Naming Conventions For Individual Samples

[dannthr]

Hey fam,

I'm curious if there is a common naming convention for individual samples such that it cleanly organizes them between velocity and pitch key regions. Are there conventions or a consensus or is it the wild west?

Especially naming conventions that are supportive of auto-mapping scripts, etc.

Thoughts? What's your name scheme?

Thanks,
- Dan

Click to read more...

 

[Uiroo]

There is a convention for Kontakt auto-mapping I believe, though I don't use it.
Naming conventions can vary according to what you use the samples for, but my naming convention is usually:
Mic_Articulation_Pitch_Dynamic_RoundRobin

What's your usecase?

 

[d.healey]

Always use MIDI note numbers in your file names, don't use note names. C3 is that middle C? or is C4 middle C? Doesn't matter, use 60 because that's always middle C. Also when you're scripting it's much easier to deal with numbers than strings for mathematical calculations.

With Kontakt you're limited to the number of segments your filename can have, I think it's 7 if I remember correctly. So include all the important info in the file name and for further categorization you can use the folder structure. If you're using a different sampler such as HISE then this limitation isn't a problem and you can add more info in the filename.

The info you'll include will also vary depending on the instrument. For example a guitar sample filename might include the string, the picking direction, the pick position, etc.

So I'll give some examples of what I do.

Guitar:
"instrumentName_articulation_pickDirection_stringName_noteNumber_dynamic_roundRobin_mic"

Violin (bow direction isn't necessary unless you recorded more than one):
"instrumentName_articulation_bowDirection_noteNumber_dynamic_roundRobin_mic"

Frame Drums:
"instrumentName_drumName_stickName_hitPosition_direction_noteNumber_dynamic_roundRobin_mic"

Orange Tree Samples posted an article years ago about their system. They rely on folder structure for organizing their samples. I prefer to have all the data within the filename so that if I ever take it out of its folder structure I know what it is. But for a Kontakt based library I think a detailed folder structure is often necessary - https://www.orangetreesamples.com/blog/easy-sample-library-mapping-method

 

[Uiroo]

The order can be an important consideration too.
For example, in Kontakt, if I want to filter my samples so it lists all samples from a specific articulation and a specific mic, articulation and mic would need to be next to each other in the filename.

 

[dannthr]

There is a convention for Kontakt auto-mapping I believe, though I don't use it.
Naming conventions can vary according to what you use the samples for, but my naming convention is usually:
Mic_Articulation_Pitch_Dynamic_RoundRobin

What's your usecase?

Great question! Over the past year, Epic's recently acquired Harmonix team has been porting and developing tools for Unreal Audio Engine to support new Fortnite experiences like Fortnite Festival, Fortnite Jams, and Patchworks for Creative/UEFN. In UE 5.4, a plugin of these tools are releasing/have been released (Preview 1 is available now) and it includes a multi-sampler node for MetaSounds.

I've been learning what I can about the tool for my own experiments and to hopefully throw down a couple tutorial videos and naming conventions to automap keyregions was something I asked them about and they didn't have anything in the works, but asked back what I thought would be a good naming convention.

What I offered in return was the SFZ format which has metadata for expressing key and pitch ranges per sample, which could be converted to a naming convention--if a bit explicit. And I started thinking about conventions I might've seen, etc. and so that drove me to post here.

Where all the experts in sample libraries will see.

Always use MIDI note numbers in your file names, don't use note names. C3 is that middle C? or is C4 middle C? Doesn't matter, use 60 because that's always middle C. Also when you're scripting it's much easier to deal with numbers than strings for mathematical calculations.

With Kontakt you're limited to the number of segments your filename can have, I think it's 7 if I remember correctly. So include all the important info in the file name and for further categorization you can use the folder structure. If you're using a different sampler such as HISE then this limitation isn't a problem and you can add more info in the filename.

The info you'll include will also vary depending on the instrument. For example a guitar sample filename might include the string, the picking direction, the pick position, etc.

So I'll give some examples of what I do.

Guitar:
"instrumentName_articulation_pickDirection_stringName_noteNumber_dynamic_roundRobin_mic"

Violin (bow direction isn't necessary unless you recorded more than one):
"instrumentName_articulation_bowDirection_noteNumber_dynamic_roundRobin_mic"

Frame Drums:
"instrumentName_drumName_stickName_hitPosition_direction_noteNumber_dynamic_roundRobin_mic"

Orange Tree Samples posted an article years ago about their system. They rely on folder structure for organizing their samples. I prefer to have all the data within the filename so that if I ever take it out of its folder structure I know what it is. But for a Kontakt based library I think a detailed folder structure is often necessary - https://www.orangetreesamples.com/blog/easy-sample-library-mapping-method

Awesome reply, thank you! Also for the Orange Tree blog post, I will definitely check it out and forward it to the team.

For dynamics, are you using traditional notation? (pp, pf, mp, etc.) Or are you expressing a velocity value? Like the top value or something?

Thanks,

 

[d.healey]

For dynamics, are you using traditional notation? (pp, pf, mp, etc.) Or are you expressing a velocity value? Like the top value or something?

I use "dynamic1" "dynamic2" "dynamic3" etc. The higher the number the higher (louder) the dynamic.

HISE's automapper can extract the numbers from these strings and use them to automatically spread the samples across a velocity range - for example all samples with dynamic1 in the name will be auto mapped to velocity 0 - 42, dynamic2 will be 43-92, dynamic3 will be 93-127 (I can't remember the exact numbers, but it does an equal spread).

Or it can auto map them to different groups, or it can do something else. It's also possible in HISE to script your own mapping behavior.

 

[tcollins]

Always use MIDI note numbers in your file names, don't use note names. C3 is that middle C? or is C4 middle C? Doesn't matter, use 60 because that's always middle C. Also when you're scripting it's much easier to deal with numbers than strings for mathematical calculations.

I know this is a thing, but it is lot easier for me to deal with a sample that says D#3 in the name rather than note number 51 (or is it 63?), especially when recording, editing and preparing the samples themselves. Most samplers' auto-map functions work with both anyway.

But, yeah, once you start scripting it's all note numbers.