• Register
  • Help
Page 1 of 2 12 LastLast
Results 1 to 10 of 13

Topic: Dumb Question

Share/Bookmark
  1. #1

    Question Dumb Question

    I've been rescearching GS3, and I was wondering--is this the type of program where you create a VSL or a GPO or a GOS or something like that? I think that's the case, but I just wanted to make sure.
    And about that hard disc streaming stuff: would that work through a USB hard drive? How big of a hard drive would I need for a full orchestra? How would I route MIDI through GS3? I don't have a VST sequencer, only a somewhat old but trusty sequencer that supports only DXI.
    Any help would be greatly appreciated.
    Oh, and I'm running a system of Win 2000 pro, Pentium IV 1.8 ghz, 1gb RAM, 20 gb internal hard drive, 120 gb external hard drive (music drive).
    Will GS3 work on my computer?

    Chris

  2. #2

    Re: Dumb Question

    Caveat: I've sampled some stuff, but haven't made anything very big or even moderately commercial.

    I think the most important software for making a lib is the editor: SoundForge or something equivalent.

    At the front end you do your sound design, planning, getting equipment, a place, talent etc. When you start the recording process, you just need something that can capture the audio. Vegas is my favorite for capture right now. Sony offers a 30 day demo. You don't need anything special here, just something that will reliably capture all the tracks needed, and that provides features that make it easy to catalogue things.

    The next step is the chopping and processing. Here's where the editor comes in. Anything that can process in batch or automate the process is potentially good. But only if it really works accurately and realiably. Make sure to only process copies of your raw captures. Don't lose your originals. Keep each step of the batch process in macros where you can, so you can re-process, if needed.

    Part of the secret is to use smart naming. That helps you stay sane, and makes the editing in the sequencer go more quickly.

    It's best to process a small subset of samples, then test, test and re-test. Once you know exactly what you want to do, then process the whole lot.

    Finally, it's time to use the sampler. GS3 and Kontakt are the main ones. Get both. Choose one or the other to start. I'd start with Giga first. Once complete, import it to Kontakt and fix what broke. Now you serve both markets.

    I don't mean to minimize the programming work in the sampler, but if the capture and editing is bad, it will be bad in every sampler. If you do a good job at the front end, the lib has a chance.

    I think there are two paths for sampler programming. One is to keep it simple. Just provide great articulations, and make sure it's tuned, balanced and consistent. Piano libs are like this. The programming is relatively simple. It all comes down to quality.

    The second approach still demands quality, but here you get creative to add playability, or maybe specialized impulses or other "tricks" to make your lib easy to use. The sampler programming could get much more complex here, depending on what you're trying to do.

    All the best...

    -JF

  3. #3

    Re: Dumb Question

    Thank, JF!
    That really gave me a good idea of how to create a sample library. But is my 120 gig hard drive enough? It still has 110 gigs left, but I was looking at VSL at over 250 gigs!!! That amazes me how many samples must be in there!
    In terms of complexity of my library, I was thinking of, for the only complexities, keyswitches (especially for the strings!) and different controls controling different things, like volume, attack, dynamic level, stereo pan, legato (eliminating the attack portion of the sample), and stuff like that. Something similiar to GPO but with more realism.
    I know I have a lot of questions, but bear with me. I have all the time in the world--I'm home sick
    When I record my samples, I want to use a completely dead room, right? And when I edit, what exactly do I do? Do I seperate the attack and the release from the rest of the sample? Or is this just EQing and stuff?
    Well, that's all the questions I have.....for now, anyway!!

    Thanks for your help,
    Chris

  4. #4

    Re: Dumb Question

    OK, I thought of another question:
    What's all this about sequencing in GS3? Is GS3 a sequencer as well? Can I route it to external sound moduels too? Is that sequencer a "normal" sequencer, or is it specially designed for GS3?

    Chris

  5. #5

    Re: Dumb Question

    I will also chime in here, and give suggestions.

    I am not sure about the sequencing question. I use something else myself.

    As for the strings, if you want realisam, sample various bow lengths-up and down. This is how real string players play. I watch orchestras play and they are not always using legato...they have a way of making a very smooth transitions, but many times they are simply playing shorter notes with the bowing direction changing. For long sustain patches, you may want to offer a patch that has the start of the attach removed, but be certain to keep the start of the recording so the user has that option. I think you should also provide a long sustain starting with an upbow too.

    For variation, have your string players record notes while they play different bow directions simultaniously. This would be recorded in additon to the other up-down bowings.

    If possible try to record divsi patches...so many libraries do not have this. VSL was the first to try and offer it. I think the number of players should be rather small to begin with.

    The important thing is to record the basics first. Save the FX and extended techniques for an update.

  6. #6

    Re: Dumb Question

    Quote Originally Posted by cptexas
    Thank, JF!
    ...But is my 120 gig hard drive enough? ...
    It's certainly enough to get started. It depends on the final size of the library, and your process. If you make lots of backup copies, and copies of different versions, you'll eat that up more quickly.

    I would recommend an additional hard drive. Not so much for size, but for backups. You wouldn't want a disk failure to lose all of your work.

    Quote Originally Posted by cptexas
    In terms of complexity of my library, I was thinking of, for the only complexities, keyswitches (especially for the strings!) and different controls controling different things, like volume, attack, dynamic level, stereo pan, legato (eliminating the attack portion of the sample), and stuff like that. Something similiar to GPO but with more realism.
    This is all pretty strightforward. You might contact Dave Govett to find out about his planned tutorial video for GS3. He released one for GS2.5 that is really great for learning the details of editing.

