It becomes more apparent when you listen to the soundtracks than when watching a film or series. I should have listed some examples, but I've already deleted the folders. We're still far off from samples being able to sound just like the real deal (although it's getting closer every year), so it just sticks out to me. It's one thing to add a few sampled instruments on top of something that is more synth-driven, that's OK. But when it tries to mimic a big, old-fashioned orchestral pirate score or whatever, with zero dynamic, abrupt ends to chords and so on, it just grates on me.
I completely agree with that last sentence – I'm open to accepting any form of music for what it is, but when it falls so short of what it's aspiring to, it's much more frustrating than something that sets its goals more modestly but achieves them.
But your first sentence probably goes a long way to explaining why it's done this way. If the shortcoming is not so apparent in the film, well, that's the score's primary function, isn't it?
Nobody likes real orchestral scores more than me, all of my favourite film composers used them and there is nothing out there that will equal the quality of their sound, but, and it's a big but...
...there is a danger here that we could morph into similar beings to those who deride or look down on film music as something lesser.
If I happen to like a score produced on keyboards that simulate real instruments and if it does the job it is supposed to do I have no problem at all with it.
I have to be honest. Having streamed countless films/series at home over Covid time I can't recall a single case where I sat back and thought, nah, that's no good because it was played on keyboards rather than with a real orchestra.
I have no problem with non-orchestral music, it can be synthesizers, it can be electronic, it can be chamber music, it can be a jazz trio, it an be all kinds of things. What Thor is referring to is though something that just tries to sound like an orchestra when it really isn't; sampled orchestras (as of yet, that may one day change) sound lifeless, they have a bland, clinical sound to them.
I've deleted the scores in question now, but some off the top of my head include EMILY AND THE MAGICAL JOURNEY (Kraemer), NINJAS DOWN THE STREET (Kieboom), PLUNDER QUEST (Sammi), THE LOST CITY (Toprak), THE GREEN PLANET (Merrison & Slater), QAZAC (Silliotto), OUR GREAT NATIONAL PARKS (Schweitzer), BLASTED (Westad), others I can't remember right now.
I think a lot of scores match the content now which is a lot of mass-produced streaming junk. These examples share some of that in common. They're either random streaming film content or television shows. But a lot of them just aren't great compositionally to begin with, which is going to make things worse.
EMILY AND THE MAGICAL JOURNEY sounds the worst, and really sounds like Kraemer used samples from 15-year old music libraries he uses to make mockups. The film also seems like a random low-budget streaming piece.
NINJAS DOWN THE STREET uses slightly better samples but the compositions aren't that great - skim through the playlist and note how almost every song starts with sustained strings or features taiko... It's also for an equally low-budget production.
PLUNDER QUEST also looks like some bad independent production and apart from using poor samples, the music is also just bad. It's sort of a copy of Bruce Broughton doing a pastiche of a western, so effectively a pastiche of a pastiche... never a good idea.
The others are passable and bother me less because they're for documentaries: THE GREEN PLANET appears to be for a planetarium and OUR GREAT NATIONAL PARKS is expected fare now for over-scored documentaries.
LOST CITY actually sounds decently realistic, but it features extremely conventional Remote Control-esque compositions for everything from the action to the comedy. The comedic stuff in particular sounds the worst and falls into the same compositional traps I hear frequently in comedic scores in low-budget independent films. But I also hear Jeff Richmond-30 ROCK rip-off, so again when it comes to composition it's just not that great.
I'm not a proponent of synthetic orchestras - I think they're a means to an end for composers starting out... but ultimately you need to be good at COMPOSITION. Synthetic orchestras can be wrangled to sound pretty good, but it requires time, skill, and attention to detail. A lot of composers now don't focus on composition but rather the tools of the business, so they jump into programs with synthetic libraries while having very little knowledge of how to compose. Even a live orchestra can sound bad if the composition is poor.
A big part of this is the orchestration. If you have no knowledge of how actual instruments work, you'll write stuff for instruments that would be how a keyboardist plays, not how the actual instrument player can/would play (you hear this a lot in Remote Control music). There's also improper use of instrument groups that are common with inexperienced composers, which synthetic orchestras exacerbate. You really need to have the knowledge and spend the time to properly orchestrate your music.
