A User Tutorial For EWQLSC

[nlamartina]

It's taken me some time, but it turned out pretty good. I'm just hoping that when the updates are released that the whole thing won't become obsolete (or maybe I do :p ). At the very least, I'd like to continuously update this so that all SC users can benefit from it, new and otherwise. Oh, and to try and save space and bandwidth, all my MP3 examples are 96 kbit. They're not spectacular, but they're good enough. For this tutorial, I was working with the F Bass WB Dyn Mod multi. Also, the whole tutorial exceeds the maximum post length, so I'll do a series of posts. Also, please keep in mind that the entirety of this tutorial is my opinion on using the program. This isn't endorsed by anyone but me, so I present it as-is. I know there are some spelling mistakes and ambiguities in here, and I'll work toward ironing those out.

So without further ado, here it is. Before we get into the nitty gritty however, let's take a look at a few startup things...

Hardware Considerations and Other Misc. Tips

If your machine is capable of dual-channel memory transfers, USE IT, even if it means rearranging the memory sticks in your machine. You’ll see a dramatic difference in performance.
If your sequencer has an option to not close open MIDI ports when it loses window focus, turn this on when running WB externally. This will prevent EWQLSC from ‘burping’ each time you change focus between windows.
Try working with one choir at a time, not only to keep your focus when tweaking your Votox, but also to leave more processor overhead for auditioning and recording your choir.
Before doing any serious work with your choir, make sure to plug your MIDI controller into WB (after opening Kompakt and loading your choir multi) and twiddle the mod wheel to ‘clear the choir’s throat’. If you fail to do this, all the consonants will be pronounced out of balance, and any work you complete during the tweaking process will be a waste of time.

For example, here's what SC sounds like without twiddling before beginning:
http://www.lextechmedia.com/tutorial/nomod.mp3

And here's after twiddling
http://www.lextechmedia.com/tutorial/modtwid.mp3
I prefer to move the non-pitched consonant volume slider about 3 or 4 notches to the right. This is mostly a pedagogical consideration, so you should play around with this setting to suit your tastes. However, be sure to DO THIS FIRST, and once you set it, LEAVE IT ALONE. Adjusting it later in the middle of your tweaking will ruin the relative volumes between your consonants.

EastWest\Quantum Leap’s new Symphonic Choir (EWQLSC) package represents a revolutionary breakthrough for musicians, producers, and composers with its highly realistic and thorough presentation of the first believable synthetic choir. Like other EastWest programs, its interface is clean, well-designed, and intuitive. However, because the methodology of this software is so radically different from other sampling products available today, it can present the end user with unique challenges in terms of how easily it can be learned and how quickly one can get meaningful output from it. However, through both education and practice, it is not only possible to render perfectly believable and rich recordings from it, but it can also happen very quickly. The purpose of this tutorial is to help users get to that point of comfort, where they can both skillfully and quickly get believable results from the software, with deft and confidence. I will cover a variety of topics, but the EWQLSC process can be broken down into four very simple steps:

Setting text
Adjusting velocities
Adjusting timings
Rendering for output

That’s really all there is to it. There’s nothing painful or difficult about it. As long as you have your workstation, music, and text ready, this whole process can be completed in a matter of days (or even hours, depending on the complexity).

[nlamartina]

1. Setting Text

Materials:
MIDI file: http://www.lextechmedia.com/tutorial/tut1.mid
Finished CHO file: http://www.lextechmedia.com/tutorial/Tut1.cho

Setting texts for real performers can be one of the most difficult yet rewarding parts of choral writing. It gives the composer not only an opportunity to breathe life and authority into his or her music, but also to tell a story inside the story by painting the text over specific musical devices. In this, the music truly comes to life. While I’m not going to talk about text considerations, I will highlight one of the most important parts of text setting, and that is syllabic separation. Since we’re working with a virtual choir, we won’t have to deal with some of the disciplinary problems associated with a real choir. EWQLSC won’t talk when you drop your hands from the keyboard, it won’t mutate or spread its vowels, and it’ll sing on key regardless of how much it’s practiced. However, despite this we still must be careful about how we break up syllables in our text when typing it into Word Builder (WB). Let’s use an example word, such as, “Disparity.” When setting a word such as this, it is common practice to always avoid ending a syllable with a consonant. Thus, on the page:

This would be correct: Di-spa-ri-ty
This would not be: Dis-par-it-y

It’s the same number of syllables, and the same word, but with the latter example, a real choir will be tempted to close to the consonant and sing on it. This is bad technique, and thus, settings like the former should be enforced.

