• Register
  • Help
Results 1 to 10 of 10

Topic: ... and Sec's "Frankenstein" Choir: "Bist Du Bei Mir"

Share/Bookmark
  1. #1

    ... and Sec's "Frankenstein" Choir: "Bist Du Bei Mir"

    Ok, guys.... do not attempt to take seriously this one, not even for 5 minutes. My wife listened to this an hour ago and she is still laughing and picking on me

    This is a very shy attempt of mine to a SINGING choir. Singing in German, too!
    You should know Bist Du Bei Mir, Stolzel/J.S.Bach. I had this old-ish midi I wrote while learning staff view in Sonar. It was a strings arrangement... now 1st violins left their place to my "Frankenstein" female choir.
    The idea to patch the choir was simple: Convert Miroslav's Choir vocals in wavs to import into Independence and steal some pure consonants' samples here and there. I found some again in Miroslav's samples, a few more in "Halleluja" library, a choir and voice samples' set that I have... and grab the missing ones from youtube, browsing pieces sung by female choirs "a cappella", cutting and pasting tiny audio clips. Yes, this shows exactly HOW RIDICULOUS this whole thing is!
    Then ... I mapped everything into Indy. There's 5 key-switched layers with the vocals, plus some more layers - the consonants' layers - always active, independently from the active vocal. The problem is that the consonants' samples were so unclear, dirty, clipping and snapping that I had to make for most of them a dedicated layer with a dedicated volume envelope so to try to clean as much noises I could.
    That's a pic of the mapping:

    https://app.box.com/s/4qdf609a3w1qp1l4ibcq

    on the left you see the layers, on the right the actual mapping with the block of keys on the left being the vocals, keys on the right are the consonants.

    In this way this thing is even playable in real time... in THEORY, at least.
    It is actually rather easy to use... I pulled this rendering out this evening and yesterday evening, but mainly getting crazy finding offset points and attack curves for each sample... so the net time for the rendering was not too much.
    The funny part is to look at the track in Sonar. I've never filled up a (monophonic) track with so much data:

    https://app.box.com/s/xlwciz3mc6aayqmhpl9i

    The little dots in the the upper part are all the consonants, in the middle section are the actual notes -the vocals- and the lower part are the key-switches. A real MESS

    And now the mp3. The verses this thing SHOULD sing are:

    Bist du bei mir, geh ich mit Freuden
    zum Sterben und zu meiner Ruh
    zum Sterben und zu meiner Ruh.

    Bist du bei mir, geh ich mit Freuden
    zum Sterben und zu meiner Ruh
    zum Sterben und zu meiner Ruh.

    Ach, wie vergnügt wär so mein Ende,
    es drückten deine schönen Hände
    mir die getreuen Augen zu!

    Ach, wie vergnügt wär so mein Ende,
    es drückten deine schönen Hände
    mir die getreuen Augen zu!

    Bist du bei mir, geh ich mit Freuden
    zum Sterben und zu meiner Ruh
    zum Sterben und zu meiner Ruh.
    Remember that this is german, so, for example, "meiner" is pronounced like the english "miner", or "bei" is pronounced "by".

    And the file:

    https://app.box.com/s/qlyg5jbv9jf5k18gqi2s

    EDIT. NEW LINK.

    I guess I could do much more to make this patch work better... but it would be useless without some proper clean consonants' samples so I am giving up. The idea, tough, is not that bad. I mean... with a certain amount of imagination and self-suggestion this thing sings a bit... and it is a patch of less than 50mb... imagine what could happen with some serious samples.

    Anyway... this was fun. Especially for my wife....

    Fab

  2. #2

    Re: ... and Sec's "Frankenstein" Choir: "Bist Du Bei Mir"

    Sounds fabulous, Fabricio!

    I could need such choir samples, they certainly sound more imaginative than the "Ah" and "duh" that I am limited to work with

  3. #3

    Re: ... and Sec's "Frankenstein" Choir: "Bist Du Bei Mir"

    ..well, Reinhold... now you have to come here and tell that to my WIFE!

    Just joking. Thanks for the listen, glad you enjoyed it. I'd let you download the patch, but there are legal issues as I do not own the samples. I mean... I am pretty sure that the "Wishful singing" female ensemble wouldn't sue me for stealing one of their "N", or Paragita choir wouldn't also do so because I've grabbed a "S" ... but at least "Miroslav Vitous" would get a bit angry

    Thanks again.

