Sevana Oy Blog » on the fly

Sevana announces partnership with Dialogic

admin — Fri, 17 Dec 2010 11:14:58 +0000

Hi,

We are happy to announce that Sevana became Dialogic Application Partner!

Best Regards,

Sevana Oy

http://www.sevana.fi/partners.php

How can I test this voice quality measurement software?

admin — Mon, 30 Mar 2009 19:09:20 +0000

Sorry for the pause with our new posts on voice quality! This temporary information gap will be over soon and meanwhile we would like to give an answer to such questions as:

- how can I compare two audio files to test voice quality?

- how can I test my codec for voice quality it provides?

The answer to the questions about is: please use contact information on these matters from here: http://www.sevana.fi/about.php

- is there a demo version I can play with?

Yes, it’s available for download from here: http://www.sevana.fi/audio_speech_codecs_quality_analysis.php

- demo version doesn’t allow comparing files! What can I do?

Please use contact information from here: http://www.sevana.fi/about.php and you may get a fully functioning software for your testing purposes!

- are there any updates to the functionality presented in the demo version?

Yes! There is a new version that actually tells you why there is a quality loss providing quantative feedback! And… you know where to go to inquire that version

Stay tuned!

Confused on how to measure voice quality after all? Here is the answer – signal + noise + speech model + analytics = state-of-the-art approach!

admin — Tue, 24 Feb 2009 06:32:31 +0000

Today we would like to propose a yet another concept for voice quality testing and estimation. We believe that a successor for PESQ and other methods should have a generator of test signals that allows sound signal generation according to one of the sound flow models (we explain them in further posts). It can be either a particularized set of sound signals or a signal, received as an output of statistical speech model. Generator’s signal can either be saved for follow-up usage or be exposed to processing and estimation. Bank of signals stores sound data, received as a result of signals’ generator work or from some external sources (you can ask for sample speech models from http://www.sevana.fi). Accordingly, an input of estimation block is a signal of generator directly or one of the bank of signals. Test signal is the input of the synchronizer or of the device under test, which can be for example, a vocoder, compressed/decompressed audio, just an audio/voice file or a communication channel. The output signal of the device under test is an input of synchronizer as well. The synchronizer matches in time initial signal and a processed signal. The synchronized signals in chunks are input into analytical module, which determines the degree of similarity for signals and issues the quality estimation as the measure of similarity between the initial and the processed signals. Misty? The point is that a successful voice quality testing system should have:

- Test signals generator
- Noise signals generator
- Statistical speech model
- Signals syncronizer
- Analytical module

Stay connected and we describe all of these modules in our upcoming blog posts!

P.862 Perceptual Evaluation of Speech Quality does not always percept degraded speech well – what to do?

admin — Mon, 23 Feb 2009 04:08:10 +0000

PESQ stands for Perceptual Evaluation of Speech Quality, an ITU standard specifying how one should test speech/voice signals to obtain objective measurement on the quality of tested voice sound and predict the results of subjective listening tests on network and telephony systems. Relax, even this method is not perfect…

PESQ is meant to be an objective measurement method that predicts the results of subjective listening tests on telephony systems. PESQ uses a sensory model to compare the original, unprocessed signal with the degraded signal from the network or network element. The resulting quality score is similar to the subjective “Mean Opinion Score” (MOS) measured using panel tests according to ITU-T P.800. The PESQ scores are calibrated using a large database of subjective tests. The method takes into account coding distortions, errors, packet loss, delay and variable delay, and filtering in analogue network components.

And again… although being one of the most popular tools PESQ has a number of disadvantages such as demanding test signals to be speech-like because many systems are optimized for speech and respond in an unrepresentative way to non-speech signals (e.g. tones, noise, ITU-T P.50). PESQ test signal is to be set by tester and thus vendor estimations may vary from end customer estimations. The approach performs signal level equalization what theoretically is not that good because when speaking different sound volumes may have different spectrums. PESQ cannot catch significant quality loss, which occurs when the voice is equalized such that there is far less low frequency and high frequency energy when compared to the original voice file (this case is described in Microtronix article devoted to PESQ: http://www.microtronix.ca/pesq-disc.html)

Hm… what should follow this post then? What is the outcome of all previous posts? What are we talking about? You’ll learn that tomorrow. Stay tuned!.

Speech Intelligibility is not enough – let’s go for speech transmission index! The truth is out there…

admin — Sat, 21 Feb 2009 12:44:15 +0000

Today we would like to present two more methods to measure voice quality for example degraded by using a lossy codec or transmitting via mobile or fixed networks. These are methods similar in the approach and differ in implementation scheme, here they are:

STI (Speech Transmission Index). We may approximately consider speech signal as broadband signal modulated by low-frequency signal. Articulation speed determines modulation frequency. When modulation depth decreases, speech signal becomes similar to noise and its intelligibility decreases. Accordingly, intelligibility decrease can be estimated according to modulation depth decrease also. Whole speech range is divided into 7octave bands. An octave noise signal is the input. The test signal intensity distribution agrees with the distribution of speech signal intensities. The modulating signal frequencies vary from 0.5 to 12.5 Hz with one-third-octave interval (14 frequencies in all). The STI measuring method is stated in the International standard IEC 268-16.

RATSI/STIPA (Rapid Speech Transmission Index). The STI method needs a lot of measuring procedures and calculations. A simplified method was developed, which provides for measuring only in 2 bands with 5 modulation frequencies and reduces the number of measuring procedures and calculations. For good intelligibility RASTI values must be not less than 0.6.

As these are analogous methods they have common disadvantages: speech transmission index as well as rapid speech transmission index imitates speech production process by means of noise model, but to take into account the properties of speech production and hearing in such a way is far from optimum.

Anticipating why we are posting short methods description and their disadvantages? That will be discovered in the next post about C50 – factor of clearence method.