Another part is the mix, which most of Thor's examples suffered from. It takes time to mix a synthetic orchestra and in many cases the samples are already "pre-mixed" so a lot of people won't bother. But if you really want to get the most from a synthetic orchestra, you have to focus on mixing which a lot of inexperienced composers also don't have the skills for (and on these projects, the time either).
The expressiveness of a real orchestra is also a huge challenge. Some sample libraries are really good, but it takes A LOT of detailed tweaking and finessing, and frankly nothing beats real musicians playing their instruments with the variety of expressiveness and dynamics that come naturally.
And no mention of the PERFORMERS. A sampled voice made to sound like Laurence Olivier doesn't mean it can present any of the unique nuances and performance skills of the actor himself. But people who aren't particularly bothered about that might say "Wow, that sounds just like Olivier performing Hamlet!". Same with instruments. Does a sampled violin allow for the nuances and performance skills of the performer of a real violin, or the superb craftsmanship of the instrument maker? Of course not. Only those with a peripheral awareness of those aspects of the art, would think so. Maybe if 20 of those sampled voices of Olivier were massed together, they would sound something like a bland male choir. Just like 20 sampled violins together might sound like a mass of expressionless violins. Sad.
I would rather LISTEN to a score performed by an orchestra (unless that score was written specifically for synths, such as NO MAN'S LAND). But that is only because, as a musician, I enjoy listening to a bunch of other musician's playing together. That's part of the charm of listening to film scores for me.
A sampled voice made to sound like Laurence Olivier doesn't mean it can present any of the unique nuances and performance skills of the actor himself... Maybe if 20 of those sampled voices of Olivier were massed together, they would sound something like a bland male choir. Just like 20 sampled violins together might sound like a mass of expressionless violins.
There are music libraries that sound pretty good and have realistic expressiveness BUT - you can't beat the real-time performance of a live orchestra.
You can tweak and prepare the synthetic music to get it close - you can even play it on your keyboard with a bunch of crazy keyboard shortcut switches as you play to cycle through the different expressions... but no matter what, you can't get a synthetic orchestra to be played and recorded in real time.
Just like your example with Laurence Olivier - it may fool some people, but getting there overall will require a lot of time and pre-planning that doesn't occur in real-time. The magic of what happens in-the-moment with real people is a level of unpredictability coupled with the instinctual responses to real-time feedback. You can't get that organic liveliness digitally.
I'm not a proponent of synthetic orchestras - I think they're a means to an end for composers starting out... but ultimately you need to be good at COMPOSITION. Synthetic orchestras can be wrangled to sound pretty good, but it requires time, skill, and attention to detail. A lot of composers now don't focus on composition but rather the tools of the business, so they jump into programs with synthetic libraries while having very little knowledge of how to compose. Even a live orchestra can sound bad if the composition is poor.
A big part of this is the orchestration. If you have no knowledge of how actual instruments work, you'll write stuff for instruments that would be how a keyboardist plays, not how the actual instrument player can/would play (you hear this a lot in Remote Control music). There's also improper use of instrument groups that are common with inexperienced composers, which synthetic orchestras exacerbate. You really need to have the knowledge and spend the time to properly orchestrate your music.
Another part is the mix, which most of Thor's examples suffered from. It takes time to mix a synthetic orchestra and in many cases the samples are already "pre-mixed" so a lot of people won't bother. But if you really want to get the most from a synthetic orchestra, you have to focus on mixing which a lot of inexperienced composers also don't have the skills for (and on these projects, the time either).
The expressiveness of a real orchestra is also a huge challenge. Some sample libraries are really good, but it takes A LOT of detailed tweaking and finessing, and frankly nothing beats real musicians playing their instruments with the variety of expressiveness and dynamics that come naturally.
Over the past couple of years I've been teaching myself orchestration and composition, while picking the brain of a talented orchestrator friend and studying scores to the best of my ability (I play by ear and I read music on a remedial level, which I plan to change soon by starting music lessons). I've been composing using a keyboard and many different sample libraries, some of which are free, some low cost, and some subscription-based.
