Composing : EWQL Hollywood Orchestra by Joel Irwin

I've got different templates for different kinds of projects. I still keep my HO template current, even though I haven't used it for a few years. Bottom line is that the sounds are incredible, even given its age (dude...Shawn Murphy). The downside, for me, is that I have to use a separate track for every articulation (and you might want to see how that translates to Sibelius, putting different string articulations on different MIDI channels). My workflow is faster, and somewhat more organic, when I use TouchOSC/keyswitches to change articulations (one instrument-one MIDI track). But HO Strings, Brass and Perc are still contenders with more recent libraries (WW not so much, imo).

As far as getting a more realistic sound without doing complex midi settings...you're going to have to do those with any library. There's no silver bullet for not putting the time in. I use controllers (WX-7, TEC breath controller, Handsonic, V-Drums, sliders, modwheels, keyboards, iPad/TouchOSC) to record every track into a DAW (DP) the way I would if I were recording studio musicians. If you look at the MIDI data from my Flute 1, for instance, there's a whole lotta MIDI going on. If I had to draw that out, it'd take some time, and it'd take some skill/experience in knowing how to translate MIDI data into a believable performance. If I perform the flute part on a WX-7, it may take me a few takes to get it solid, but I know the MIDI is a faithful translation of a wind player's performance, cuz I used a wind player to create it. If you don't go the composer/performer route, then you gotta put in the time to draw out the MIDI data. Until some better AI comes along, that's what's needed to make it sound musical.

I've matched HO with other libraries, but it takes spatial plugins. The process is to play a note on a HO instrument and then play the same note on a comparable instrument you're using from another library (using a closer mic, instead of one with baked-in room sound, or a combo of both). Then modify the space and placement until they sound like they're sitting next to each other in the same hall. I use MIR PRO, but I've also used Spaces (which is prob'ly your best bet since HO uses some of the same IRs), Proximity, Valhalla, etc. The trick is to get the early reflections and the hall reverb to match. That's a whole discussion in itself, but that's the goal.

As far as working in an engraving program for music production, if I had to do it, I'd record in every instrument's performance, rather than enter the notes through mouse/keyboard. But that's me. If you step-entered them, then you need to make sure the MIDI data sounds musical, by drawing in the expressions and tweaking the velocities.

One caveat to what I've said above, re: artificial intelligence and Sibelius, is the program NotePerformer 3. They've got a 30-day free trial you might want to check out. It still doesn't reach the production-grade I'd need to deliver to a client, but imo it's a great sound for a demo. And who knows, maybe by NotePerformer 5, it'll give you that silver bullet. https://www.noteperformer.com

The Diamond version of HO is the way to go, imo. When I used to produce studio soundtracks, I'd use the LsRs stems as you might a reverb send. You can use the surround channels in Diamond in much the same way, and with much the same effect.

Spitfire is awesome as well. It's my go-to template more often than not, especially the Chamber Strings as well as some of the Albions (Tundra, Iceni, and Uist). For me, it boils down to workflow. I can load in all the Chamber Strings and use TouchOSC to trigger different articulations. So instead of 16 MIDI tracks for Violin I articulations (which is what I'd load in for HO), it's only 1 track. And, I can usually play in a whole cue without stopping, by triggering the articulations as I go (unless it's really complicated and I need to punch in). There are arguments that it's better to have the articulations on separate tracks, so it's up to your personal needs. If you're handing your file off to an orchestrator then separate tracks are definitely cleaner. Also, for mixing/balancing, a case can be made that separate tracks give you more control. I'm of the school that if I can perform the part in a balanced fashion, I'm good to go. And having one track is a cleaner/faster way for me in the end.

I only have some of Cinesamples specialty instruments; no full-orchestra template. But I've heard some great demos from them. I mean, it's the Sony stage, right? Best stage in the world if you're a Goldsmith adherent. You might want to pick up a TEC breath controller and see what it does to your strings dynamically. My personal philosophy is that the best way to breath life into an electronic instrument is to play it like an instrument. Not that I haven't heard some great tracks that have been drawn in with a mouse. But those composers definitely have "MIDI patience."

Also, if strings legato is your main concern, you should check out Cinematic Studio Strings. https://cinematicstudioseries.com/strings Some of the best legatos out there. They also have great brass and will hopefully be releasing their woodwinds this year. The legato can be a little tricky in timing, but it's worth the effort. Combined with their Solo Strings (as first chairs), it's a killer sound. I use these strings for darker material, like in my score for OUIJA HOUSE (after the minute-mark: https://soundcloud.com/jonathanprice/ive-been-good).