    BTW, GS3 should run fine on your computer. The processor is a bit weak, but if you're not running GigaPulse, it's enough. GS3 Ensemble is enough for what you plan.

    Of course, to make great demos, you will want GigaPulse. That would mean that you would want GS3 Orchestra and a faster processor (2.4 GHz or more in the case of the Pentium).

    Quote Originally Posted by cptexas
    ...When I record my samples, I want to use a completely dead room, right? And when I edit, what exactly do I do? Do I seperate the attack and the release from the rest of the sample? Or is this just EQing and stuff?...
    There are different schools of thought regarding the room. VSL's is dead. Kirk Hunter recorded in a space, but didn't use release samples, so the instruments sound full, but you still need to add reverb. EWQL recorded in a hall with multiple mics, phase aligned things and included release samples, so no other verb is needed.

    If you weren't allowed to add reverb, EWQL would sound better than KH, which would sound better than VSL. But, we can add processing. So VSL's is the most flexible, EWQL the least, and KH somewhere in between.

    Most of the editing is aligning the start of the sample with the start of the attack, and ending the sample after the note has died away. You will probably want to do super quick fade in/outs to make sure the samples start/end on zero. (no clicks!). Compression, EQ, amplitude balancing, noise reduction and auto tuning can all be considered. One school of thought is to fully process it for a great out of the box sound. The other is to leave things more raw, so there's no info lost, and the composer can tweak the sound later. I prefer something in between.

    Regarding the sequencer, you will want Sonar or Cubase. If money's tight, a cleaper version of Cakewalk is fine. It does everything that Sonar can in MIDI, though Sonar has some convenience and time-saving features.

    All the best. And get well soon.

    -JF

  7. #7

    Re: Dumb Question

    Quote Originally Posted by esperlad
    I will also chime in here, and give suggestions.
    As for the strings, if you want realisam, sample various bow lengths-up and down. This is how real string players play. I watch orchestras play and they are not always using legato...they have a way of making a very smooth transitions, but many times they are simply playing shorter notes with the bowing direction changing. For long sustain patches, you may want to offer a patch that has the start of the attach removed, but be certain to keep the start of the recording so the user has that option. I think you should also provide a long sustain starting with an upbow too.

    For variation, have your string players record notes while they play different bow directions simultaniously. This would be recorded in additon to the other up-down bowings.

    If possible try to record divsi patches...so many libraries do not have this. VSL was the first to try and offer it. I think the number of players should be rather small to begin with.

    The important thing is to record the basics first. Save the FX and extended techniques for an update.
    Thank you very much, esperlad!
    I was intending this library to be one of fixed issues with other libraries. That's why I'm here at this forum! I was hoping to improve on Garritan's, but all other suggestions will be greatly appreciated.
    Don't you worry about the strings. I am VERY nitpickey about strings. I'm a cellist, so I see exactly where your coming from. I was planning on using Garritan's concept of layering a bunch of solo strings into a string section. That way, the size of the section is totally controlled and divici's are very convincing. Although I was thinking about waiting for the SampAllign "Sonic Morphing" technology to develop some more. I don't think it is a part of GS3 yet. You can take a look at it at http://www.garritan.com/stradivari.html. It's great sounding stuff!

    Chris

  8. #8
    Moderator
    Join Date
    Jun 2000
    Location
    Chandler, Arizona
    Posts
    4,045

    Re: Dumb Question

    The Sonic Morphing technology is Gary Garritan's and is not a program that will be available to other library developers. It is not based on GS3.

    Creating a sample library is not an easy thing to do and can be very expensive to create. Ask any developer who's developed a string library. There is a limitation to todays technology which accounts for most of the issues with the current libraries. VSL and many others use the brute force method of giving you a large amount of samples with almost any kind of articulation available. The problem with this approach is combining all the various articulations to get a realistic performance. Libraries such as GPO us programming to get the most out of fewer samples. It is designed for quick sketching without having to dig through articulations. Neither method is going to be perfect.

    Before starting on any library you need to know what the focus of the library is going to be, what features are to be included, etc. Then you need to figure out which sampling platform (Giga, Kontakt, etc.) have the features available to develop the library with. You may discover that you need to create your own engine as a couple companies have done.

  9. #9

    Re: Dumb Question

    Quote Originally Posted by Haydn
    You may discover that you need to create your own engine as a couple companies have done.
    How do I do that?
    I understand creating a library isn't easy, but I'm trying to learn as much as possible about sampling technology. Someday I hope to complete a sample library for a full orchestra. But I REALLY want to use sonic morphing! Maybe Garritan will sell his idea? I really don't know about patents and stuff like that, so I'm not trying to steal anybody's ideas. Just some of them are REALLY good!!


    Switch topics......

    In a large orchestra (for expandability), how many of each stringed instrument are there? I'm guessing about 42 violins, 16 violas, 12 cellos, and 8 basses. Is this well balanced? Am I on the right track?

    Chris

  10. #10

    Re: Dumb Question

    I would not suggesst using anything type of sound morphing or physical modeling. If you record enough material, the realism will come through. That is not to say that you cannot take certain notes and alter them to make sure that most techniques are matched. Cresc. and Dim. for example.

Go Back to forum

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •