alastairp
Joined 15 years ago
Karma
167
Submissions
17
Member since
Jun 23
Recent submissions
Comment · 6 months ago
Comment · 11 months ago
On a similar note Matt Parker recently released a video about Perlin Noise, which won an Oscar (for Technical Achievement) in 1996: https://www.youtube.com/watch?v=JrLSfSh43oA
Comment · 3 years ago
Another anecdotal description of old staircases that I've heard of before is from Burgos castle in Spain, where (it's said) that the stairs to the bottom of the well change direction half way down to prevent you from getting too dizzy [1]
> Se accede al interior por unas escaleras de caracol. Para evitar el mareo, los 4 primeros tramos se hacen en el sentido de las agujas del reloj y los dos últimos tramos en sentido contrario.
[1] https://rutasparatodaslasedades.blogspot.com/2019/07/el-cast...
Comment · 3 years ago
A similar embeddings model based on Discogs genre/style data is the Effnet-Discogs model made at the Music Technology Group at Universitat Pompeu Fabra: https://replicate.com/mtg/effnet-discogs
Comment · 3 years ago
In fact, ListenBrainz (partner project to MusicBrainz) is doing some stuff similar to what you mention about listening histories. We're using the data to generate similarity based on when songs are listened to each other in "listening sessions" along with other songs.
Follow the troi-bot user with a ListenBrainz account, and we'll generate you a daily playlist: https://listenbrainz.org/user/troi-bot
This is still very much work-in-progress, but we're doing as much as possible out in the open to solicit feedback from people.
Comment · 3 years ago
Hi, Freesound developer here, this is great! We're a small research team at a university[1], and we're also working on auto-captioning tasks that include trying to describe sound effects based on the audio content, although we're not doing anything with OpenAI or summarisation of metadata at the moment.
A reminder, we also have acoustic similarity using signal processing analysis and nearest neighbour lookup for all sounds[2]
Could you update the sound displays to include a link to the sound page on Freesound, for citation purposes? You can use https://freesound.org/s/[soundid] or use our embedded player. See the "Embed" option in the share box on the right -hand-side toolbar of any sound page.
Are you interested in integrating this into FS? Talk to us![3]
[1] https://www.upf.edu/web/mtg/
[2] https://freesound.org/people/lebaston100/sounds/243627/simil...
Comment · 3 years ago
I encountered something similar to this last month. I was visiting Canada from Spain, and got an esim from https://www.airalo.com/. Due to what I guess is airalo obtaining transit from the cheapest provider, the data connection ended up being bounced through Czechia.
This meant that my android phone was a bit confused at times. When on mobile data the weather widget would default to showing me weather for random Czech cities, and search results would be Czech-localised. When on wifi connections, results were in English and Canada-localised, unless I bounced through a VPN exit node at home.
After 3-4 days, ads on Youtube started becoming Canada or Czechia-localised depending on if I was watching it via wifi or mobile data. It seems that google eventually decided that I was in Canada, continuing to play me Canadian ads even when I got home to Spain, on both wifi and mobile data. What was even stranger is that my partner started getting Canadian ads on youtube on her ipad too, which doesn't have any details about my google account (other than us living on the same internet connection). It took about 2 weeks for google to start playing us Spanish ads again.
Comment · 4 years ago
The tickets are definitely pitched as "for regular commuters", but I'm going to take advantage of them for getting further out of Barcelona to cycle. I've spent 70€ on 4 tickets and expect that I'll receive refunds for 2 of them (by travelling at least 16 times on a ticket). A return ticket to a location 2 hours away ends up being close to 15€, so even if I get no refunds due to not travelling enough, it's still economical to buy the tickets if I'm planning on using them 2-3 times a month.
Comment · 4 years ago
hi, I'm a freesound developer! We'll reach out to them.
Comment · 4 years ago
This is actually still a difficult task and still under active research. The name of the phenomenon is "tempo octave error". Typically an algorithm looks for evenly-spaced strong pulses of energy, and infers the BPM from that. If there is a strong beat at multiple of the actual BPM (half, double, 4x, etc) then it could be mistakenly identified as the BPM. As alin23 points out in a sibling comment it seems like the Spotify algorithm at least has a confidence level here. There is some more information about BPM computation and octave errors at https://www.audiolabs-erlangen.de/resources/MIR/FMP/C6/C6S2_...
Comment · 4 years ago
For key and BPM data, the MetaBrainz foundation also has AcousticBrainz: https://acousticbrainz.org/ (I'm a developer on this project).
Unfortunately, I would say that the data that we have in AcousticBrainz isn't as good as what's in the Spotify API, although the algorithms that we use are completely free (available in the Essentia signal processing library
- https://essentia.upf.edu/). Over the last few years the algorithms in Essentia have improved and we're hoping to release a new version of the tool used in AcousticBrainz to improve the database.
Comment · 4 years ago
Do you mean if the song has a vocalist singing? At the Music Technology Group (https://www.upf.edu/web/mtg) we have some classifiers that do something similar there are some demos at https://replicate.com/mtg/music-classifiers, see the output of the "voice_instrumental" classifier. This is built with a dataset of examples of instrumental songs and songs with singing in them. Normally classifiers are trained on timbre of the song to identify this.
Earlier versions of this classifier are also available at AcousticBrainz: https://acousticbrainz.org/
Comment · 5 years ago
Interesting, thanks for the report. I can't see this behaviour when testing myself. Do you mean that there is a network load of the file every few milliseconds, or that the file disappears and reappears in the DOM? From what I can tell, this file is added with static CSS, so it shouldn't be reloading all the time. Are you playing from the search results page, or a sound page? Bug reports are welcome: https://github.com/MTG/freesound/issues
Comment · 5 years ago
I'm a researcher at the Music Technology group who works on freesound (https://www.upf.edu/web/mtg)
We also maintain Freesound labs, which lists a lot of projects and research made using content from freesound: https://labs.freesound.org/
FSD50k is our hand-curated dataset of sounds designed for research in sound and event recognition tasks https://zenodo.org/record/4060432#.X3xrgi8RqL4, http://dcase.community/challenge2019/task-audio-tagging
To me one of the clear things that GitHub gave us was a structure around a person rather than a project. To me it felt liberating to quickly create a repository attached to my name than it was to go through the (what felt to me) very serious process of coming up with a project name and reserving it on sourceforge just to get a cvs or svn repository (along with website, mailing lists, issue tracking(?), etc, etc...). It felt like the mental load of "oh this is just a quick thing" was a lot easier with github.
> It gave projects issue trackers, pull requests, release pages, wikis, organization pages, API access, webhooks, and later CI.
Although it didn't give us this all at once. I still remember when we created a new user account in order to simulate an organisation, before they existed. I distinctly recall discussing with friends if we wanted to set up a bug tracker software for our project with the assumption that "GitHub will probably release one in a few months anyway". In the end we just kept a text file committed in the repository. Issues were announced a few months later.