Having said that, I agree 100% with what you say here and want to emphasize your point about programming, equalizing, and mixing the sample libraries so that they blend together and "breathe." It takes an INCREDIBLE amount of time to get the dynamics and expression right, and I often have to use several different tracks of articulations to get things just the way I want.
Danny Elfman on his Masterclass said that long strings are always going to sound synthetic, a statement with which I mostly agree, depending on what you're writing. I've been pleasantly surprised by a couple of brass libraries I've been experimenting with, and piano and percussion are pretty convincing. Woodwinds I've found to be a mixed bag. A violist I know recently told me her quartet will often get mockups from composers that sound super realistic, though, and she sometimes wonders why they would even bother hiring real players, so it does seem possible to program at a very high level if you have the skill (and the time).
Regarding the larger question of why, I think it's related to budget and the fact that there's SO MUCH programming coming out that sample libraries make it faster and more convenient for the production process. The emphasis on "content" (I hate that word) has only amplified the longstanding problem that the entertainment industry functions like a machine, cranking out movies and shows and games on a relentless assembly line.
A sampled voice made to sound like Laurence Olivier doesn't mean it can present any of the unique nuances and performance skills of the actor himself... Maybe if 20 of those sampled voices of Olivier were massed together, they would sound something like a bland male choir. Just like 20 sampled violins together might sound like a mass of expressionless violins.
There are music libraries that sound pretty good and have realistic expressiveness BUT - you can't beat the real-time performance of a live orchestra.
You can tweak and prepare the synthetic music to get it close - you can even play it on your keyboard with a bunch of crazy keyboard shortcut switches as you play to cycle through the different expressions... but no matter what, you can't get a synthetic orchestra to be played and recorded in real time.
Just like your example with Laurence Olivier - it may fool some people, but getting there overall will require a lot of time and pre-planning that doesn't occur in real-time. The magic of what happens in-the-moment with real people is a level of unpredictability coupled with the instinctual responses to real-time feedback. You can't get that organic liveliness digitally.
Make no mistake: artificial emulation of orchestras through samples and synths has come a long way, no doubt, and I can definitely see how they are useful and give you the impression of a composition, but they are still a long way from sounding like an actual orchestra, recorded together in a room. Sampled orchestras, even the best, just give you an approximation.
JEEPERS CREEPERS 3 by Andrew Morgan Smith fooled me into thinking it was a live, acoustic, orchestral score, when in fact it was performed with samples by AMS. So they've already got me once. Even now when I play it, with the knowledge, it STILL sounds real.
JEEPERS CREEPERS 3 by Andrew Morgan Smith fooled me into thinking it was a live, acoustic, orchestral score, when in fact it was performed with samples by AMS. So they've already got me once. Even now when I play it, with the knowledge, it STILL sounds real.
Yes, the libraries that were used are good quality. This type of music is easier though because it's mostly percussive (literally with percussion and then playing the winds/strings/brass percussively. It's also a challenge because a lot of these rips and runs are pre-templated, so once you play them they would sound exactly the same again.
What stands out the most though is a slight mis-match in mixing between the elements, likely because they were using multiple libraries from different sources. The weakest stuff is when things quiet down and they play more traditionally, like at the end. The strings really stick out there, and you first notice them at 1:59 - very synthetic.
The composition is a bit frenetic but overall it plays well to the strengths of the sample libraries.
Regarding the larger question of why, I think it's related to budget and the fact that there's SO MUCH programming coming out that sample libraries make it faster and more convenient for the production process. The emphasis on "content" (I hate that word) has only amplified the longstanding problem that the entertainment industry functions like a machine, cranking out movies and shows and games on a relentless assembly line.
Time is a huge factor, also because studios want to hear mockups in advance so now it's far easier to take that rough draft mockup and spruce it up slightly to be the finished work.
This isn't the best example because I made it in 45 minutes as part of a response to a thread about THE ROCKETEER, but overall it sounds reasonably realistic:
Again, not the best example because I didn't mix it and there are little things here and there that I let slip by, but if this was for a real project, it would be very tempting to just say that was good enough and move on. And the more "content" that gets cranked out, the more the craft of music scoring will slip into quick, synthetic, templatized work.