    Fab

  4. #4

    Thumbs up Re: ... and Sec's "Frankenstein" Choir: "Bist Du Bei Mir"

    Hey Fabrizio, it's great!
    what you did is amazing, due to the tools you were using... I'm really surprised by the nice sound you got, congratulations.

    I was a mad-scientist of Frankenstein choirs myself, and I know what you mean! Before giving-up, I did some experiment recording the consonants and a very low volume solo of my own voice to inflate a bit of reality and Italian sound to EWQL Symphonic Choir library:

    Dio del Cielo

    If you like this type of work i recommend you think to buy it (it's on sale today, even if still a bit expensive).

    the best of this library is singing with Orchestra (it works decently) but with a bit of effort, a cappella is not exciting but still possible:

    Oh Tannenbaum

    Kyrie Papae Marcelli

    Anyway, it's an expensive and professional library... your choir was free... and you had a huge amount of fun... that has no price.

    ciao, alla prossima
    Fabio

  5. #5

    Re: ... and Sec's "Frankenstein" Choir: "Bist Du Bei Mir"

    The entertainment value of this thread of yours is Huge, Fab - Fantastic stuff! - Why can't I take it seriously? I don't find anything to laugh at!

    The recording sounds very nice indeed - From what you describe, it sounds like the assembly time was rather enormous in comparison to the length of the results. But as Fabio said, you obviously had a great time putting it together, and that gives the project incalculable value.

    You're one of the few people who post about putting together do-it-yourself music projects. Most of us have our hands full simply loading whatever Libraries we have and trying to make music with it. It's admirable the way you roll up your sleeves and dig in to do custom editing of samples and twisting the arms of programs to make them do more than originally intended. I love it.

    I looked at the screenshots, read your text several times, but I really don't have a clear picture of what you actually did to make this work.

    I think you mean you first assembled syllables, vowels, consonants from a variety of sound sources, and that audio track of assembled bits and pieces then triggered the choir samples which originally are just "ahs"--? Confusion for me is that in the Sonar shot, it looks like you have seven MIDI tracks - This must be because you're using "simple instrument" tracks which I never use, but which combine audio and MIDI on the same track display?

    Isn't there some extra plugin you needed for the audio clips to work as volume envelopes for the MIDI?--Or is this ability built into Independence?

    Once in awhile I've used a morphing plugin which yields rather artificial, robotic results, but is fun - The source audio creates the envelope shaping the Ah so words can be heard, and the original audio is no longer an audible part of the results.-- Is what you're doing here something like that?-- If so - couldn't you record your own vocal track rather than dealing with some "unclear, dirty, clipping and snapping" bits you found on YouTube and elsewhere?-- With the morphing plugin I've used, I just record a track of me speaking the words as I listen to the backing track. After making any adjustments needed, that spoken track then drives and shapes the choir. Maybe you're doing something different here, but as I said, I couldn't really understand your process.

    Come back and tell us more! Your test track is very successful - Maybe there's a way to streamline the process more?

    Randy

  6. #6

    Re: ... and Sec's "Frankenstein" Choir: "Bist Du Bei Mir"

    Quote Originally Posted by Fabio View Post

    ...

    I was a mad-scientist of Frankenstein choirs myself, and I know what you mean! Before giving-up, I did some experiment recording the consonants and a very low volume solo of my own voice to inflate a bit of reality and Italian sound to EWQL Symphonic Choir library:
    ...

    ciao, alla prossima
    Fabio
    Ciao Fabio,

    thanks for your words, I truly appreciate.
    I know QL Choir, I do not have it but I've heard many demos... also something by yourself... a song maybe, I do not recall well. And, of course, your demos up here. QL Choir is the top VST choir and will probably will be for long. There are other nice choirs on the market... but they can only sing nonsense phrases so QL remains pretty much in a category of its own..... UP TO THIS MOMENT! Ok, ok, I am just joking
    QL is not perfect, its words are not always very intelligible, but, for what it is, a VST, it is really smooth and musical. "Oh Tannenbaum" was really convincing.... but I do prefer your own tricky "Frankenstein" version: it is a bit like cheating, but it works!

    As you say, working on this patch was very interesting. There's the sheer fun, as you mention (and by fun I mean some really explosive laughter ) and there was the special taste of the "challenge" to add a few flavors to this weird recipe.
    Remember when I said that this was ok and I was giving up? Well.... I was lying! Especially to myself. I thought about this and I understood a few sources of many of its problems. More about this in the answer to Randy's message, he asked some more info so there will be some talking to do.
    So... I think I'll will keep having fun with this guy a bit more.

    Alla prossima!

