- Lijo Jacob
- March 2, 2021
Simulating Real World Audio for Voice Command Engine
Despite the ubiquitous presence of voice assistants in our homes and workplaces, the technological intricacies of how automatic speech recognition works continue to amaze us.
One of the most crucial aspects that determines the accuracy of a good voice assistant is based on testing it in generalized real-world environments, which unfortunately is very difficult to do. Hence, engineers implement the test infrastructure to simulate these environments. In this blog, we will see how the Audio AI team at Ignitarium tests its deep learning models on real-world simulations.
Dataset Preparation and Collection
A real-world audio signal is very challenging to create in a simulated environment. In order to do so, two sets of audio are needed namely: noisy signals (background noise) and specific audio keywords of interest (KOI).
Noisy signals are collected in different formats from thousands of sources like work locations, industrial shop floors, bus stations, birds chirping, computer-generated noises, etc.
KOIs are collected through Ignitarium’s cloud-based data collection component, AudioFarm, which is part of our comprehensive SeptraTM Audio ML platform. Once the two sets of audios are collected, they are made into a common format (“.wav”) at a fixed sampling rate of 16 KHz as it aids in different data manipulation operations.
Data Labeling
Data Labeling is an important part of any machine learning model. A good model requires quality data, which in turn requires meticulous labelling especially in the case of audio. To ensure the quality of real-world simulated audio, which for simplicity we will be calling as ‘long audio’, our KOIs need to be properly labelled. The KOIs collected usually have a lot of unwanted noise, so the audio files need to be listened to and the starting and endpoints of the speech region marked. These points are then saved as json files, which in turn are referenced, during long audio generation.

Fig 1 Audio Labeling Tool
Long Audio Creation
For creating a long audio file, a random noise sample is picked and a 10 second duration clip is cropped. Or if the noise sample has duration less than 10 seconds, it is repeated to meet the criteria. The 10s noise sample is randomly set to a minimum loudness value and maximum loudness value (in dB) in a config file.

Fig. 2: Random Noise Sample
Next, a random KOI is selected based on a set of preconditions. The preconditions are kept in a configuration file for ease of accessibility. The preconditions include minimum and maximum sample duration and loudness, maximum allowable signal to noise ratio (SNR) etc. Based on these conditions, the KOI is embedded randomly anywhere within the noise signal respecting the max allowable SNR and max allowable noise dB.

Fig. 3: KOI embedded noise signal
The next KOI will be embedded by maintaining a random distance between minimum allowable duration and maximum allowable duration. Special attention is also given to ensure that the noise clips don’t get attenuated because of speech embedding. These steps are repeated till it becomes impossible to further insert a KOI into the long audio.

Fig. 4: Multiple KOI embedded noise
The above steps are repeated multiple times to generate long audio of any desired length. Along with the generated long audio wav file, a metadata file and a csv is also created. The metadata will capture the properties of the contents of the long audio file. The csv file contains the start end location of KOI(s) in the long audio.

