Create Ensembles from Single Instruments

You can certainly do this if you want to create totally unrealistic electronic sounding music. sorry - but that PDF is rather pointless.

I'll try to explain.

Ensembles from single instruments? I suppose we're talking about real instruments here. If you have a single violin, it sounds a bit weak and feeble. half a dozen sounds great, and big section wonderful - IF - they are real. Real musicians playing strings play with nice sounding musical errors. They play an A. The player at the end may well hit A-440Hz, but the person next door will start on maybe 435, or 445, and once they are playing together, one will feel the sharpness or flatness and gently compensate, but at the same time the other player does the same. Add in a third player and it's yet another different pitch. In terms of timing they will also be hitting each note at slightly different times and slightly different volumes, and even with different tone from bows at different distances from the bridge. Each one also uses vibrato, which has their finger pulsing away at the speed they always use. Some may be a bit more gentle with it, others more harsh and distinct. This is what makes a string ensemble real. The electronic slip a bit one way or the other is a very dysfunctional solution as it caters for just one thing - an advanced start or a delayed start. It cannot advance AND extend a note. It can't introduce vibrato, it cannot introduce subtle pitchblende, it can't introduce modulation or other individual playing features.

As a way to create ensembles from single instruments it's the kind of thing 1st year students do in music technology before they discover why it sounds naff!

I've been producing string sections this week, and frankly the best way is to play the same music 6 times to get 6 different versions of the same music, while using volume, expression and pitchblende at the same time. I'd love to open a box, click on something and have it sound real - but it just doesn't happen. This way just produces comb filtering and a kind of chorus effect as the tracks merge, which is an interesting effect, but not found in real life. Real ensembles sound amazing because they are all evolving around each other.
 
The entire point is that musicians deliberately adjust their tuning on fretless instruments continuously - and it is this that create the sound of an ensemble. you use a finger tip to form the note, and then you wobble - some do it better than others, but we're not talking out of tune, we're talking microtones - and delay and fixed pitch shifts do not do this. If you spend thousands on sample libraries, the quantity of samples is huge and you may have 5 different violins, 4 different violas - then same with cellos and basses - you do not always select the first one and repeat it. They need to be different, that's my point. sorry to grumble away on this, but creating the sound of multiple players needs effort. You can record string parts using less good samples or synths and few people ever bother to use the pitch bend wheel while they play. This can make them come alive. Real instruments have science behind them. you sometimes hear pitch bend on clarinets and oboes - clarinets can pitch bend, oboes can't (well, a tiny bit really) but people simulate them badly. strings have fixed open perfectly tuned notes, and then everything else will be off - just a bit. Some notes will slide to the next, others will have gaps. vibrato comes in after the note has formed - but how long? Every player will be different. Ensembles are just groups of differently sounding instruments from exotic to student, and every player does it differently. Each one listens to the others - hence why they find one eared headphones essential. However, the biggest errors come from looking at the lines played with the key editor most DAWs have. Is every single note velocity and note length different? If they are all the same, that's a mistake. if you have a slow sounding patch or sample, then they need to be played slightly ahead of the beat. A more strident sound later. As a piece travels from ppp to fff, their beat placement changes. Is a drop from F down to E two notes, or one note with a semitone pitchblende suddenly applied? Does the violin sound play below the lowest note of the real thing? Beginners really abuse this one. A viola sound goes lower, but some synths let the violin sound go lower than real life and the tone of the string section changes as the content gets lower.

All this stuff is obvious to a string player - and alien to others. To create realism takes work, and money - let alone time. There is no preset you can click in Cubase to make a solo line an ensemble. you might just as well add a chorus effect, it'll work just as well. If you want a realistic ensemble you need to be able to play keyboards and have patience. Even if you spend 900 quid on a spitfire wonderful string collection, you HAVE to play them!!!
 
nope - if you ever play a string instrument solo, your tuning can be off by quite a bit unless you are blessed with perfect pitch, which few are. My double bass and electric fretless basses are good examples. If you are playing with few open strings because you are playing higher on the fingerboard, it's very easy to be half-way between say a B and a C. As soon as you play with other accurate instruments you instantly snap back to concert pitch, but usually the few open strings will keep you locked, but micro tuning is the norm. I do some work with an ELO tribute and on stage the string players - usually just two live ones with the rest on track as required have quite serious issues with tuning if monitoring isn't perfect. Real orchestras sound wonderful because of it, and it even spreads a little to things like big bands, where all the saxes will be nearly in tune, but again, each player will tweak their tuning with embouchure changes to stay in with their neighbour. Even multiple timps in the percussion sections of the big orchestras are all slightly differently tuned and this helps them sound bigger too. If you record in stereo with accurate techniques, you can actually detect the tuning differences as a left-right spread. String tuning is clever stuff.
 
A bit of a joke, but I always loved the way Brian built his tracks to imitate various "ensemble" sounds with layering. The thought of trying to create an ensemble mathematically by adjusting pitches and timing in midi is totally foreign to me.
 
Many moons ago I was Principal Examiner for A Level Music Technology, and sequencing wise (which was pretty well MIDI only back then) the students had to do a pop piece - typically recreate Abba, Gary Moore, Lionel Richie, Santana and stuff like that - which was done pretty well, and a classical piece of music which was universally dreadful. Even those with nice sounding strings and orc gestural sounds rarely got their 'orchestra' to work properly. just a few did, and when you picked it apart, if there were 6 violins, there were 6 different MIDI tracks - all unique. Quantise, copy and paste and pitch and time shifts never worked properly. Quite a few said things like "I played in the string part by hand for realism and then because my timing was poor, I quantised it! One I remember well - she'd left the old versions of the file in a folder, and while we weren't allowed to mark it, I did play it - the unquantised version and it was hugely better. Sloppy playing, but it would have earned marks. All the machine tweaks we now have to humanise rigid playing only scratch how bad a professional musician is at playing in time and on exactly the right note. One of those musical rules. Thou shalt not play out of tune or time unless it makes it better.
 
Back
Top