Facebook PixelShazam for spoken word
Brainstorming
Brainstorming
Create newCreate new
EverythingEverything
Sessions onlySessions only
Ideas onlyIdeas only
Idea

Shazam for spoken word

Image credit: https://mashable.com/article/identify-song-apps; https://www.lifehacker.com.au/2017/09/use-the-microphone-when-public-speaking/

Loading...
Povilas S
Povilas S Jan 04, 2022
Please leave the feedback on this idea
Originality

Is it original or innovative?

Feasibility

Is it feasible?

Necessity

Is it targeting an unsolved problem?

Conciseness

Is it concisely described?

Identify quotes, bits of conversations, lines from movies, etc. Any type of speech samples mixed into songs or other types of media played anywhere through speakers that you can "Shazam" with a dedicated app.
The app displays you original source of the speech, identifies the speaker, gives some background information about it, or, if it can't find the original source, it displays reciting sources (sources sampling the speech) in chronological order.
This could be a Shazam extension to identify sources of spoken word samples in songs, but it could be developed into a separate app to identify any kind of spoken word sample used in any medium.
Why?
You like a quote/speech sample mixed into a song - you'd have to google to find the source/speaker and often googling this is unsuccessful. You like a speech sample used in a youtube video and want to know the author/listen to the full speech, know the context, etc.
4
Creative contributions

Shazam for movies and tv shows

Loading...
JN
J. Nikola Jan 10, 2022
That's a great idea and I think it could work easily with existing technology.
I would suggest expanding the function to all the movies, tv shows, or anything recorded with sound!
The key thing we need is, as Povilas S and Darryl Koh Yuan Jie mentioned, a database. The song library is huge and ever-increasing, but the number of movies, tv shows, and Youtube videos is enormous. One guy on Quora tried to calculate the length of all songs on iTunes and reached an astonishing number of 1.300,000 hours of music. Since the movies, tv shows, and Youtube videos are much longer than the average song (3 min), this database would be several times larger and thus, a bigger challenge.
Please leave the feedback on this

Plagiarism checker

Loading...
Spook Louw
Spook Louw Jan 14, 2022
Theoretically, this technology could easily be paired with something like Turnitin to become a real-time plagiarism checker.
The app would simply convert the audio it receives and check it against its database to be able to detect whether or not the speaker has stolen material from another source.
This has obviously been optimized for academic papers, but there's no reason that it wouldn't work just as well for jokes, poetry, lyrics or speeches.
Please leave the feedback on this
Loading...
Povilas S
Povilas S11 days ago
You suggestion is similar to what J. Nikola has proposed in his contribution.
Please leave the feedback on this

Similarity search function for spoken word

Loading...
JN
J. Nikola Jan 10, 2022
There is a one really cool thing I would like to do with the song Shazam, but it's not available (tell me if it is). I would like to Shazam or search for a song and see how many songs of similar patterns are there. It would be like matching the test sequence of DNA or RNA to the genome database. The result should be a list of songs sorted by similarity (beat, frequencies, length, word selection, genre, ...). I am sure the music industry uses similar tools, but I would like to use it for fun, too.
It would be cool to see the same thing with the spoken word. The idea would not be there to give the credits and financial support to the "owner", but to prepare original public speeches, movie dialogues, or commercial jingles. I bet you can think of many more applications. Of course, not every word or a quote would be in the database, but only the ones played many times.
Please leave the feedback on this
Loading...
Povilas S
Povilas S18 days ago
Yes, that's pretty much exactly what the idea is about, if I understood correctly, what you mean.
Please leave the feedback on this
Loading...
JN
J. Nikola17 days ago
Povilas S Sorry, I rephrased it. Can you check now if it makes any sense?
Please leave the feedback on this
Loading...
Povilas S
Povilas S17 days ago
J. Nikola So you mean you would like to be able to search for similar speeches, quotes, etc?
Please leave the feedback on this

Shazam solves a real problem

Loading...
JW
Jaballah Wajdi Jan 21, 2022
Although it is simple, the problem solved by shazam is an annoying problem for many people around the world, we can say for all of us.
Sometimes we suffer when we spend a lot of time looking for a song or a full speech. Then using this app we get what we want in seconds.
Moreover, the app can also be used for other services like plagiarism, and avoid confusion between songs with similar words for example.
Please leave the feedback on this

Add your creative contribution

0 / 200

Added via the text editor

Sign up or

or

Guest sign up

* Indicates a required field

By using this platform you agree to our terms of service and privacy policy.

