Monde greens tend to be annoying and funny sometimes. For those who don’t know what they are, it is the misunderstanding of song lyrics that usually turn out strange or funny. A transcription firm known as Verbal ink decided to test their professional’s ability in transcribing songs and how many errors they would make or if they’d make any funny monde greens.
They decided to pit their professionals with IBM Watson which is a speech to text application made by International Business Machines Corporation. This was not done for any scientific research; it was just for fun. Verbal ink tried their best to make sure that this test was as fair on both sides as possible. They did this by:
- Finding acapella versions of the songs
This was to ensure that Watson would not be confused by the rest of the music
- The Transcriptionists were made to simply swear that they haven’t heard the songs before and that they wouldn’t use Google to double check their work.
The experts chosen are Colby and Schenae. The songs that were used for the test were:
- Taylor swift- I knew you were trouble
- Van Halen- Runnin’ with the devil
- Elton john- tiny dancer
- Brad Paisley- ticks
The challenge was divided into 4 parts with each expert getting two challenges each while Watson would participate in all the challenges. The results were as follows:
-
Lyrics Challenge 1
Watson made some clear mistakes and missed about 6 words. It produced some monde greens such as “You got me a loon” instead of you get me alone. Schenae on the other hand did a good job with zero errors and words missing.
-
Lyrics Challenge 2
Colby did very well with 0 errors and words missing. Watson on the other hand made big mistakes and missed 15 words. This is a lot worse than in the previous challenge. Some monde greens were such as “You have some live n. n… Please start’ instead of Yes, I’m living at a pace that kills.
-
Lyrics Challenge 3
Just like the previous challenges Watson missed some words although it was less than before. It missed 3 words and made 8 errors with lyrics such as “Blue gene bear bear yeah” instead of Blue-jeaned baby. Schenae did a good job as usual with zero mistakes.
-
Lyrics Challenge 4
The trend stayed the same even in the last challenge. Watson kept making mistakes with 5 words missing and 12 errors while Colby did a perfect job.
CONCLUSION
The main aim was to see how many monde greens would be obtained from this experiment. From the observations done, Watson was too random and the experts were excellent at their job. This is not to underestimate Watson; it is still an amazing application.
