Sai Krishna Rallabandi

Am a PhD student in Language Technologies Institute at Carnegie Mellon University. I am currently working with Prof. Alan Black. I actually consider myself the CTO of Alan Black Research Lab. I will design most of the pipeline in our submissions to speech related challenges so you know whom to blame if CMU doesnt do well. Talk to me if you want to be a part or take up a significant portion from me for a particular challenge.

Most of my thought process right now is occupied with excitement for PhD at CMU.

My hyperparameters:
Email Id: srallaba@andrew.cmu.edu
Resume: Sai Krishna (Updated December, 2017)
Statement of Purpose : PhD, CMU (Thanks Shrimai, Sunaina, Pallavi, Sumit, Khyati and a thousand others!! who helped turn my rubbish into this)

Productivity Reports:
Spring 2017 Fall 2017 Q4 2017
Q1 2018

Posts with suggestions for prospective students :
Optimizing returns from coursework in LTI

Video Reviews of workflow :
Summer 2018

Publications

2018


"Submission from CMU for Blizzard Challenge 2018", Sai Krishna Rallabandi, Pallavi Baljekar, Peter Wu, Evangelia Spiliopoulou and Alan W Black, Blizzard Challenge Workshop 2018.

"Automatic Detection of Code-switching Style from Acoustics", Sai Krishna Rallabandi, Sunaynana Sitaram and Alan W Black, Third Workshop on Computational Approaches to Linguistic Code-switching, 2018

"Investigating Utterance level Representations for detecting Intent from Acoustics", Sai Krishna Rallabandi , Bhavya Karki, Carla Viega, Eric Nyberg and Alan W Black , Interspeech 2018

"An Investigation of Convolution Attention Based Models for Multilingual Speech Synthesis of Indian Languages", Pallavi Baljekar, Sai Krishna Rallabandi and Alan W Black, Interspeech 2018


2017


"On building mixed lingual speech synthesis systems", Sai Krishna Rallabandi and Alan W Black , Interspeech 2017

"Speech Synthesis for Mixed-Language Navigation Instructions", Khyati Chandu, Sai Krishna, Sunayana and Alan W Black, Interspeech 2017

"CMU's entry to Blizzard Speech Synthesis Challenge 2017", Sai Krishna Rallabandi, Pallavi Baljekar and Alan W Black

"CMU's entry to Blizzard Machine Learning Challenge 2017", Pallavi Baljekar, Sai Krishna Rallabandi and Alan W Black


2016


"Experiments with cross-lingual systems for synthesis of code-mixed text", Sunayana Sitaram, Sai Krishna Rallabandi, Shruti Rijhwani and Alan W Black, Speech Synthesis Workshop (SSW9), 2016

"IIIT Hyderabad's entry to Blizzard Challenge 2016", Sai Sirisha Rallabandi, Sai Krishna Rallabandi and Suryakanth V Gangashetty, Blizzard Challenge 2016

"Sonority rise: Aiding backoff in syllable-based speech synthesis" , Saikrishna Rallabandi, Ayushi Pandey, Saisirisha Rallabandi, Tejas Godambe, Suryakanth V Gangashetty, NCC 2016

"Investigating Signal Correlation as Continuity Metric in a Syllable Based Unit Selection Synthesis System" , Sai Sirisha Rallabandi, Sai Krishna Rallabandi, Naina Teertha, R Kumaraswamy, Suryakanth V Gangashetty, International Confence on Speech and Computer, 2016.

2015


"IIIT Hyderabad's entry to Blizzard Challenge 2015" , Sai Krishna Rallabandi, Anadaswarup Vadapalli, Sivanand Achanta and Kishore Prahallad, Blizzard Challenge 2015

"Audio Rendering of Mathematical Content", Venkatesh Potluri, SaiKrishna Rallabandi, Priyanka Srivastava, Kishore Prahallad, CSUN 2015

"Developing a unit selection voice given audio without corresponding text" , Tejas Godambe, Sai Krishna Rallabandi, Suryakanth V Gangashetty, Ashraf Alkhairy, Afshan Jafri, EURASIP Journal on Audio, Speech, and Music Processing

"Learning continuous representation of text for phone duration modeling in statistical parametric speech synthesis", Sai Krishna Rallabandi, Sai Sirisha Rallabandi, Padmini Bandi, Suryakanth V Gangashetty, ASRU 2015

2014


"Significance of Paralinguistic Cues in the Synthesis of Mathematical Equations" , Venkatesh Potluri, Sai Krishna Rallabandi, Priyanka Srivastava and Kishore Prahallad, International Conference on Natural Language Processing, 2014

Some previous projects

Google Summer of Code 2015 - Forced Alignment of Speech and Text in Long Videos


IIIT Hyderabad's System for Blizzard Challenges

2015 - Synthesis in Indian Languages 2016 - Synthesis of Children's stories

Interested in Synthesis?

Learn Check the evolution

Now that I am associated with CMU, I suppose I should change this some time soon.


Current Projects

BILINGUAL SPEECH SYNTHESIS


These are a series of projects that we are working on with an aim to build voices capable of rendering codemixed content.

Project Portal


  • VOICE CONVERSION


    This is a project I took up, and can be seen as a example of a project that surfaced following from the famous 20% rule by Google. I was working on a lame version of voice conversion and looking at its success we(Alan and me) thought we could take it up more seriously. Currently, I am working on a submission to Voice Conversion Challenge.
    You can track the updates in the project on ResearchGate.

    Project Portal

  • LANGUAGE DETECTION IN FOUND DATA


    It is easy to today find a large amount of data via lectures, youtube videos, etc. But, obtaining reliable transcriptions for the audio still remains a hurdle. While manually transcribing the audio data is a time consuming and tedious process, obtaining automatic transcriptions using commercial large vocabulary continuous speech recognition systems proves expensive, and such systems are also not available for many languages. We investigate a specific framework to use such found data and build synthetic voices seemlessly and effortlessly. Project Portal



    Interested in any of the projects and want to be a part? Email Me