Are we composers, or are we collectors and assemblers of different musical sounds?
Composers have since the beginning of time been both. It's just that there are far more resources and instruments (and technology) for us to play with now.
Brass instruments have been around for centuries, but it wasn't until the 18th century that composers started using them more prominently in music (as opposed to being a novelty / gimmick).
Additionally, Ravel used the soprano saxophone to great effect in some of his music long before anyone else was using it (and before it became a staple sound in jazz). There are numerous other examples.
We also live in an age where you have to be able to sell your music. Back in those days, there was no way to play a record, so you HAD to go out and find an orchestra to play your work. Now, anyone can upload a tune to Soundcloud, and if you want to stand out, you'd better make sure you can sell it. It's no different to how it was centuries OK, just the methodology (and terminology) was quite different.
When purely talking with music for film / games / TV, you are just a cog in the machine, and you are trying to accentuate and ultimately sell a product. So - the role of a composer is not just to be able to write the notes, it's about having a vision for the overall sound. That often requires at the very least, a basic knowledge of engineering.
The best and most successful composers are usually the ones who have the best or most effective concepts, and then find a way of hiring the best people (or as many as they can afford) to execute those concepts. Sometimes they don't need to hire anyone, but in the case of orchestral music - it's essential.
Obviously this is hard when you're starting out - it's especially tricky to find a good mixer, but...it's not tricky to find a talented instrumentalist. I hear constant whining about how expensive musicians are, but it's nonsense as they really are not expensive at all - considering what you can get from them. The amount of time (for example) that I save from hiring a guitarist to play some chords rather than struggling with some crappy fake strumming on a guitar sample library is ALWAYS worth it.
I bet most composers are probably spending $1,000 - $3,000 a year on new libaries. Many composers (such as myself) spend far far more. But actually, if you looked at what you spend on libraries and divide that by 2 the following year and use the other half for live musicians, let's see what you'd get. Or, you could do sample libraries one year - and live performances the next year.For example.
Gareth spends $5,000 on sample libraries. So, next year, he'll spend $5,000 on musicians because he already has great libraries having just spent $5,000 on them! So what can $5,000 get you?
$5,000 will get you a 3 hour recording session with 30 musicians in Eastern Europe. This includes the studio space, cartage, and a bilingual producer. Depending on the difficulty of the music, you could probably get about 15-20 minutes (possibly more, unlikely to be less) of recorded music done with a 30 piece string section. No sample library can touch the quality of that, the musicians in Eastern Europe are quite capable.
$5,000 will get you 45-50 hours of recording time with 1 professional musician in Los Angeles. (Or 22-25 hours with 2 musicians, etc...). These figures drop dramatically once you go out of the big cities.
Additionally, music students as music universities / conservatoires are a FABULOUS resource. They LOVE playing on composers' work, are usually really talented, and are nowhere near as expensive as a union musician. I once did a 20 piece 2 hour recording session with students for $600. This is one of the pieces (I recorded 7 in total) from that session - http://soundcloud.com/garethcoker/continuum
And so on....
My point is, if you plan things out not just for the next 3 months, but for the next 3 years, you can find better ways to stretch your finances and resources. I understand that not all composers have $5,000 to drop, but equally if composers are serious about their business, then honestly, they probably need to be spending AT LEAST that on (pick any one or more of the following) new tech gear, new instruments computer upgrades, live performances, mixing, mastering. Of course, none of this guarantees anything, the music and concepts in the first place still has to be good - but when you start hiring people to execute your concepts, a real transformation can occur in your music. Even just adding one real player makes an ENORMOUS difference to the emotional content of your music if everything else is samples. A vast majority of the best selling tracks here have live 'something' on them, be it a vocal, ukelele, electric guitar, acoustic guitar. There is quite often a live element on there.
There are of course, exceptions to everything I'm writing, but I'm citing real-world stuff from an American/British standpoint. Also, if you're a hobbyist, as many others here are - then it's a completely different matter. You guys can spend as much or as little as you like!
HOWEVER...to get back to your original question - if you really do need to do it all yourself, then generally yes, you need to be a good engineer, because that's an important part of the music making process. You're working in an ultra-competitive and crowded business, with both talented composers and total hacks joining every single day - all competing for the same jobs. You have to do whatever you can to stand out from the crowd. It's a lot of work, and it's a pain, but that's how it works nowadays!
EDIT: Sorry, I write way too much. I'm procrastinating from doing what I'm supposed to be doing!