General comments

Loading...
Darryl Koh Yuan Jie
Darryl Koh Yuan Jie23 days ago
I think this could work for popular songs, quotes or speeches but it may be harder for relatively unknown ones. On wikipedia, it states that "Shazam identifies songs based on an audio fingerprint based on a time-frequency graph called a spectrogram(...)Shazam stores a catalogue of audio fingerprints in a database. The user tags a song for 10 seconds and the application creates an audio fingerprint. Shazam works by analyzing the captured sound and seeking a match based on an acoustic fingerprint in a database of millions of songs." Link here
I found this article that describes it in a much more extensive manner. Shazam makes use of audio frequencies to generate audio fingerprints unique to the particular audio. However, it seems that for Shazam to work well, it requires many users to submit the same song or snippets of the same audio recording to ensure it exists in the database in order for Shazam to match them together. As what you mentioned, it could work if there is a melody or tune mixed into the spoken word, however I am not sure how feasible it would be for say a speech, especially one that is not popular or likely to be recorded by other users.
Please leave the feedback on this
Loading...
Michaela D
Michaela D23 days ago
That would be a cool feature if you surpass the practical difficulties. It may be challenging to recognize a bit unless you are able to replay it, or it is long enough. Shazam takes a few seconds to recognize a song, if we add to that a few seconds to activate the app, then you would need something that lasts at least 10 seconds. In case of a quote that is a bit long. However, if you are able to play it over and over then it would work.
Please leave the feedback on this
Loading...
Povilas S
Povilas S23 days ago
Michaela D, Darryl Koh Yuan Jie Regarding identification, I have a feeling that with spoken word it might be even easier than with songs. Many songs have similar bits - chords, parts of the melody, rhythm, so it's not easy for an app to identify the exact song, especially if it's a remix, cover, a song using samples of other songs, etc. Once you start thinking about this it becomes obvious that what Shazam does is a pretty difficult task to achieve. However, despite this, many songs Shazam identifies in the first few seconds or even seemingly instantly after you press the "listen" button.
In the case of spoken word, on the other hand, you have a voice of a certain person which is very distinctive and you have a certain combination of words that you may not find anywhere else except that speech, it's like a code of sounds (letters, words, arranged in a specific manner pronounced with a specific timbre of someone's voice) so all the algorithm has to do is to read the code.
What is not an easy task to do, however, is to make a large enough database of spoken-word files. With music, it's easier, cause there are already databases combining a lot of known songs, so all you have to do is get access to those, join a few together, etc. With speeches and other forms of spoken word, databases that exist are rather slim and scattered all around, you might find a database containing all speeches of one or a few famous speakers, but that's about it, I think.
Spoken word samples might come from many different sources, e.g. movies, podcasts, public speeches, lectures, etc. So to combine all this into one database would be a huge task to do. Also, the database would have to be continuously updated with new data.
Please leave the feedback on this
Loading...
Darryl Koh Yuan Jie
Darryl Koh Yuan Jie22 days ago
Povilas S Yes, you are right, Shazam is definitely a innovative form of technology. I was watching this video by Real Engineering on youtube that basically describes the audio fingerprint and hash function Shazam uses in its algorithm. From my understanding however, remixes of songs actually won't be a problem. Since every song(even similar ones) have very different audio fingerprints, it is still possible for Shazam to differentiate them. The most important factor in finding the answer to the audio you are looking for would be whether it exist in Shazam's extensive database. If the algorithm is unable to match the audio to their database, it is not possible to identify it.
Indeed, speeches and unknown spoken words might be the real challenge we face here. Currently, I believe Shazam includes Spotify as part of its database which helps tremendously. Perhaps, if Shazam was able to include Youtube and other platforms that consist of many samples from different sources as what you mentioned, it could prove useful. Currently, youtube is arguably the most popular video/audio streaming platform and definitely contains many movies and speeches. It also houses many unique one of a kind videos and memes which can only be found on exclusively there, so it would be a good place to start.
Please leave the feedback on this