# Topic: OT: How do midi values relate to traditional scoring?

1. ## OT: How do midi values relate to traditional scoring?

Hi all, I have been studying orchestration in more detail lately and this has lead to some questions...

I know this may vary with different sample libraries, but is there an approximate midi velocity value that relates to the dynamics in a score?

For example, I would imagine that; fff = 127 velocity. But how about a simple forte? 80?

Also, since 0 = off, ppp must start at some audible level above zero?

Thanks.

2. ## Re: OT: How do midi values relate to traditional scoring?

Hi Scott,

Certainly a question to ponder. My first thought was that the variances in out of the box sample volumes and instrument articulations - as you said, would make this very diffcult. But then when you think about a live orchestra, you have this same propensity for variances, albiet in a different form. I believe orchestras in practice deal with this in relative terms. Maybe looking at it from relative percentages in midi might be doable. fff is 100% of your loudest boldest parts, forte is reduced by xx% percent from that. Hmmmm...

3. ## Re: OT: How do midi values relate to traditional scoring?

There\'s no standard as far as I know. And there are various sine, cosine and s-curves that are applied to some loudness levels.

But...

As a reference, here\'s what Sibelius does. It\'s in percent, so you\'d have to multiply by 1.27 for MIDI velocities:

fff = 100%
ff = 89%
piu f = 83%
f = 77%
meno f = 71%
mf = 66%
mp = 56%
meno p = 52%
p = 48%
piu p = 44%
pp = 40%
ppp = 34%

I don\'t know if it\'s based on anything concrete, but I\'d imagine that they researched it fairly thouroughly before they chose these values.

However, I find that the dynamic variation isn\'t as strong as I would like in practice. Scaling this such that ppp is closer to 20% with fff remaining at 100% would likely prove to be more dramatic.

4. ## Re: OT: How do midi values relate to traditional scoring?

To paraphrase Piston.... Dynamics are relative. Meaning mf will be different depending on who\'s playing it and who they\'re playing it with.

5. ## Re: OT: How do midi values relate to traditional scoring?

Things like the number of players, the orchestration, the venue, the size of the audience all have an influence on the relative dynamics of a piece. Even the harmonic content of an instrument makes a difference (dull=seems quieter than bright sound). Then, there is interpretation, what went on before and so on. I\'d say the your ear is your best bet! [img]images/icons/smile.gif[/img]

Rudi

6. ## Re: OT: How do midi values relate to traditional scoring?

From a practical, hands-on perspective, I\'ve found that even something you\'d imagine to be as relatively consistent as piano libraries differ in mapping and curve so much that a sequence played on one to great effect can make another equally good library sound horrible.

With more disparate or variable instruments the variation is more pronounced.

So, I don\'t think any formula is applicable. The variations are just too great, and it is probably even better that way.

7. ## Re: OT: How do midi values relate to traditional scoring?

I agree with Bruce. In fact the only Libraries I have that offer that amount of flexiballity( and it\'s why I use them so much) are Dan Deans Wood and Brass. I have been racking my brains working on a 3d diagram for percieved loudness for orchestral instruments. It would show Compass and volume .

ed

8. ## Re: OT: How do midi values relate to traditional scoring?

Thanks everyone. Seems like one mans forte is another mans FORTE!!

I dont suppose you could come to an approximation by decibels?

For instance fff would be closer to 0db?

Ah well, perhaps it is simply better to chose the tonal characteristic of the sample that you like. You might start with a ppp sample that could end up as mp or something else after normalisation, limiting , compression, etc, etc.

9. ## Re: OT: How do midi values relate to traditional scoring?

Well, a full orchestra comes at around 90dB to 100dB. So you could use that as a guide as indicated by John Fairhurst.

When you refer to fff at 0dB, you probably mean the maximum level on your DAW - this indicates a signal strength\'s relative to a reference value, not the absolute actual output from your speakers (which depends on how loud your amplifier is and how close/far you are from it). In that case, the lowest level would be -45dB or so (ppp), and the maximum around 0dB (fff) - actually slightly less as most mixes leave a 0.1 to 1dB leeway to compensate for less than perfect converters on consumer equipment)

There is a page that gives you some examples of sound measurements at:

http://www.btinternet.com/~j.doyle/Mus-Tech/Units/Unit4/Properties.htm

Rudi

10. ## Re: OT: How do midi values relate to traditional scoring?

You know, the fact that Sibelius has mapped those velocities based upon intensity of the orchestra is actually pretty cool, and you could use that to your advantage in a sense.

Of course, it\'s totally up for interpretation, but today\'s samples have a lot of cross-fading. If you set the velocities to say, 86% (being forte or whatever), then you can use the mod wheel to adjust intensity and dynamic from there. It seems like it would work practically, and I look forward to trying that. But, yeah - maybe you can use the basic velocity percentages as a basis for your composition sections, and rely on the cross-fading to get you where you need to go. Thoughts?