running right before our eyes and see that its very much possible - simply awesome! You know how its like when you watch movies or shows like CSI or Alias and they show you some techie stuff and you’re like ‘oh yeah, its a movie - science fiction!’ but this one is real - and fed and i saw it working real-time!

The following are the demos that we witnessed:
1. The direction detector that shows a video wherein people are in a meeting and there are arrows pointing to where the person is directing his attention to or who he’s conversing with - so you can see arrows following the movements of a person’s head.
2. The face-tracking system that identifies the eyes of the person (so Dr. Waibel presented a video where everytime a door opens and someone goes inside, you can see two circles where the person’s eyes are)
3. Directional sound system consisting of ultrasonic speakers that convolve with the air. Basically, how it works is that if the speaker is directed to you then only you can hear the sound. If its not directed to you, then you wont be able to hear a thing.

But the demo that fed and i can relate the most is the language translator. This project is a part of the CHIL (Computers in the Human Interaction Loop) project that aims to aid human to human interaction by breaking the language divide. Dr. Waibel showed us the ‘Long Tail of Language’ - a graph depicting the number of speakers of a certain language. On top of that list is Mandarin, English, Spanish with hundreds of millions of speakers, and then there’s a plunge on the number of speakers of the other languages. And there are 6000 languages in the world! Anyway, this project, like i mentioned aids human to human interaction by breaking the language barriers. And how this thing works may sound simple, but being a speech processing person myself, let me tell you, its not as easy as it sound

There are three legs to this system: (1) Speech recognition - where the system identifies what the speaker has said and outputs it in text form; (2) Translator - translates the text from one language to another; and (3) Speech synthesis - the translated text is converted to sound output. What’s amazing is that Dr. Waibel had a demo of an english-to-filipino translator However, the synthesizer he used is Spanish so it sounded like a Spanish guy speaking Filipino words (close enough though!)

And so there comes an opportunity for us. Dr. Waibel is really interested to continue that project for the Filipino language and for the other languages in the Philippines - both on the recognition and the synthesis end. They already have the translator (one language to Filipino) but it still needs improvement since no one at their lab speaks Filipino. And that’s where we come in. I am proud to say that our lab is the pioneer in our country to do speech signal processing and we have the ability to help Dr. Waibel and the people at InterACT (International Center for Advanced Communication Technologies) to get this going (at least for the Filipino language). There already have been projects and theses on Filipino speech recognition and synthesis done at our lab. (And Dr. Waibel said that he’s impressed with our efforts in the lab, and he’s quite amazed that these researches are being done in our country. nice!)

So yes, im excited for all these opportunities to collaborate with great minds! It will be great for the lab to be expanding our horizon, so to speak, and to be working with people who are very good at this field, at least this way, we can leap from where are now to where we want to be Like i said, we’re inching along, but we’ll get there, with all the help we can possibly get!

all in all, this experience is truly awesome and one-of-a-kind! oh wait, did i mention, Dr. Waibel’s hobby is flying helicopters?!

Here’s our picture with Dr. Alex Waibel

had a photoshoot this 25th of august and it was lots of fun my thanks goes to the photographers who still decided to go through the shoot even if i was the only ‘model’ who made it: thet, marc, buddy, tintin, vic, eric and milton; to the hairstylist who i have worked with in my first-ever photoshoot krissy; and to the makeup artist who works magic (hehe!) kaye

the multiply sites of the photographers/makeup artist/hairstylist are:
marc and thet: http://marcdrizrosales.multiply.com
buddy and tin: http://buddyilao.multiply.com/
http://tintoot.multiply.com
http://bugnotandkulot.tintoot.com
vic: http://pogita77.multiply.com/
eric: (to follow)
milton: http://miltonching.multiply.com/
krissy: http://krissymira.multiply.com
kaye: (to follow)

looking forward to having more photos to be added to my collection

here’s my favorite pic (more in amadeapaula.multiply.com)

Had our DSPiktyuran moment (4pm, July 9, 2008). So who are in this photo:

1st row (L-R): Dae (that’s me!), Bogs, Leslie, Ma’am Gev, Sir Franz, Paulo, Elaine, Liz, Fed, Mike, Lito, and Noel

2nd row (L-R): Ted, Keone, Kel, Joseph, Manu, and Oliver

Here’s another one of me with my very lovely Vicky I dont really work like this.. Just for the photo shoot

And by the way, the big black box in the background is our own “whisper room” Very interesting - you can scream and scream inside the room, but to the people outside the room, your voice seems like a whisper…

So yes, that’s our lab and the bright people who are part of it

Dae

I have encountered the ADSP linker error [Error li1040] more times than I could remember while working on my thesis. This error message goes something like

[Error li1040] “.\EZLAB_21065L_debugger.ldf”:97 Out of memory in output section ’seg_pmco’ in processor ‘p0′
Total of 0×88e word(s) were not mapped.

