KVR Audio

pw · Post by pw » Wed Mar 12, 2008 7:14 pm

Architeuthis wrote:ARRITE, WE HAVE A PROBLEM! What are your opinions?

The music box has no sharps. I could record each note two times as much as I was originally intending, take half of the C notes make them C, the other have make them C#.

What do you think? Edit: I'm calling people around the web right now, see if anyone has any advice on alternative musical boxes.

it is not really a problem, you just expand the range of the natural notes to cover an adjacent sharp or flat. No need to have half the C's as "C" and the other half as "C#", and it doesn't give any practical advantage to do so.

Architeuthis · Post by **Architeuthis** » Wed Mar 12, 2008 7:17 pm

pw wrote:it is not really a problem, you just expand the range of the natural notes to cover an adjacent sharp or flat. No need to have half the C's as "C" and the other half as "C#", and it doesn't give any practical advantage to do so.

I'm having a hard time understanding your post. What's the difference between "expanding the natural range" and having half the Cs and C and other half as C#?

pw · Post by pw » Wed Mar 12, 2008 7:23 pm

Architeuthis wrote:
pw wrote:it is not really a problem, you just expand the range of the natural notes to cover an adjacent sharp or flat. No need to have half the C's as "C" and the other half as "C#", and it doesn't give any practical advantage to do so.
I'm having a hard time understanding your post. What's the difference between "expanding the natural range" and having half the Cs and C and other half as C#?

The difference is you don't need to approach it as "half/half". You map all the samples you have for "C" to "C" as the root (because you want to trigger as many samples as possible for realism, right?), and you extend the playable range of those samples to include both "C" and "C#".

Do you have much sampling experience? I'll be happy to help if you need, once you have made and cut the recordings - it sounds like a really interesting project.

Architeuthis · Post by **Architeuthis** » Wed Mar 12, 2008 8:39 pm

Alright now I see what you are saying and it seems like you are suggesting to do it this way because it will be less work, but I don't mind the workload. I think 15 alternatives per note will be good. Playing the music box, you do get some anomalies, however it is very impractacle to try to separate the anomaly from the vanilla because it's so varying. Better to keep the alts random.

Architeuthis · Post by **Architeuthis** » Wed Mar 12, 2008 9:05 pm

After trying the music box on various surfaces, it looks like cardboard produces a nice clean tone. However, maybe that's because it's soft, I have yet to screw it onto a piece of wood and so there's more buzzyness on the wood.

Any suggestions for resonators?

Edit: Also sounding nice is a tub and hollow piece of marble.

Architeuthis · Post by **Architeuthis** » Thu Mar 13, 2008 2:02 am

Ok, so I'm definitely going with the box. Next step is to figure out mic placement. Here's what I'm going to try. I'll post sample recordings tomorrow, maybe some of you can tell me which microphone placement sounds best. However, I have a problem, my recordings still have a mechanical buzzing. The buzzing is not constant so it might be possible to record full length notes without buzzing.

Mics in ear, instrument at arms legnth distance
Mics in ear, instrument almost touching my nose

Mics pointed at instrument, almost touching it, 45 degree angles (above)
Mics pointed at instrument, almost touching it, 90 degree angles (from the sides)
Mics pointed at instrument, hand-length distance, 30 degree angles (above)
Mics pointed at instrument, hand-length distance, 70 degree angles (from the sides)

Mics pointed forward, almost touching it, 45 degree angles (above)
Mics pointed forward, almost touching it, 90 degree angles (from the sides)
Mics pointed forward, hand-length distance, 30 degree angles (above)
Mics pointed forward, hand-length distance, 70 degree angles (from the sides)

Architeuthis · Post by **Architeuthis** » Fri Mar 14, 2008 7:00 pm

Mics in ear, instrument at arms legnth distance
Mics in ear, instrument almost touching my nose

Mics pointed at instrument, almost touching it, 90 degree angles (from the sides)
Mics pointed at instrument, less than arm's length away, 90 degree angles (from the sides)
Mics pointed at instrument, less than arm's length away, 70 degree angles (from the sides)

So listen to these, tell me what you think, I can also do more recordings for example, if you liked the sides_close.mp3 I can try more close positions to fine-tune the recording position.

Edit: Oh, and don't worry about the grinding sound. That won't be in the final product.

Architeuthis · Post by **Architeuthis** » Fri Mar 14, 2008 7:31 pm

Sorry to double post, but this is important.

For those of you who have musical box sample libraries, please post some demos using them or give me some links to those free ones. Yes, I said I've tried them all but those musical box sample libraries are elusive IMO, hard to find, and of course, by "all" I mean some.

Also, I almost certain that the buzzing in the microphones won't be a problem. It is remedied by being careful with the microphones since the buzzing is caused by wires being close to the recorder.

Architeuthis · Post by **Architeuthis** » Sat Mar 15, 2008 4:44 pm

After listening to the Free Boulder Music Box Sample Library, it's possible that the best position would be slightly farther away than the close 90 degree samples to cut off some of the low end.

Edit: Also I want to keep in mind that I want the music box to sound good with an orchestral setting.

Edit2: Oh yeah, I'm getting a mic replacement, it *may* fix the buzzing problem. If not, I know how to avoid the buzzing.

Edit3: You know, I didn't plan on sampling the grinding noise caused by the music box crank, but I find it to be slightly pleasing. I think I'll do a multi-sampled crank!

Edit4: OH SWEET! I just realized, in the sides-close sample you can hear a panning effect! You can also hear it a little bit in the sides-far sample too. Maybe I like the sides-far the best... eh maybe sides close is still better.

Megakazbek · Post by **Megakazbek** » Sun Mar 16, 2008 9:39 am

As for samples with anomalies, I suggest making a preset in which they will be randomly triggered with roughly the same probability they occur in real music box. That way it will be more realistic and there will be no fuss with CC control which always greatly decreases playability. Not sure if it's possible in all samplers, but it's surely possible with Kontakt.
As for the sharps, if you sample some notes twice and use them as sharps, not only you'll be doing more work, you will be also doing unnecessary work, e.g. waste your time. You will still have to transpose the samples up or down a half step to make them sound like sharps/flats, and that's exactly what will sampler do for you automatically if you assign, for example, a single C sample to a zone which consists of two keys - C and adjacent C#. You will get exactly the same result, there's no need to sample notes twice and also you'll save lots of disk space.

Architeuthis · Post by **Architeuthis** » Mon Mar 17, 2008 4:17 pm

Megakazbek, after listening to the music box for a while, it's clear to me that it's impossible to separate anomalies from normal samples. Each time you hit the same note, it sounds different. However, it doesn't sound different enough to say one is more normal than the other.

I'm going to get a replacement for my mics and I should be able to start recording either next week or two weeks from now. Who knows when the replacement will get here.

My final decision on # of alternative is 7 per note. For example, C with be stretched to B and C#. It will cover three notes. 3 x 7 = 21. The second note on the music box is D which will be stretched nto D#. 2 x 7 = 14.

Edit: As of now, I have no idea how I'm going to do round-robin. Seriously.

Planning on creating a musical box sample lib.