I've always wanted to buy CSS, but haven't done it yet. A few weeks ago Spitfire had the Studio Strings on sale and I bought the basic package for $125 and I'm happy with them. next sale I'll upgrade and get all the mics. I did this the day after I got them. I needed to write a string arrangement here to a pre-recorded group. I locked DP to sibelius and wrote the chart then export the midi into DP and overdubbed cc. For 125 it was a bargain. (http://linwoodbell.com/clientdownloads/Esta%20Tarde%20Vi%20Lover(2).mp3)

Jonathan - your feedback is awesome not just for me but for everyone. I get what y'all are doing. I wish there was a way to sketch things out to sheet music like composers have sketched on paper for decades and then let the orchestrators provide the live performers basic information how to play and articulate leaving some room for the conductor to provide last minute guidance based on the instrument and microphone placements in the hall and the way he/she hears the sound in rehearsal.

But we are not there and creating such higher level 'expert systems' (I did my PhD research in expert systems back in the 70s) will likely imho remain embryonic perhaps this whole decade. Our eventual goal is to provide basic sheet music notation with textual instructions and let the players figure out what to do with having 10+ tracks with dozens of midi instructions which are specific to the sample set we have purchase. Imagine for example, the difficulty of say scoring for cinesample strings and then deciding to use spitfire strings instead.

My current philosophy is quite simple - if I were scoring a feature with a significant budget and the possibility of live performers, if the score was being mixed loud enough that the listener could hear the difference between good strings and significantly more realistic strings, between good articulations and really intricate articulations, I would spend the time doing exactly what you are doing. The questions becomes an economic and time one for me - what is the best use of my time and money. The filmmakers I currently deal with have little to no budget, my 30+ projects have all been shorts and my customers (filmmakers and viewers) already like the sound. I don't and my teacher/mentor does not. Some of it I can fix through using multiple channels changed through 'programs' (P1, etc...)which Kontakt, for example implements using 'instrument banks'. That enables me to use a single or small number of staffs and to change instruments via 'programs'. Some of it can be done with articulation control. But I don't currently have the time and frankly the interest, to 'tweak' the sound with advanced commands via the controls, sliders, etc. The 'incremental' improvement may not be heard in the final mix and my economic benefit of spending two, perhaps, three times the scoring time is not there.

One may argue that my implementations are currently not in the ballpark. I would think, that perhaps if that is so, it could be the original composition and voicings, it could be the arrangement the way I spread the sound across a single instrument to a whole section. I spend a portion of my time improving in that realm but it is I (and my teacher) who are no satisfied with the sound.

So in the time I now have (perhaps months) to take a score and find ways and approaches to learn and improve how it sounds, I am willing to invest (though not for Vienna :) ) some money in trying out and experimenting with other sets to see if I can at least currently get a more realistic sound with perhaps less 'effort' in having to work with MIDI to get it to sound the way it 'needs to'. Now is the time to invest perhaps in a product or products than CAN be used the way you describe but also have the basic functionality and 'sound' to be more realistic with a minimal amount of tweaking (mic placement, room shaping, articulation, etc.). But what I am not currently going to do, for example is to exactly control the bowing - for example, the difference in sound between and up bow and down bow. I may hear the difference and like it but my customers won't hear it and if they do - will not expect me to spend a lot of time doing that.

BTW - I have had Note Performer since version 1. I used to use it to demo my jazz lead sheets but for the last year or so, found it easier just to give it to the ensemble to read. They usually nail it by the first or second try and don't need to hear anything.

Here is a postscript - I never bothered to check the resource requirements - they are enormous. Again I purchased EWQL Hollywood Orchestra Diamond plus Solo. Here is the data:

Folders - 51,124

Files - 1,342,087

Space - 749 GB (804 GiB)

My machines had two 1 TB internal SSDs each. I don't have enough space on one SSD for the files I would need to clear out on the other one to make space. So I need another 2 TB SSD.

Here are my options -

1. If internal, should I go again with an ATA and swap a 1 TB for it. Or should I go with something like a SABRENT Rocket 4.0 which has a very high Passmark number, 32,356 (https://www.harddrivebenchmark.net/high_end_drives.html). Seems like if I have an available slot, the SABRENT PCI solution would be faster and allow me to keep the other 1 TB SSDs.

2. Or should I go with an external SSD solution. I would assume there would be a performance hit as the drive goes through one USB port (I could not for example find a Passmark result for the new Samsung T7 but I would assume its 5,000 to 10,000 at best). However, the tech support people say they use an external Samsung T5 SSD (the newer and faster T7 came out last month) and it works fine even if the samples are used/cached directly from the SSD and not first put into memory (though both of my machines have some room since they are 32 GB memory each). Using an external SSD would allow me to share between both machines without having to also look for a storage solution for my laptop - though an iLok solution (versus transferring the license back and forth would require two USB ports.

Any opinions?