    Fabry

  7. #7

    Re: ... and Sec's "Frankenstein" Choir: "Bist Du Bei Mir"

    Quote Originally Posted by rbowser- View Post
    The entertainment value of this thread of yours is Huge, Fab - Fantastic stuff! - Why can't I take it seriously? I don't find anything to laugh at!

    ...
    Come back and tell us more! Your test track is very successful - Maybe there's a way to streamline the process more?

    Randy
    Randy! Thanks you for chiming in, you are always so supportive.
    By the way, don't you know that you should NEVER push an italian into ....talking? You did it, be ready to bear the consequences!
    The point is that is a bit difficult to explain the logic behind this... and probably will be VERY verbose. I'll try.

    The idea was simple. I know many of us tend to think that talking - or singing - is a complicated thing involving morphing of sounds, modulation and so on. If you think about it, though... there's nothing more monophonic than the human voice. No way it can produce two sounds together, not even a vowel and a consonant. Also... when we sing, we really just sing the vowels, while the consonants are sort of immutable and are also "weird" sounding. When used in speech - or singing - they are really just some "noises" between vowels. Take a "M". When we name it, we have to say "EM" in english, it has to be named using also an accompanying vowel, otherwise it would be a sort of muted noise. And that muted noise is exactly what we produce when we speak - or, again, sing.
    The idea used in this process was simply to build phrases according to this logic, just connect a vowel with a pure, "muted noise" representing the consonants.
    Getting the vowels was easy. I found clean ah, oh and uh samples in Miroslav's Choir. Eh and ih were derived by its "Portato Re" and "Portato Si" samples by "cutting their heads".
    Consonants were difficult... as already said I had to search, audition files, scan choir performances and so on. Remember that I wanted "pure" consonants, without any hint of a vowel in them, so I had to cut tiny, tiny bits of audio. When you split audio, the remaining wav file usually tends to clip a lot, producing a lot of noise. That was the main problem that I tried to solve by fading in and out the files, offsetting their start and acting on dedicated attack envelopes. I did all the cutting and fading in Audacity, while the normalizing, offsetting and the volume envelopes were done in Independence. Yes, Independence has its own wave editor and lets you handle wavs as you want (slicing, offseting, looping, envelopes and so on)... Aria does it too... it is only more difficult to program.

    So, once the material was grabbed, I did the mapping, which I showed in the "mapping" pic in the previous post. As said, voices just sing the vowels. These therefore need to have the full range of pitches and they have to be interchangeable. 5 layers for them, one per each vowel, with relevant key-switches. Then I mapped the consonants, which are immutable, on the higher octaves. These are always available, key-switching does not apply to them. So, if I key-switch to "ah", I'll have the consonant "B" on the key F5. If I change and key-switch to "eh", the "B" will always be there on F5.
    How to use the patch? Simple, let's go step by step. Our midi part is ready, with the right notes. First verse is "Bist du bei mir". We sing the vowels, right? So, the notes will be keyswitched to "ih", then "uh, then "ah", "ih" and "ih". These are the vowels of the phrase: (B)"ih"(ST)(D)"uh"(B)"ah""ih"(M)"ih"(R). That' s German pronunciation, remember. Once the vowels are in place, I just add the consonants in PRV by hand, right in the places they need to be "spoken".
    It is easy.
    It took me a lot of time to build the patch, as I had to find the material, edit it, normalize and so on, but to use it is not very difficult. Rather intuitive, I'd say. In fact, if you look at the track pic I posted earlier, you notice the consonants-dots at the top, the actual notes - the vocals - in the middle and the key-switches at the bottom.

    Ah... your other observation: yes, I do most often work using simple instrument tracks. I do not need post-production when I am using Independence as everything is done internally, using its own effects and routing everything through its internal buses to its own mixer. I never look at Sonar' s mixer, actually... Independence's internal mixer is more accurate and more customizable. In this case, what you hear is pretty much all you can see on the pic, I am only using a master FX console on Sonar's Master bus with just a "coloring" IR reverb in it. Actual reverbs are handled internally, separately on each instrument.
    EDIT: Ops, Randy. I made confusion yesterday. Those are NOT "instrument tracks", they are simple midi tracks. There is just one audio track in there, the one named "Independence", and that track handles all audio and routes it directly to Sonar's master bus. That's because, as I said, I usually do everything internally of Independence (when working with it).

