Freesound Forums

Hi
Just read today that the recently-active 'modularsamples' is to be a database of over 40,000 sounds. This means Jovica will finally lose his top spot after (about) 10 years.
Here are some other ways statistics will change on freesound:
NOTE: all stats apply to freesound before all samples were uploaded.
-At present, the most common tag on Freesound is 'field-recording'- around 24000 sounds have this tag, about 10% of all sounds on freesound. When all 40,000 samples are uploaded, 'multisample' will become the most common tag- there's 12448 pre-modularsamples, after there will be around 52448, or 18% of all sounds on freesound. 'field-recording' will shrink down to 9%.
-16% of all sounds on freesound will be by modularsamples.
-The 0 licence is used presently on 31% of sounds, around 76000. Attribution is used on 118000, around 49%. After modularsamples 0 licence will account to 41% of sounds, though Attribution is still merely in the lead with 42%.
-1 in 8 sounds on freesound will be by modularsamples. This means one may be selected every 8 days.

Just thought this might be an interesting look at it
Thanks
kwah

Thank you for taking the time to draw up these stats

Appreciated, and interesting to view.

Hi Kwah,

thanks for your post!
we're thinking of the impact of the modular samples library in Freesound, the stats you show are right.
I think this is in fact a good opportunity do discuss some aspects like how the "all time most active users" chart is computed.

One thing we already changed to avoid a "negative" impact of the library is the way in which the tagcloud is computed in http://www.freesound.org/browse/tags/. Instead of tanking into account all tags for computing the tagcloud, we now "discard" all the tags that appear more than once in a single pack. Hence, if a pack has 200 sounds with the tag "multisample", that tag only counts as one for that pack. In this way the tagcloud has been significantly improved.

Regarding the users ranking, I think that we should change the way it is computed. Right now, we basically count the number of uploaded sounds and the number of forum posts (number of forum posts has less relevance). I think we could also consider ratings and comments performed, and compute it in a way that number of uploaded sounds is not completely determinant.

Furthermore, regarding the random sound of the day, I guess we could add some limitations like not picking a sound from a user that already got a sound selected as the "random sound of the day" in the last XX days.

What do you think?

I love this type of statistical analysis. Great community service Kwahmah.

Frederic, your optimizations make sense to me, especially the ‘sound of the day’ one. I tend to see the same sounds reappear every so often, although I don’t know if this is something to do with browser cacheing or something.

In any case, this modularsamples partnership is a great idea. I love a bit of sample and synthesis, which is a very underappreciated method of sound design (though not among the Freesound community).

frederic.font wrote:
Furthermore, regarding the random sound of the day, I guess we could add some limitations like not picking a sound from a user that already got a sound selected as the "random sound of the day" in the last XX days.

What do you think?

Seriously, now. As for the most active users.
Most people like to get credit / recognition for something. Why not have an expanded section, looking at the different aspects of Freesound?
"All time most active users" - as we already have
"Most downloaded users"
"Highest rated users" (most stars)
"Most active commentators"
"Most active forum users"
etc...

We could also have something recognizing the most active users in the last week / month.
Maybe these features could encourage a bit of competition between users and make the community more active?

Of course, all of this is very nice, but does not necessarily add or aid to the main purpose of Freesound. - So maybe low priority if there are other more useful features to implement.

IMHO most active users list should be considered as combination of:
- 60 % downloads of user sounds
- 20 % number of uploaded sounds
- 10 % number of forum posts
- 10 % commenting activity

now those stats are just useless

Sorry guys

If there's a way to remove me from the stats entirely I would encourage that. I'm not exactly an active user, it's only due the amazing efforts of Fredric that the MS library is on the site in the first place.

modularsamples wrote:
Sorry guys

If there's a way to remove me from the stats entirely I would encourage that. I'm not exactly an active user, it's only due the amazing efforts of Fredric that the MS library is on the site in the first place.

I don’t think anyone’s unhappy about the samples themselves, least of all me. They rule.

But if people really feel strongly about ‘fixing’ the stats, then maybe just leaving MS out of the stats calculation is indeed the best option.

jamesabdulrahman wrote:
But if people really feel strongly about ‘fixing’ the stats, then maybe just leaving MS out of the stats calculation is indeed the best option.

Thanks modularsamples for making your huge collection available at Freesound!

Indeed, I wouldn’t call it a high priority. And as I said, the samples are great.

Synth on!

Hi,

I kind of like the idea of being able to show most active users depending on different stats, and still I think the "overall" most active users should be a combination of users' ranking in the different facets. When I have some time to actually work on this I'll let you know and maybe we can decide among two or three options. In any case, it is great to have this discussion with you!

Hi Fred

Speaking of skewing statistics...
When we count posts to give points to active users, is there a (easy) way of discounting the posts in the off-topic section?
There is a fair bit of nonsense there, especially my "random word association game" that has been going on for 1 year now... Would not bee too surprised if half of my posts are on that thread. And, although it is fun, I am not sure it adds a lot of relevant contents

Hi,

we could do that but I think off topic is also part of Freesound and it gives "life" to the site so I think it is totally fine to count this posts

I had the same thought AlienXXX a while ago, and the majority of my post count has come from that thread, for which I feel slight dishonour.

However, I don't think what you suggest would be the right thing to do. It seems a bit drastic. Apart from the game, there is not much regular activity in that sub forum, so It doesn't really make sense to negate any accrued post count.

We could lock the word game thread if you feel like it's cheating.

Headphaze wrote:
We could lock the word game (...).

AlienXXX wrote:
NOOOOOoooooooooooooooo!

Didn't think so

you know you'll have to record all these words and upload them to freesound at the end right?
could be a dare for the freesound 20th anniversary, and by that time there will probably be more words on the thread than sounds by modularsamples (to link with the initial discussion of this thread)

frederic.font wrote:
you know you'll have to record all these words and upload them to freesound at the end right?
could be a dare for the freesound 20th anniversary, and by that time there will probably be more words on the thread than sounds by modularsamples (to link with the initial discussion of this thread)

ACTUALLY... without knowing, you have just sparked the next dare!
(but it is not what you think! and you have to wait until 1st Aug for it to be announced officially...)

Bugger... derrailed the thread again!
So, back to something related to the thread topic:
Yes, there will also be a Modularsamples dare. It is not every day that you add gigabytes of contents to Freesound from a single user. That fact has to be celebrated!

AlienXXX wrote:
It is not every day that you add gigabytes of contents to Freesound from a single user. That fact has to be celebrated!

Agreed