Sonora STILINGUE

Index:

1. New possibilities for Social Listening
2. Where is Sonora?
3. What are the limitations?
4. How to acquire?

Discover relevant insights on social media through speech recognition in audio and video.

The consumption of audiovisual content on social media continues to grow in Brazil and around the world: over 40% of Brazilians with internet access already consume podcasts, and it is estimated that 23.5% of the world's population will start listening to this format by 2024. Regarding video consumption, it is estimated that 99% of Brazilians watched some type of video on social media or streaming platforms in 2020. 

Another recent Kantar Ibope Media study indicates that 31% of the population prefers watching streaming videos together, demonstrating an increasingly consolidated social habit. With the growing demand, there is also a need to gather and analyze information contained in videos and podcasts. 

With this in mind, STILINGUE developed Sonora, a technology that is part of the Listening module and allows "listening" to audio and video in YouTube and Twitter posts to bring more richness to analyzes within your research. 

Customized Services: need help obtaining exclusive reports for your analysis? We can assist you! Our team of experts is ready to assist you. Contact your STILINGUE Account Manager to access this and several other services.

New possibilities for Social Listening

Like other products that use STILINGUE Artificial Intelligence (AI), Sonora's speech recognition technology transcribes audio and video content to generate insights about brands and products. This means you will have even broader coverage in your research by discovering mentions and potential influencers that could not be found otherwise. 

It is important to note that initially Sonora is not responsible for collecting publications: its role is to act on audio and video content already collected (based on the groups in your Search Configuration) and transcribe these contents into text format within STILINGUE. This way, you have a more comprehensive setup with improvements in classification and more refined analyzes in your Listening research.

For the technology to work correctly, it is necessary for the terms of interest (such as mentions of a brand or product) to match the terms existing in the groups or themes of your Search Configuration. In other words, a brand not previously configured in a group or theme cannot be identified by Sonora.

Check out some of the applications that Sonora offers:

• Discovery of mentions in review and comparison videos: a significant portion of content creators (found on platforms like YouTube, for example) focus on the analysis and comparison of various products. The brand or product is not always mentioned in the title or caption of the video, but with Sonora, it's possible to extract this information directly from the audio;
• Identification of influencers: in addition to finding narratives mentioning your brand or product, Sonora's speech recognition can identify new influencers - who may be content creators on a large-scale video channel or podcast - with the potential to become brand ambassadors in future actions;

• Time-saving in relation to content: monitoring crises and specific campaigns requires speed in analysis. Sonora's technology can be an essential support for these and other fronts thanks to the spotting feature, which allows identifying exactly when a mention was made in the video. The time for analysis decreases considerably;
• More accurate brand perception: by providing insights from a new source of content, Sonora can reveal different contexts, trends, and scenarios that truly represent the scope of your brand on social media.

Where is Sonora?

The technology that makes up Sonora is present within Listening and can be found in posts with videos. You can filter transcriptions by the "Sonora" tag in the Publication Filter and view them on the Publication Wall.

Transcriptions from videos are divided into 1-minute segments. Each card includes the following information: 

1. Spotting

Inform the time intervals (1 minute each) in which the extracted texts were identified. Thus, if a brand is mentioned in the video, the time will be shown with the citation.

2. Transcription

Textual recording of the identified speeches with Sonora. 

What are the limitations?

• Sonora is available only in Listening and initially covers publications from YouTube and Twitter, for compliance reasons. However, it is possible to customize the functionality to extract audio from podcasts and other formats according to the need, as a separate service. Contact your account manager to check the possibility with the STILINGUE Professional Services team;
• Sonora does not act on retroactive publications, but you can hire a specific service to transcribe publications prior to activation;
• Transcriptions do not include punctuation in the text. This means that only the words from the audio will be extracted and presented on the card;
• The audio quality and other factors such as pronunciation and the use of foreign terms may compromise the generated transcription in some cases, as Sonora was developed specifically for brands in the context of Brazilian Portuguese and is still in the evolution phase.

How to acquire?

Sonora is a Service Delivery feature, meaning it allows customization by including other proprietary audio sources. If you want more information and access, please contact your account manager.

For more information, visit the discussion on the subject at our community or videos on our channel. 😃