crazy idea: Open-Source HD Orchesta Library

[supercoco74]

Hi, just thinking, maybe it's a crazy idea, but here it is:
In my ignorant opinion, there is very good sintesizers in linux, but no open source High Quality Orchestal Library (better than Sonatina. like Hollywood Strings, for instance). This may be because such a thing requires a huge investement in money, facilities, musicians, technicians, etc that no small team can do, opposite to sinthesizers. Recently I read about the succesful fundraising campaign Musescore launched in their Open Goldberg Variations Project (http://musescore.org/en/node/9622). With a goal of 15.000 USD, the fundraising has got to 23.000 USD.
So I was thinking that an appropiate Kickstarter (or similar) campaing to hire the studio, the technicians and the orchestra to record and produce an High Quality free-software orchestral sample library (in gig, sf2, sfz?), could raise more than that. Think that it would be something every digital musician needs sooner or later, and it would interest not only linux musicians, but windows and mac ones too. Everyone, in fact. I cernainly would give some money to it.
I have absolutely no idea how expensive that could be. But it would be nice to see it happen!

[bhilmers]

Yes, I would definitely donate to that. I'm working on something orchestral this very moment, and even though the sounds in FluidR3_GM SoundFont are great it is still seriously limited. At the end of every project I always feel like I have a sketch and not a finished product.

[i2productions]

I don't actually think it would be to overly expensive if done correctly. Rather than doing massive fundraising why not make this 'Open Source Library' a completely 'Open Source Project'. Put out a call for orchestral musicians and engineers to donate a little time. As long as the recording of each instrument is done to relatively the same standards, and the final renders were done in the same studio there should be very little loss of quality and consistency. So and engineer in Spain records the double bassoon in short medium and long notes using a set of approved equipment. An engineer in Ireland records the pan flute with approved equipment and forwards the raw data to an engineer in the US that mixes it all down. Just a thought. I'd be totally willing to donate some time in the recording end of this!

[supercoco74]

i2, clever idea!

[varpa]

There is already this project: http://opensymphonia.sourceforge.net/ Its not clear whether anything is actually happenning with this project, however.

[GraysonPeddie]

I'd be very happy to support that. Maybe the OpenOctave team would be happy to include it in the download section of oostudio-2012 and replace Sonatina.

[raboof]

I don't actually think it would be to overly expensive if done correctly. Rather than doing massive fundraising why not make this 'Open Source Library' a completely 'Open Source Project'. Put out a call for orchestral musicians and engineers to donate a little time. As long as the recording of each instrument is done to relatively the same standards, and the final renders were done in the same studio there should be very little loss of quality and consistency.

I think this is a very cool idea, but I expect it's very easy to underestimate the complexity and requirements to achieve a good result.

For example, speaking for myself as a sax player, I guess you'd want the whole range of saxes (at least sop/alt/ten/bar), and it'd make sense to have them played by the same person in the same room with the same equipment. I'd love to contribute, but personally I'd only be able to provide the tenor track. Plus it'd be tricky to audition the musicians...

I think an 'aggregate' project like Sonatina (which takes its instruments from various sources) seems like a good idea, perhaps we should see which of those instruments are not up to standards and improve those specifically, contributing them back to Sonatina?

[i2productions]

The trickiest part would be the sectional and full orchestral sounds. Would actually have to have a full orchestra and probably pay for them and and maybe the space. As far as auditioning for single sounds, I think it's a moot point. You don't need to know anything about the instrument except where each note is really. I realize there might be some difference in timbre and such recording from different sources in different rooms. But if the same engineer does all the mixdown, a good one can make them all sound compatible. Just my 2 cents. I realize it won't come out quite as nice as the millions of dollars put into some of these sample collections, but for the purposes of a good quality free product, I think it would be more than adequate.

[GMaq]

Hi

I somewhat disagree that it can be that uniform a process, which really is unfortunate... I think the attention to detail that raboof is referring to is really what separates a 'good-enough' library to an amazing one especially with strings and woodwinds where timbre and nuance are everything.

