Should I delete all my samples / start a new library?

[Xenon20]

Hi All,

I'm faced with a difficult decision as to whether I should take on the colossal task of reviewing and sorting through 500GB of samples to keep the good stuff, or whether just to delete the lot.

I know about sample managers like Sononym, XO etc etc, but I honestly think that I would be better off from having a small but well organised sample library of quality samples to limit choices and boost productivity.

A second question is whether I will really use them. I mean, there's no point in spending weeks, potentially months sorting them out for them just to end up sitting unused on a disk.

Right now I'm thinking it would be best to to delete them all and feel free of the weight. Then just slowly start reviewing one pack at a time, keeping the samples I like the most and slowly build up a library of a sensible size (like maybe 80 gig, not 500!).

Has anyone else had the same problem? What did you end up doing and how has it worked for you?

[mjolnir]

I'm not sure if this will work for your particular situation, but what I started to do was to create a "TIME OUT" folder for some of my used samples.

Instead of deleting samples I just put them in there. And when the folder gets "full" (or when I get enough new samples that I like, stored elsewhere) I'll delete the entire "TIME OUT" folder. The principle is to get them outta sight / outta mind and ready for the dustbin without depriving myself of the ability to find and use the samples. Also, it makes the list of samples easier to browse since it's smaller than it would be with the (over-)used samples out of the way.

In your instance, you probably want to do something different, but at least hopefully this helps you get some ideas of other ways to handle it. Maybe move the samples to a flash drive for special occasion use?

[Xenon20]

I have them all backed up so I wouldn't actually be deleting them forever. I think the size of the task to sort through them all is just unmanageable to be honest. I think I'd lose the will before the task was done.

What I'm going to do is delete the samples but keep the backup. Then I'll start a new library from scratch. Once I'm happy I've built up a decent amount of quality samples, I'll see how I get on using those, and I can always go digging in the backup if required at any point.

[mjolnir]

All, I get it. That's a good idea. I did something similar with my VST/LV2/VST3 overload.
It sounds like you'll be okay. Please post up your tunes too. Samples are a great way to work, in my opinion.

[eLawnMust]

Yeah... Samples ARE a great way to work, 'Pre-Bounced Instruments' using almost no CPU so you can have a buttload going on...

500 GB is alot of samples but if you are agonizing over it maybe there's something else not musically related on the HDD you can delete instead... That's what I do when space becomes a concern... could be anything perhaps 80-100 GB of porn you could delete...

That's the only bad thing about samples is the space they use & the better they are the bigger they are but in the end be grateful we have such large HDDs these days I remember having a 10 GB HDD back in the late 90's thinking I was really topping off shoot nowadays I'd fill that up in a minute...

For me it's the 'mixes' that take up space... AKAI for SkaleTracker, SF2s for alot of other, Looping samples for this & non-loopers for that, REX, RX2, ACiD WAV for ACiD Pro & Project5...

[mjolnir]

eLawn, I know what you mean.

I ran a scan on my drive to find out what was taking up the most space and it was my audio projects including my mixdowns.
I started rendering to WavPack instead of WAV lately because of that. I had thought that my downloaded tune jukebox would be hogging the space, but it wasn't that. I already save some tunes as FLAC too.

I stopped hoarding the unmentionables earlier in the year. my DAW is for music making and that's that. LOL
but they made some funny jokes about me over at DOA at one point, something about hundreds of tik tok girlfriends contacting me instead of me working on music (which is FALSE). LOL

Anyways, interesting topics.

[jamcat]

The only sample libraries I even have installed at this point are EastWest Symphonic Choirs and Hollywood Choirs.

Everything else is physically modeled: SWAM, Pianoteq, Organteq, MODO DRUM, MODO BASS, B-3X.

That is really the way to go in 2024→2025.

[abi]

That is really the way to go in 2024→2025.

For you.

[jamcat]

It's the way to go for anyone who wants to keep a small footprint without compromising their sound quality or articulations, and actually expand their expressive and dynamic palette, as well as their control over performance nuance.

[eLawnMust]

It's the way to go for anyone who wants to keep a small footprint without compromising their sound quality or articulations, and actually expand their expressive and dynamic palette, as well as their control over performance nuance.

