On 16/11/2011 03:50 AM, Darren New wrote:
> On 11/15/2011 8:56, Invisible wrote:
>> It's actually made up of a bazillion samples, with lots of scripting
>> to tell
>
> That's pretty cool.

Yeah.

In the old days, you'd have a sample of one note, and change the 
playback speed to get different notes. If you were really flash, you'd 
have samples of several different notes, so you don't have to bend the 
pitch too far. Trouble is, then as you run up or down the scale, you can 
hear where the software switches from one sample to another.

Of course, those days are long gone. Each Kontakt instrument has at 
least one sample for every single individual note. For something like a 
piano, where every single note naturally sounds very slightly different 
(and has a different stereo location), the result is very impressive.

In addition to that, you have "velocity layers"; rather than just 
adjusting the sample playback volume in response to harder or software 
keypresses, you actually play a different sample. (Of course, MIDI 
allows 127 different velocities, while a typical instrument has only 
perhaps 6 different velocity samples for each note, so some volume 
mapping is still used to go between. Personally I would have thought 
cross-mixing would be more accurate, but what do I know?)

Then of course some instruments have a characteristic attack, or even 
release (harpsichord). And some have infinite sustain (generally wind 
instruments), so you have a looping sample in the middle. Rather than 
have text boxes to type in sample names, Kontakt has some sort of 
scripting language to control what the sample engine plays. This 
obviously allows a lot more flexibility.

As if all that wasn't enough, for some instruments they record them with 
4 sets of microphones, placed at different distances from the sound 
source. So you can hear a piano up-close and personal, where you can 
hear the sound of the keys, the muffled clunk of the dampers, and 
everything. Or you can hear a distant piano, where the notes are as 
clear as day, but not the mechanics. Again, this isn't filtering, this 
is multiple sets of samples.

On top of *that*, you can have the piano lid open, half-open or closed, 
which obviously radically alters the sound.

So for the piano instrument, you have: 3 lid positions * 4 microphone 
distances * 12 key velocities * 88 notes on the keyboard, yielding a 
total of 12,672 individual samples. (Can you *imagine* how boring it 
must have been to record all of those??) No wonder each piano comes on 
its own DVD! (There are 4 different pianos, remember...)

The pianos also support different tunings - but *that* is done in 
software. (E.g., you can have a piano with Pythagorean tuning if you so 
desire. In solemn truth, it makes scant difference to the sound.)

A similar thing happens with, say, violin. 4 sets of microphones. The 
players play each possible note, as softly as they can. They play them 
all again, slightly louder. And so on, until they're playing as loud as 
they can. Then they play again, tremolo style. Then again, pizzicato 
style. Then again, sforzando style. And so on.

For harp, there is "normal" and "flageolet" (slightly muted). Some of 
the instruments have just one articulation. Some (e.g., harpsichord) 
have only one velocity layer.

>> With the stage now set, it seems an awfully long time since any "new
>> products" came out.
>
> I found Melodyne to be an amazing-seeming program. Not that I've played
> with them.

I'll take a look when I go somewhere with sound capabilities.

>> Just to be fair, I sat down and listened to some of the audio demos
>> for the
>> compressors. I was very, very hard-pressed to actually tell the
>> difference
>> between the compressed and uncompressed sounds.
>
> Maybe I'm confused, but isn't that the goal of an audio compressor?

The goal of a compressor is to very subtly enhance the sound. Which 
makes it all the more amusing to see a "fully authentic emulation of the 
iconic sound of this vintage compressor". If the compressor is any good, 
it doesn't *have* a recognisable sound! :-P

Also, I suspect that if all of your sound sources are algorithmically 
generated, you probably don't *need* a compressor in the first place. 
You can just change the volume envelope of the sound source directly. A 
compressor is more useful if you're trying to record live real-world sounds.

>> Bizarrely, there is now an iPhone version of Maschine. It costs 4.99 €.
>
> Sadly, that is the way of the software these days.

So I can pay a hundred pounds for the software on my PC, or 1/20th of 
that for the exact same program on my phone?

Either one product is vastly overpriced, or the other is an absolute 
bargain. :-P

(Then again, given that an iPhone presumably costs more than even the 
most expensive PC in the first place... maybe not.)

Post a reply to this message