Popular sampling conventions that don't sound realistic?

[koolkeys]

I'm just curious here about what different conventions that are commonly used in sample libraries that don't sound realistic or right to you.

Here's an example, and one that I don't understand: panning a piano soundset across the stereo field based on high or low notes. I mean, audiences just don't hear a piano that way, so why sample it that way(or program it)? Yeah, that's how it almost sounds to the player, but even then it's not drastic enough where you really notice it due to all the soundwaves bouncing around inside the piano. So why do people do this?


What other conventions annoy you? Or do you think just aren't accurate or right?

Brent

[liars&ashes]

---Ha-ha, ye olde panning piano trick, I think we can blame 80's synths for that can't we ?

Jeff

[Cordelia]

I'm so glad someone mentoned the stereo piano problem. It drives me nuts.
OK, here's one common one I dislike-
Normalising all the samples in multi velocity drum sample sets. I have to contend with making sure the switches are smooth in terms of velocity of the drum strikes anyway, but this adds another layer of trouble. Now I have to compensate for changing room sound levels, and more often than not (for obvious reasons), the quietest hits end up with the most room sound information.

[gravehill]

I don't see anything wrong in having a sampled piano to use the whole width of stereo field. You can always narrow it down in the sequencer and even flip the sides and pretty much do what ever you want but unless it's sampled that way, you'd be locked to what the person doing the programming had in mind and that might not be the right thing for every situation.

[Chuck E. Jesus]

I don't see anything wrong in having a sampled piano to use the whole width of stereo field. You can always narrow it down in the sequencer and even flip the sides and pretty much do what ever you want but unless it's sampled that way, you'd be locked to what the person doing the programming had in mind and that might not be the right thing for every situation.

+1

[stk]

What other conventions annoy you? Or do you think just aren't accurate or right?

people who think 4 or 5 velocity layers for acoustic drums is okay

[marce]

What other conventions annoy you? Or do you think just aren't accurate or right?

people who think 4 or 5 velocity layers for acoustic drums is okay

You are right. Who needs more than a single layer?

[pw]

Normalising all the samples in multi velocity drum sample sets. I have to contend with making sure the switches are smooth in terms of velocity of the drum strikes anyway, but this adds another layer of trouble. Now I have to compensate for changing room sound levels, and more often than not (for obvious reasons), the quietest hits end up with the most room sound information.

I am interested in this - normalising drum samples is done so that the sampler's velocity->volume settings can work as expected, no? If not normalised, the samples with a quiet signal (quieter drum hits) will be too quiet as the velocity triggering them will reduce their volume even further, making an unrealistic dynamic range, in a relative sense, across all the samples assigned for a particular strike (hi hat or whatever).

Am I missing something here?

[pough]

You are right. Who needs more than a single layer?

For those of us from the Animal school of drumming, nuanced hits simply means you're a weakling.

[koolkeys]

I don't see anything wrong in having a sampled piano to use the whole width of stereo field. You can always narrow it down in the sequencer and even flip the sides and pretty much do what ever you want but unless it's sampled that way, you'd be locked to what the person doing the programming had in mind and that might not be the right thing for every situation.

But I'm talking about being realistic. And panning a piano isn't realistic. It doesn't sound natural, and in any situation I've ever heard it in, it was almost distracting. But I think you may be misreading the post.

I'm not talking about using the whole stereo field. I'm talking about panning it so that the high notes on a piano are towards the right, and the low notes towards the left, meaning that as you play across the piano, the sound shifts as you play towards one direction or another.

I've played the piano for 17+ years, and not once have I ever played a piano that sounded that way. And I'm sitting right in front of the thing.

Brent

[gravehill]

^ Yeah, I know. My point is that having the piano keys spread across the stereo field gives you much more control in how you want the piano to sound. You can still mix it completely mono if you want but going the other way around it's more difficult. But if you want it spread stereo, at least you have the possibility to do it.

[liars&ashes]

But I'm talking about being realistic. And panning a piano isn't realistic.

----Eh, at this point in time, for popular music, a panning piano sound is pretty "realistic" as far as majority opinion goes. Ie = think brainwashing, as in, we all got used to it sounding like that because that's just how the synth/sample players have been doing it for 2+ decades now. As far as for "real" piano oriented music, well, they aren't using samples or synths, so no worries there eh. Certainly there are worse examples of what this thread is about eh, so moving on to other things to bash on ?

Jeff

[dougsyo]

What other conventions annoy you? Or do you think just aren't accurate or right?

people who think 4 or 5 velocity layers for acoustic drums is okay

You are right. Who needs more than a single layer?

You need more layers for when the guy yells "more cowbell".

Doug

[Cordelia]

I am interested in this - normalising drum samples is done so that the sampler's velocity->volume settings can work as expected, no? If not normalised, the samples with a quiet signal (quieter drum hits) will be too quiet as the velocity triggering them will reduce their volume even further, making an unrealistic dynamic range, in a relative sense, across all the samples assigned for a particular strike (hi hat or whatever).

Am I missing something here?

No, you're right, especially when dealing with sample sets with few velocity layers. If you only have four layers for your snare then of course you have to scale the samples, but unless the hits are very close mic'd, the problem of changing room sound volume still exists.
I prefer to use many velocity layers (I can do OK with 16) and turn volume scaling off. Ideally, I like left and right hand hits as well. That means each hit (32 in all) falls within a velocity range of 4.

All of this is kind of moot these days anyway, with so many great drum VSTis.

But for this reason I love my Ludwig sample set, which is set up as I describe, with 16 layers each of left and right hits, matching stereo room mic samples, stereo overhead samples, and none of them are normalised. It's slightly less flexible, maybe, but sounds more realistic to me.

[electro]

Normalising all the samples in multi velocity drum sample sets. I have to contend with making sure the switches are smooth in terms of velocity of the drum strikes anyway, but this adds another layer of trouble. Now I have to compensate for changing room sound levels, and more often than not (for obvious reasons), the quietest hits end up with the most room sound information.

I am interested in this - normalising drum samples is done so that the sampler's velocity->volume settings can work as expected, no? If not normalised, the samples with a quiet signal (quieter drum hits) will be too quiet as the velocity triggering them will reduce their volume even further, making an unrealistic dynamic range, in a relative sense, across all the samples assigned for a particular strike (hi hat or whatever).

Am I missing something here?

A good library will not reduce or increase the volume of any sample.