That certainly sounds good, like a commercial advertisement but from alot of what I hear nowadays all of what you stated isn't working for the masses that are using such methods... What I mean is their music sucks! People need to learn music first & this is a big problem... If Hanz Zimmer was using Impulse Tracker today it would blow away everyone using a 'Pro' DAW & the includeds... Peeps now use way too much FX to mask inability for music, take the FX off & it's almost always Playskool...

I use physical models too, older ones & samples too & mine are samples I make myself so quality is not 'compromised' Plus I don't use modern OS can't stand them so XP for me forever... I won't compromise on what OS I use that's for damn sure...

As for articulation much can be done on a sample anywhere in a tune that is just as good as synth generated it just takes a bit longer but uses less CPU for the wild stuff I use HourGlass>>>





I can get sounds outta that I can't even begin to engineer in a synth & in a more human organic way...

But to each, well... whatever...

[enroe]

I'm faced with a difficult decision as to whether I should take on the colossal task of reviewing and sorting through 500GB of samples to keep the good stuff, or whether just to delete the lot.

...

Has anyone else had the same problem? What did you end up doing and how has it worked for you?

Haha, yes Xenon, there is a lot to say about that, here are
my first three thoughts. But I'd rather keep it short:

1. Real samples or loops?

I would like to make a very clear distinction: There are
different types of samples: On the one hand, there are (1)
samples that can simply be used as a loop in a song. And
there are samples (2) that are already finished
multisamples that you can play on the keyboard like an
instrument.

I wouldn't include case (1) in a library at all because they
are too song-specific (tempo, pitch) and don't really fit in
a library. If you have samples like that, I would never
include them in a library. They are only specific to a single
song!

The real samples are only those of type (2)!

2. Format of the samples

Since samples are always at the beginning, they do not
need to have a predominantly dynamic range. This means
that sample formats in 24-bit or 96 kHz are completely
superfluous. I always convert all (multi)samples to 16-bit
44.1 kHz. This reduces the size considerably.

3. Size of the multisamples

If you download all the sample libraries available on the
internet like a vacuum cleaner, you will soon have a huge
library and no longer have an overview. My experience is
that very small multisamples, mostly from the sf2 era, have
often been recorded very lovingly and beautifully and are
much more usable than giga-sized commercial
multisamples. So it's better to look very carefully and
choose!

4. Jamcat's suggestion: Not bad, but wouldn't be
enough for me.

Physical modeling is certainly one way to bring more life
and articulation into the music. But when it comes to that,
I think live recording is even better. At least with guitars,
you should always do that. Because any music that is
meant to be captivating needs to be lively and dynamic.

============================================

Therefore: If you take (1) - (3) into account, you will not end
up in the bad situation of having a library that is far too
large, excessive and no longer manageable.

[jamcat]

4. Jamcat's suggestion: Not bad, but wouldn't be
enough for me.

Physical modeling is certainly one way to bring more life
and articulation into the music. But when it comes to that,
I think live recording is even better. At least with guitars,
you should always do that. Because any music that is
meant to be captivating needs to be lively and dynamic.

Yeah, for some reason, guitars will never sound realistic when sampled. Maybe they can with physical modeling, but so far no one has come out with that. IK Multimedia has owned the trademark for MODO GUITARS for nearly a decade, and yet it still does not exist.

I always record real guitar parts, though samples or (one day?) physical modeling has a limited usefulness for composition. But real guitars need to be recorded for the final version. I do that with bass guitar sometimes, writing and arranging with MODO BASS before I record a live bass guitar.

[martinjuenke]

IMG_9186.jpeg

[No_Use]

My samples hard drive recently died and of course I had no backup.
So in my case the decision is easy.

[Songwriter101]

I'm in a similar situation. Started yesterday to weed out cheap sounding samples. Lot's of GB's but not as much as you have to swift through.

I created a folder called favourites and got enough kicks, snares, claps and snaps already in there (about 100MB not GB).

I think I can delete the rest (about 45gigs) and forget about it.

Keep it simple and functional. Most sample packs contain 2 to 4 sounds per instrument which resonate with you and the rest are fillers anyway.