I often suggest to people who have strong opinions about this sample set or that- a significant part in virtual orchestration is simply understanding your tools, what we call the 'idiomatic behavior'. Each sample library has its own strengths and weaknesses, and learning to lean into those strengths and away from those weaknesses is what separates a skilled virtual orchestrator from an amateur. Like any instrument, it must be practiced and studied for the best result. As Daru, who experienced my 'Deserted Island Contest' a few years back knows first-hand, sample count doesn't mean much if you're willing to be creative with what you have and adjust your style to fit best what tools/sounds you have available:
That being said, although I have heard many pieces done with freeware which sounded better than some I've heard with $1000 libraries, I don't think it's possible to argue that freeware is perfect, especially at a professional level. It is, at best, an option for those who wish to avoid spending on what may be a hobby or the start of their professional career, or an artificial constraint for those looking for a creative challenge or a unique sound.
The reason in my mind why freeware sample libraries cannot substitute a commercial sample libraries, at least at a professional level, is the development budget available. As someone who has developed both, it's no different than the economy of a budget airline versus the lavishness of a business-class flight. For a developer to economically create a piece of freeware, then to maintain access to it through tens or hundreds of thousands of downloads, to provide demos and support, is no small feat. I make no secret of the fact that my main commercial library, VSCO 2 Pro, cost about $10,000 to produce, much less the cost of distribution, licensing, and of course the cost in hundreds of hours of time.
This cost came out of my own money and luckily there has been a sufficient demand for the product to the extent that I could recoup that 'investment' and make a fair amount of income to support myself and my future projects on the side. Even so, this budget is absolutely minuscule compared to recent top-tier commercial libraries; you would be lucky to hire a pro string ensemble for an day at a fancy studio for this amount. For a major commercial sample library today, tens or potentially even hundreds of thousands of dollars can go into a product because they have a good sense they will get their investment back on that, especially from top-tier commercial film/trailer composers who may have a budget of several thousand dollars a year so they can get the latest and greatest samples. Sales and 'sketching' or 'light' products then bring these samples to the average professional and hobbyists alike.
In addition, due to size, bandwidth is another limiting issue for freeware developers. A not-too-insignificant amount of that income from VSCO 2 Pro went into keeping the free VSCO 2 CE online until I moved it to GitHub. Bandwidth costs recently forced the developer of SSO, Mattias Westlund, to cease distributing that library from his server. Even if they can develop a product on a shoestring budget, self-distributing online is a serious issue. For some, it means dodging between various hosting platforms' free accounts, for others it means finding other sites wiling to mirror and host, for others still, options such as distributing via torrents and so on.
Synthesis and effects/utility plugins are a bit different, as they don't require the massive amount of recording and organization to pull off. Many fantastic synths and plugins are made by sole individuals today, while there are only a small handful of solo sample library developers out there; most are at least partnerships or small companies.
I should note, a new approach I've been trying recently is something I'm calling 'Versilian Community Sample Library'. Essentially with each new commercial product I make, as well as 'test' projects and libraries which are not sufficiently detailed/featured for a serious commercial release, a small portion of the samples get added to this library. It's already at 5 GB and I'll be continuing to add to it over time. I hope to encourage some other developers to contribute to it as well in the future-
Combined with new promising developments in freeware such as those by Airwindows, Christoph Hart's HISE sampler, KVR's annual contests and so on, I think there's at least a decently bright future for those seeking free options. From the triangle of time, budget, and quality, there will always be some unusually kind and dedicated people willing to sacrifice an excess of time to make something budget-friendly and quality.