Record 'Raspberry' & 'RaspberryPi' for a distributed dataset

Hey guys if you have some time please record yourself maybe 10x instances of either of the above as centred 1 sec wavs and zip / upload to a drive and share the Url here.

1 Like

Not many postings of raspberry! a big fat one to u all!

We can have a “Hey Marvin” crispy style though.

Could you please explain, what you mean with “centered 1 sec wavs” exactly?

Check the above dataset as its a collection of 1 second .wav files 16kHz mono where generally as much as possible the voice data is approx centred (placed or situated in the centre.) of the recording.

Also contributing a ‘Hey’ & a Marvin would not go amiss as really we should have a document or thread of distributed datasets that are stored on accessible drives.
We could quickly and easily for no extra cost create large distributed datasets that are merely a collection of distributed zip files of audio datasets held together by shared metadata.

I will do a ‘Hey’ & a ‘Marvin’ in the next couple of days.

Like this?

1 Like

Hey marvin 23 recorded samples

Yeah great a zip file with 10 or 20 would be great.

That is your file but what I mean as centred in the 1 sec wav but fine as it was

Ah, got it. Any tips on how to center the audio with Audacity?

I create two new tracks. One with 1s silence. The other one is where my recorded voice goes. Then I copy&paste to voice to the “silence-track”. But no idea, how to get it centered.

To be honest its not that important and I guesstimate as I grab them.
I just watch the start grab point and paste out till 1sec + then duplicate.
You use the <.> tool to move any wav.

1 sec wavs of a good recording like you supplied is the main thing and having some than not having any is the main thing

14x Raspberry

1 Like

Not sure how this would go but a marvin-stop with a hey-marvin2

1 Like

In there there is a dataset for the KW ‘Olivia’ as well