SWR Media Services

Tablet with screenshot of Fraunhofer IDMT Ad Quality Monitoring
© Fraunhofer IDMT/bongkarn - stock.adobe.com

As a subsidiary of the German public broadcasting corporation SWR, SWR Media Services markets advertising time and sponsorship in the SWR radio programs and on SWR television. With our audio matching software, it is now possible to automatically check whether the provided commercials and sponsor mentions are broadcast correctly in the program. 

Challenge

Advertisers expect their commercials or sponsor mentions to be broadcast in full and at the agreed-upon times. Whether the broadcast was carried out without any loss of quality was previously checked manually. This involved listening to the recorded advertising blocks after broadcasting and documenting whether the spots could be heard in full and without errors - a time-consuming and labor-intensive process.

Realization

The audio matching software significantly streamlines and speeds up the labor-intensive manual process. The software automatically detects and documents whether and how closely the aired commercials match the reference material.

The software initially creates so-called unique acoustic fingerprints for all reference commercials and sponsor mentions. These fingerprints are then compared in real-time with the live stream of the radio programs. The software even distinguishes nearly identical sponsorship announcements, where often only individual words vary, such as "the weather was presented" and "the weather is presented," which nevertheless must be clearly assigned and documented.

If the software recognizes that the respective fingerprints of the reference material and the advertising block in the live stream match but do not meet the predetermined quality score, it sends a warning e-mail after a few minutes. This might occur, for example, if the broadcast was interrupted by an unplanned traffic report or overlaid by the announcer's moderation text. This ensures the prompt re-airing of incorrectly broadcast commercials.

This might also be interesting for you

Research topic

Audio and Visual Content Analysis

Extracting meaningful data from audiovisual content

 

Research topic

Automatic Music Analysis

 

Query-based Audio Matching

Detect a given audio query within a stream or file – even under noisy conditions or with a very short query