For those of you who have already programmed a DSP, this is no surprise. The ADSP 21065L’s on-chip SRAM only has 544 Kbits that can be used to store both instructions and data. This on-chip memory is partitioned to 2 blocks: Block0 has 288 Kbits (6k x 48) and Block1 has 256 Kbits (8k x 32). But for the rest, the reaction would be something like “What can I do with 544 Kbits of memory?!”. (Hehe, a LOT actually!) But I remember a software developer once told me, “Memory, to me, is an infinite resource until the SysAd  gets mad at me”. Makes sense… but not to a DSP programmer.

At first, I was only having trouble with data memory - I had a handful of arrays in my code and I have codebooks (that are used in the vector quantization). There’s an easy solution to this and that is to store the data in the external memory. In my implementation, memory is organized such that I can store data in the following address: START(0×03000000) END(0×030ffeff). The trade-off, though, is it takes more cycles to access data in external memory (It only takes a single cycle if data is stored in the on-chip memory). So, if the number of processing cycles is a concern (as it is to real-time systems) a careful selection of which data could be stored in external memory should be done.

But what if the problem is in program memory? The error message shown above is one of the error messages I have encountered  The segment seg_pmco, defined in the LDF file, is where my instructions are stored. And based on my experience, running out of program memory is more difficult than being out of data memory. Im still experiencing it actually. This is the case where code optimization comes in. My entire code is in assembly, which in itself is already optimized (compared to C or C++).

Another way to solve the problem in program memory is to implement modularity. I do that but I missed some bits of the code that can be turned into subroutines. So I had  to go back to my code (all of it!) and checked for some more optimization opportunities and i was able to save a few more program memory spaces. Right now, all the most commonly used functions such as filtering, inner product computation, math functions (square root computation, division subroutine, acos computation, log computations, etc), scalar  and vector quantizations, etc are written as subroutines such as if they are needed anywhere in the program, it only takes a function call. However, it still didnt solve my problem…

By the way, with ‘entire code’, I meant encoder + decoder. So now, the way to solve the problem is to implement the encoder on one board, and the decoder on another I’m sure there will still be problems that will arise (such as bit syncing, etc) but I’m very excited to implement it

Wish me luck!

Hey everyone,

We have extended the survey deadline to July 7, 2008 11:59 pm. This gives you more time to respond to the survey (without hurrying up) and get the chance to have a copy of excerpts of the following books: E-business made easy, Jump Start, Go Negosyo and Go Negosyo Celebrity Edition. (See previous post for more details)

Looking forward to your responses and thank you very much!

Also, we shot some videos to entice (hehe!) you to fill up the survey Here’s the link: http://www.youtube.com/watch?v=7K7VsltzdII

Again, thank you so much for your help

Dae Unisa (for the BGN 2.0 Development Team)

So I am now 24 years old and first of all I would like to say thank you to everyone who greeted me That means so much. I know everyone gets to have this day once a year but I am still one of those people who believe that a birthday is a very special day and just the fact that you’re still alive is reason enough to celebrate!

I originally thought of celebrating it on my own - well, i had a few reasons to be a bit sad - but i realized i have more reasons to be happy! I have my family and they’re all healthy and happy; I have my boyfriend who’s everything i want and everything i need; I have my friends who are always there for me; so what more can i ask for? So I was ‘convinced’ (hehe :P) to take my brothers out So the three of us went out,  had ‘Wham’ burgers, got some Starbucks goodness (yummy!), watched ‘Wanted’, attended a jollibee kiddie party, went to the supermarket and then mass! yup it was a long day but a very fun one

So for this year I only have 3 wishes: (1) My family and my man to stay healthy; (2) Finish my thesis; and (3) get a US Visa (So please pray for me too!)

Well, thats how my birthday went. Very simple but full of fun and the best thing about it is that i spent it with the people i love - whether in person or online

As promised, I am posting here the link to the BGN 2.0 Survey that we have created I hope you guys can find the time to fill it up as it only takes 12.5 mins And theres more, if you finish the survey before July 3, you get the chance to have a copy of the e-books (please read the letter below). Also, your responses will be a huge hep to us to come up with a better ‘web movement’ known as Brain Gain Network

Maraming Salamat!

- dae

I have been working with the Brain Gain Network (www.bgn.org) for 2 years already - and I must say that it opened my eyes to so many opportunities for success and growth for Filipinos everywhere. I get to work with people like Paco Sandejas (Managing Partner of Narra VC), Dickie Gonzales (CEO of Philippine Emerging Startup Open), Bill Luz (Executive Vice President of Ayala Foundation), and the likes. I even met Dado Banatao! I feel blessed because it’s not a chance every 23-year old could get

For so many months, our team (Paco, JJ, Rhea and I) have been conceptualizing on how we want the next version of the BGN website would be - admittedly the current website is not very user-friendly (as most of our users have commented). But aside from that, we still need to know what the users think, because we want the users to be part of this whole process from beginning to end And with that said, we are creating a survey that aims to gather information from more BGN users to help us improve the website - in terms of UI, content, services, etc. I’ll post the link here as soon as it’s ready and available I hope you take 10 minutes to respond to the survey

I am very excited for this project - as we are getting very serious about it, and just like Rhea described it, we’re “intense” right now There are so many things we want to get done - i cant reveal the details yet And it wont be ‘yet-another-social-network’

I tell you, its going to be great!

Will post more soon