Fig. 5: Final generated long audio
Conclusion
Validating the accuracy of voice command engines against real-world conditions is a complicated task, requiring significant time and resources. In this article, we show how representative audio is created in a fully configurable and automated manner, allowing our Audio ML engineering teams to have confidence in our model accuracy very early in their development cycles.
In a follow-on article, we’ll be describing how the generated ‘long audio’ is integrated into our automated audio test framework in Ignitarium’s ‘Sound Lab’. This allows playback and accuracy testing of our voice engines running on physical embedded boards.
59 thoughts on “Simulating Real World Audio for Voice Command Engine”
Your article helped me a lot, is there any more related content? Thanks! https://accounts.binance.com/pt-PT/register?ref=DB40ITMB
I don’t think the title of your article matches the content lol. Just kidding, mainly because I had some doubts after reading the article.
IverCare Pharmacy: ivermectin 1.87 dosage for cats – liquid ivermectin for humans
Online betting Philippines: Online betting Philippines – Jollibet online sabong
maglaro ng Jiliko online sa Pilipinas: Jiliko – Jiliko app
https://gkwinviet.company/# Dang ky GK88
Online casino Jollibet Philippines: Jollibet online sabong – jollibet login
Jiliko casino walang deposit bonus para sa Pinoy [url=https://jilwin.pro/#]Jiliko bonus[/url] Jiliko login
Abutogel: Jackpot togel hari ini – Abutogel login
Jiliko: Jiliko casino walang deposit bonus para sa Pinoy – Jiliko login
Rút ti?n nhanh GK88: Trò choi n? hu GK88 – Casino online GK88
Swerte99 online gaming Pilipinas: Swerte99 casino – Swerte99 login
Swerte99 online gaming Pilipinas [url=https://swertewin.life/#]Swerte99 bonus[/url] Swerte99 online gaming Pilipinas
Slot gacor Beta138: Beta138 – Situs judi resmi berlisensi
Onlayn kazino Az?rbaycan: Onlayn rulet v? blackjack – Yüks?k RTP slotlar
https://jilwin.pro/# Jiliko casino
Jiliko bonus: maglaro ng Jiliko online sa Pilipinas – Jiliko login
Abutogel login: Abutogel login – Bandar togel resmi Indonesia
Jackpot togel hari ini [url=https://abutowin.icu/#]Bandar togel resmi Indonesia[/url] Abutogel login
Link alternatif Mandiribet: Slot jackpot terbesar Indonesia – Judi online deposit pulsa
Situs judi resmi berlisensi: Promo slot gacor hari ini – Bonus new member 100% Beta138
Swerte99: Swerte99 slots – Swerte99 login
Situs judi resmi berlisensi: Promo slot gacor hari ini – Beta138
https://t.me/s/Official_1win_kanal/1020
Link alternatif Mandiribet [url=https://mandiwinindo.site/#]Live casino Mandiribet[/url] Link alternatif Mandiribet
Swerte99 casino: Swerte99 login – Swerte99
https://mandiwinindo.site/# Bonus new member 100% Mandiribet
Jiliko casino walang deposit bonus para sa Pinoy: Jiliko casino – Jiliko login
Bonus new member 100% Beta138: Beta138 – Bandar bola resmi
Live casino Indonesia [url=https://betawinindo.top/#]Slot gacor Beta138[/url] Link alternatif Beta138
Situs judi resmi berlisensi: Withdraw cepat Beta138 – Live casino Indonesia
https://t.me/s/Webs_1WIN
Swerte99 login: Swerte99 slots – Swerte99 login
1winphili: Online betting Philippines – Online betting Philippines
Online betting Philippines: Online betting Philippines – jollibet login
Swerte99 online gaming Pilipinas [url=https://swertewin.life/#]Swerte99 casino[/url] Swerte99 app
Qeydiyyat bonusu Pinco casino: Canli krupyerl? oyunlar – Pinco il? real pul qazan
Swerte99 online gaming Pilipinas: Swerte99 – Swerte99 casino walang deposit bonus para sa Pinoy
https://mandiwinindo.site/# Mandiribet
Judi online deposit pulsa: Situs judi resmi berlisensi – Situs judi online terpercaya Indonesia
Online casino Jollibet Philippines [url=https://1winphili.company/#]jollibet app[/url] jollibet
jollibet casino: jollibet login – 1winphili
Bandar bola resmi: Login Beta138 – Link alternatif Beta138
Link alternatif Beta138: Bonus new member 100% Beta138 – Slot gacor Beta138
Yüks?k RTP slotlar: Pinco r?smi sayt – Qeydiyyat bonusu Pinco casino
Tro choi n? hu GK88 [url=https://gkwinviet.company/#]Tro choi n? hu GK88[/url] Khuy?n mai GK88
Rut ti?n nhanh GK88: GK88 – Nha cai uy tin Vi?t Nam
https://pinwinaz.pro/# Onlayn kazino Az?rbaycan
Nha cai uy tin Vi?t Nam: GK88 – Rut ti?n nhanh GK88
Pinco il? real pul qazan: Pinco casino mobil t?tbiq – Qeydiyyat bonusu Pinco casino
1winphili [url=https://1winphili.company/#]Online betting Philippines[/url] jollibet app
Jiliko login: Jiliko login – maglaro ng Jiliko online sa Pilipinas
Casino online GK88: Casino online GK88 – Ca cu?c tr?c tuy?n GK88
Официальный Telegram канал 1win Casinо. Казинo и ставки от 1вин. Фриспины, актуальное зеркало официального сайта 1 win. Регистрируйся в ван вин, соверши вход в один вин, получай бонус используя промокод и начните играть на реальные деньги.
https://t.me/s/Official_1win_kanal/4861
MediDirect USA: southern pharmacy – MediDirect USA
Indian Meds One: Indian Meds One – Indian Meds One
cheap cialis mexico: Mexican Pharmacy Hub – order azithromycin mexico
Indian Meds One [url=http://indianmedsone.com/#]п»їlegitimate online pharmacies india[/url] reputable indian online pharmacy
https://indianmedsone.shop/# Indian Meds One