    Getting back to the patch. In order the process of connecting vowels and consonants to run smoothly I am relying on basically three parameters: velocity, that controls volume, attack and sample offset, a rather wild usage of the "master" attack envelope that is routed to CC21 (these are the frequent bars you see on Sonar's track icon) and the legato engine. As the voice is strictly monophonic, activating the legato engine is the quickest way to make sure that sounds do not overlap, so it becomes rather fast to program a "phrase". The problem is that legato gets influenced by key-switches, too, and this creates a lot of erratic behaviors: whenever I have a sequence of two vowels and I have to key-switch to the second, the legato engine gets triggered by the key-switch and I have a rather annoying pause. This is pretty noticeable on the mp3 and I'll have to solve this somehow. The easy way is to momentarily switch the legato engine off after having mapped it to some cc... I'll see.

    I already modified a few consonants' sounds, cleaned some more samples and modified (once again) lots of attack curves and offsets and the patch is "responding well", the first phrases are much cleaner now. I might be posting an updated version of the mp3 sometimes soon, so, if you want, you can help me evaluate the difference.

    That's pretty much all of it, Randy. This has indeed been VERBOSE!

    It is encouraging to know that you find my approach interesting. This place is full of talented people, I really enjoy to have the possibility to contribute with my own perspective, hoping that fellow members might find these things of some interest.

    Thanks again for your "visit".

    Fabry

    P.S. Ah, a small addendum. Most of the consonants used are grabbed by a video found on Youtube. I bumped into it searching suitable "a cappella" performances with the right ambience and timbre. But these girls are absolutely amazing, give them a listen:

    http://www.youtube.com/watch?v=RQ9-CwvrYog

    Aren't they just something?

  8. #8

    Re: ... and Sec's "Frankenstein" Choir: "Bist Du Bei Mir"

    Quote Originally Posted by sec2 View Post
    My wife listened to this an hour ago and she is still laughing and picking on me
    Hehe!!! Well, I thought this was gorgeous! I loved the singing and the overall presentation of this. Kind of reminds me of dramatic romance! One thing I have to ad to my libraries yet is a full choir. I have the basics like Olympus Micro Choir, GPO and GIO choir, Vienna Soprano choir, but nothing really complete. I wanted to get the 8dio Angel Choir but I had to choose between kontakt full version, Angel Choir, or Vienna Volume 3. I hate when I wanted all three but can only choose one. After listening to this has definitely put me on the path to getting the choir library next.

    Great work FAB!!!

    Cheers,

    Richard

  9. #9

    Re: ... and Sec's "Frankenstein" Choir: "Bist Du Bei Mir"

    Quote Originally Posted by sururick View Post
    Hehe!!! Well, I thought this was gorgeous! I loved the singing and the overall presentation of this. Kind of reminds me of dramatic romance! One thing I have to ad to my libraries yet is a full choir. I have the basics like Olympus Micro Choir, GPO and GIO choir, Vienna Soprano choir, but nothing really complete. I wanted to get the 8dio Angel Choir but I had to choose between kontakt full version, Angel Choir, or Vienna Volume 3. I hate when I wanted all three but can only choose one. After listening to this has definitely put me on the path to getting the choir library next.

    Great work FAB!!!

    Cheers,

    Richard
    Thanks Richard. But I am never satisfied... this thing can and WILL get better!

    If you are looking for a choir.... look no further than Fabio's post up there. As he mentions, there is a sale on East West QL Choir these days. I went to look and the price they are offering is pretty interesting.
    There is no choir like this one. It can do everything other VST choirs do PLUS you can nearly get it to sing proper phrases.... and these days it is even cheaper than the others . Worth a look.

    Good to have you around, Richard.

    Fab

  10. #10

    Re: ... and Sec's "Frankenstein" Choir: "Bist Du Bei Mir"

    Ok, this one definitely needed a bit more work. Just "a bit" .
    Now I am more satisfied. Actually, I am VERY satisfied. I changed a few consonants' samples, cleaned the samples a bit more, made a bit of order so the the behavior is more predictable and the patch is easier to use... and solved that really annoying issue with the legato engine. Spelling is much clearer now, it is nice and flowing, and generally the "musical behavior" of the patch is much enhanced. Plus... I learned how to use it a bit more

    Same file, updated rendering:

    https://app.box.com/s/qlyg5jbv9jf5k18gqi2s

    I'd say ...now you can take this seriously

    ..and, above all.... I succeeded to grasp one of the dreams of each and every men on this planet: ... wiping the laughter off his wife's face!
    (Hey, I am joking of course. I do love her so much... and she showed to be quite a severe critic and an efficient motivator. ONCE AGAIN! )

    Fab

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
  •