Freesound Project

Hello guys,

There seems to be a need for a bit more guidance when it comes to describing the sound effects that you upload to freesound. This is why I'm planning to write a FAQ for freesound users with suggestions on how to write clear and usefull descriptions.
It is important to have a description of a sound, it can help you find specific sounds through the search for example. What use is our major sound library going to have if you are not able to find a specific sound when you need it. There are also reasons (i believe) from a technical point of view in the freesound project, perhaps Bram would like to elaborate on that.
However, this is a problem that all sound libraries have. So perhaps its best to have a look at how they solve this problem and then turn to our specific situation here at freesound and see if we can adapt their model to ours. Perhaps we could come to form some sort of standard to which we can compare and make sure that all at least at some level fit this standard.
In sampling, especially the big part of sampling that contains field recordings, there are several layers in which we can describe sounds.

• Macro
  
• Meso
  
• Micro
  
• Nano?
  

Macro
First of all, there is the macro description. This describes the whole event of a sound. For example "this is a recording of a car crash". For somebody looking for a sound of a car crash this description is perfect.

Meso
However, what if your looking for the sound of metal and glass breaking? This car crash could be very usefull, but perhaps if you hadnt come up with that you might not have located this source. So in order for this to work you need to break up the sample in smaller events. The description now becomes :

"Car Crash: breaking of glass, crushing metal on metal, burning rubber."

Micro
this is the level in which you use the same kind of language usually found in modular synthesis. the Envelope of a sound, the timbral aspects etc. (Dennis Smalley: Spectromorphology).

Technical Description
Another important factor is the technology used to create the sample. Make sure you list:

• Recorder: MD, DAT, HDD, (if pc, what kind of soundcard)
  
• Microphone: model & type: Shure SM57 dynamic mike
  
• Processing: If so, what kind.
  

This is just a first setup, dont have time to do more at the moment. Somebody might want to add?

I would like to add one example more to this guide, i will take a random sound, one uploaded by elmomo, thanks for this file elmomo, all my respects for you. I choosed this sample randomly for the example, nothing against it.

CLICK HERE

The description here is:

"I was near the church of girona at 12o'clock pm, it sounds like that. "

A more complete description would be for example:

"Church bells "
"The bells of the church of Girona at 12 o'clock "
"some hiss sound in the background"
And also if you remember
"recorded with X microphone and X gear"
---------------------------------------------

The purpose of description is, at one hand to provide information (preferently deep) about the sound, from where it comes, it's peculiarities and the tools used to create/record it ... and at the other hand to be used when doing a search.

The tags are something very nice, even better than text searches in some cases, but anyway, the descriptions should be as much descriptive and complete as we can make them, in my oppinion.

Xcuse the blabla, have fun.

I can't stress "good descriptions" enough...

See above, and, when describing you could also try to ask yourself a lot of questions like:

• from where : the source of the sound.
  
• what: what do I hear, try to describe the sound.
  
• where: if you sampled or recorded this sound, where did you do it?
  
• method: what gear did you use to sample the sound?
  
• .........
  

- Bram

Here is papa plaga again with another of those "DON'T"

When u are uploading a sample pack of, i.e. pieces of guitar melodies, chunks of analog synth bursts, excerpts of female moans...

...do not upload them all with the same description like "excerpts of female moans" into a sample pack.

Instead, try to describe each of the differences between samples, if there are no differences it's ok, you can copy paste the description... but if one song is short and the other one long, one is high pitched and the other one low, one is rythmic and the other one not... in this cases, please, take note of the characteristic on the description and add the TAG too.

This is very important.

In the near future, we will have humongous ammounts of samples, to search efficiently between them we will need to use keywords to refine the searches.

If i search for "female moans" and there are 2000000 of them, i will try to refine the search using other keywords like "high pitch, short, female moan" or "low, long, deep, rattling, female moan".

That's why we need the good, individual descriptions for sample.

thank you for your time and xcuse me for the blahblah.

this is great. i am so intrigued with the challenge of organizing, well anything really . . . but especially sounds.

given the breakdown of "macro, meso, micro", it would be interesting to see a database that actually used fields like that instead of just "description". the first problem is that not enough people would use the database if they had to think so hard just to describe a sound. a possible solution to that would be keeping a simple "description" field, and then finding a team of moderators with nothing better to do than go through creative more acurrate "macro, meso, micro" descriptions. haha. i know i don't have the time to do that!

i find it incredibly easy to make sounds, but incredibly hard to describe the results, but i'm working on it. requiring the descriptions and tags has made me finally put in a better effort in the "description" stage for me in general (ie not just here, but also in other projects). andrew

i feel the same way, andrew. it is challenging to come up with a description for any sound. contributing to freesound has really put my mind to work and helped me express what something sounds like in a logical / word fashion.

***yesterday afternoon i spent literally hours coming up with better tags and descriptions for my contributions to freesound (122, so it took a *long* time). my wife was like "why are you so grumpy this afternoon?" and i was like "because i've got to come up with descriptions for all of these sounds i made" and she was like "that sounds like fun! let me listen, and i'll do the describing" and i was like "but you have to use a certain logical framework like '2-bar rhythmic field-recording loop processed from a water sample recorded here in Halifax'" and as soon as she realized that she couldn't just describe things as "sounds like a fart" (she really did say that!), she realized why i was grumpy and why it was taking so long to describe all those sounds i made. i think it is because making the sounds is utilizing the creative side of the brain, whereas describing them is the analytical part; so the creative part seems like fun and the describing part seems like work. but i understand as i get more involved in freesound how important good tags and descriptions are. when i first started submitting, i just used "field-recording" and "sound-design" as tags and descriptions and i didn't clue in right away why bram and anton wanted something more descriptive. now i understand how the more details are provided the better. andrew

as a sidenote: saying "sounds like a fart" is actualy a very valid description, as it tells you something about the sound!

http://freesound.iua.upf.edu/searchText.php?search=fart


- Bram

I understand; it's just that my wife thinks most of my music and sound design sounds like farting, so if she were
doing the descriptions for my uploads, there'd be a lot of "sounds like a fart" descriptions! andrew

I don't just make samples to make samples, normally, my samples have a purpose: eg. My trumpet loops were for a composition for my school musical. As was the tape flip. Putting some of this context in is helpful when browsing I find.

What I like to do when describing samples is this (this isn't what I have done so far but it is what I will do from now):

First give context
Then give specifications (What mic(s), what interface, what processing)
And then musical specifications (If applicable) (Time Signature, Key, BPM, etc.)
Finally describe the sample in a non-technical way (eg: A soaring trumpet fanfare which can be looped)