Blend and Balance

I have quite a few long-standing problems related to “Real dynamics and MIDI / sample dynamics” and “MIDI / sampled orchestraal blend and balance” Let me explain :

1. I have seen that the past masters like Beethoven almost never employ different dynamics to different instruments at the same time. When it is pp, it’s pp for all the instruments. The conductors also find it easier to handle. I follow that as a rule. do you do that too?

2. You must have seen that the balance and blending change drastically if I change the Pan setting or the channel settings of any ensemble. As a solution, I set Pan at 64, which is the exact middle in digital reproduction. Professionals do likewise. But what about channel ? I have seen a single flute in its middle register penetrating an entire string section just because I used a different channel setting. What is the solution ?

3. Sampled sounds are more real, but do they take into account the relative dynamics of different instruments ? For instance, any orchestration book will tell you that a tuba is much heavier than a contrabass at any given dynamic level : can I rely on the samples to replicate that in a composition ? What about the GM MIDI ?

These are very important issues if one wants to have his pieces performed by professionals. What you hear in your MIDI version may sound entirely different when they play it by just following the score and parts. Ideally, that should not be the case if my composition is based on sound orchestration principles. But in that case, while the performed version may sound right, the MIDI version may not!:P

I don't personally follow #1 as a rule, no.

About #2: the flute probably drowned out the strings not because of the MIDI channel change in itself, but most likely because of it being moved to a channel whose volume was greater than that being used for the strings. This might have been less-than-transparent if you were using software like Sibelius and something weird happened, because normally you don't need to even look at which channels the music is being played through, it just does it all for you, whereas with a sequencing program you often need to assign them yourself.

As for relative dynamics, I can't say I've ever come across any set of sounds that significantly varies the "dynamic centre", if you will, of different instruments. Certainly not in General MIDI, anyway. The solution is probably to have at that volume control. :P

Generally speaking, learning to orchestrate properly and write music that could actually feasibly be performed by humans just takes a little time and experience. I can even look back on stuff I wrote a year ago and immediately spot things that need attention. I suppose books may help you in this regard, if you want to expand your knowledge.

Perhaps I couldn't clarify what I want to know. This is not about books - I had read some classics on orchestration including Blatter and Piston even before I begun to compose 5 years ago. I am still reading wherever I find anything on this topic. The problem that even seasoned men face is : Even if you write a perfectly orchestrated and instrumented piece, would the MIDI version give the expected blending and balance ? Remember - by "perfectly orchestrated" I mean a piece which, when performed, will sound just the way it was intended to sound in terms of blend and balance and color.

The short answer is 'no'. Hope Mike pardons me if I get this wrong but I don't think he was talking about books.

For one thing, perfect orchestration implies a perfect orchestra and, well, there are some excellent orchestras about but you'd need to be reasonably well-known before you could get your music played by them - unless they specialise in new works or unfamiliar names. Most works I've heard performed by different orchestras in different halls have a different "colour".

M

Shekar, there is a simple answer to this question. Given the difficulty in creating an accurate "sampled" orchestra, it is impossible and impractical for the producers to have things with appropriate dynamics relative to each other.

The reason is that an orchestral sound library must be versatile enough to cover every use of the instrument from the middle of an ensemble to solo use.

As a result, they sample everything at the same volume levels and do their best to make everything as versatile as possible so that the user can tweak settings to achieve that sort of realism. They're assuming that if you're spending hundreds (and often thousands) of dollars on this package, that you know enough about how a real orchestra works to be able to set things up the way they should be in a real performance.

Remember also that many of these sampled libraries are meant to replace a real orchestra. They are made to sound so realistic and so versatile that you can make a perfectly convincing recording without a real orchestra.

In that case, a flute could easily be louder than a string section if it was specially mic-ed. For many users, details such as these are irrelevant since they're not writing to be performed - the studio recording is their final product.

Anyway, my point is that they need to make the sound libraries work for everyone's needs which sometimes leaves others out. In your case, you need to consider a real orchestra and how instruments relate to each other in a real concert setting and tailor your recording to the orchestra that will be performing it. This may mean a lot of tweaking on your part, but that's just the way things go with sound libraries.