[i2productions]

I somewhat disagree that it can be that uniform a process, which really is unfortunate... I think the attention to detail that raboof is referring to is really what separates a 'good-enough' library to an amazing one especially with strings and woodwinds where timbre and nuance are everything.

I know exactly what you mean here. However, I think a project like this could still be extremely beneficial to the community. There's a reason that orchestra libraries cost 100s if not 1000's of dollars to buy. But, I think we could get a database of sounds that would serve the community better than anything we have now.

I've long toyed with the idea of running a site that's a sample hub where users could upload any sound, and anyone else could download the sample for free.

If only bandwidth and storage for something so large were free...

[DoosC]

Aren't you describing http://www.freesound.org/ here ?

[supercoco74]

I have absolutely no idea, but maybe some of you (recording engineers, studio people, maybe someone has worked with large ensembles before) can give an estimate about the amount of money we are talking about. Can the "ensemble" sounds be recorded at least decently in, say, a week? The solo instruments could indeed be recorded by different people, wherever they are, under similar standards, and then sent to the engineer. Even the musicians playing together in the same orchestra have different sound among themselves.
Also, there is lots of faculties, music universities and so, with decent student-orchestras, and have music-technology in their curriculum. It could interest them to get involved in such a project, and would be cheaper than hiring a fully professional symphonic orchestra. Just ideas

[raboof]

As far as auditioning for single sounds, I think it's a moot point. You don't need to know anything about the instrument except where each note is really.

I think that's a bit simplistic. I'm not sure I'd be comfortable recording e.g. a soprano sax part (even if I could get my hands on an instrument), for example - I'd fear especially the higher register might sound thin and likely out-of-tune.

In any case I'm not sure I understand how sampling really works in practice. I mean, there's so much about the intonation, choice of equipment (i.e. mouthpiece), dynamics, articulation - you'd need tons of recordings to capture all of that - and even if you succeed then I'm not sure how a user of the library would get all those sounds out of it again in a usable manner.

I'm not saying it wouldn't be possible to strike a good balance here - I just wouldn't know exactly how myself .

Aren't you describing http://www.freesound.org/ here ?

I don't think so. A sample library would require a rather 'exhaustive' collection of the sounds of each instrument - for example, all notes in the range of the instruments at 3 different volumes and 3 different articulations (just an example). Freesound seems to be geared more towards longer fragments that play some melody or something afaics.

there is lots of faculties, music universities and so, with decent student-orchestras, and have music-technology in their curriculum. It could interest them to get involved in such a project

Definitely something to take into account

[i2productions]

DoosC wrote:Aren't you describing http://www.freesound.org/ here ?

I don't think so. A sample library would require a rather 'exhaustive' collection of the sounds of each instrument - for example, all notes in the range of the instruments at 3 different volumes and 3 different articulations (just an example). Freesound seems to be geared more towards longer fragments that play some melody or something afaics.

I use freesound once in a while, but I often don't find exactly what I'm looking for, nor do they have massive libraries of this kind of caliber well put together. I would be totally willing to help in any way on the development of anything we've been talking about in this thread. I'm not an orchestral engineer, but have miced my fare share of horns. (God I wish I had the money for a good ribbon mic!)

[slowpick]

Starting from scratch is like herding cats, and reinventing the wheel. Failing to support
groups with servers already up and running, and with leadership in place, guarantees mediocrity,
the very thing sought to improve on.

The old addage, 'put your money where your mouth is', seems harsh, but the
essence is, invest in what you love, cut back on the other stuff.

@ i2productions, this ribbon mic review page might be of interest:
http://www.musiciansfriend.com/pro-audi ... one#review

Those who live where there is a Symphony, may find that those who were 'almost good enough'
to get in, might love use their skills in a new way. Local craigslist might be useful to communicate,
since instruments are among the bot and sold items.