Working in WB is not much different, although our motivation for doing so is. The EWQLSC won’t try to close to the consonant in the middle of a held vowel, but it will sometimes close prematurely. There’s a subtle difference that some might not mind, but using proper text setting is ideal for two reasons. First and most importantly, it will save you time when you’re adjusting your vowel and consonant timings. If you leave consonants at the end of syllables, WB will play them as you release the keys of your MIDI device as you’re auditioning your piece. If you set the text properly, WB will simply release the vowel and not sound the next consonant until you play the next syllable. This gives you more control over your audition, and it more closely emulates a real performance and how an actual choir sings. Finally, it prevents WB from making assumptions about when the consonant has to be lifted and when to elide to the next syllable.

Speaking of elision, let’s look at one more step we can take to assist WB in its pronunciation of the spoken word. Let’s use an example phrase:


“I gave this book to my old friend. He spoke to me, and not again.”

Now imagine this will be sung over a series of long notes, very legato. Good singing technique mandates that we sing on the vowel at all times, with as little interference from consonants as possible. Therefore, we elide, or combine, consonants to keep the breath moving and keep the phrase strong and pure. So in Votox, we already have this phrase (while following the above rule about setting):


“aE giEv t!Es bUK TO maE oOld Frend. HEE SPoOK TO mEE, And naT u-gen.”



Now let your sequencer play this passage and notice the consonant chatter between the ending Ks and beginning Ts (sometimes called “the kitty sneeze”). There’s also a lift between the T of “not” and the A of “again”. But now, let’s take the default pronunciations a step further and elide the consonants:


“aE giE t!Es bU KTO maE oOld Frend. HEE SPoO KTO mEE, And na Tu-gen.”



Now it will take some practice to learn how to do this and for what word combinations, but as a general rule, it’s used when two consonants border one another across two words. We’re not trying to eliminate the consonant sounds. More accurately, we’re forcing WB to keep the consonants short, allowing your choir to sing on the vowel more without you having to go back adjust every single instance where consonants collide. It creates a smoother, more legato delivery, without having to adjust the attack key-switch. Additionally, this gives your text the ability to be sung/played faster by the software since your consonants will already be at their minimum value. Put simply, it makes your text more flexible and saves you work later.

Listen to the differance now: http://www.lextechmedia.com/tutorial/elide.mp3

The last part of this step involves selecting proper vowel shapes for our words. While the pronunciations WB provides are a good start, they aren’t always in the best interest of proper vowel formation. For instance, see the pronunciation of “again”. The schwa sound (“uh”) has been selected for the A. Schwa’s are ugly, and we need to avoid using them in this context. Good, traditional singing technique demands that we stick with pure Latin vowels, “Ah, Eh, Ee, Oh, Oo”, or in Votox, “a, e, EE, oO, O”. Now I have yet to figure out a way to make the choirs sing longer vowels (like the German ü), but for the purpose of this tutorial, the general rule of pure Latin vowels will suffice. Now of course in English we have other necessary vowels, like the short I (bit) and the retroflex R (better), which is a beast in itself. What we’re looking at is avoiding schwas and substituting pure Latin vowels whenever possible. Basically, Queens English.

Now plug in your keyboard and let’s start auditioning this phrase word by word. If we let the sequencer play everything we might miss something. Play slowly, and plod your fingers over the keys. Dwell on the words, and listen for sloppy vowels. Make liberal use of the F4 and F5 keys (start-at-selected-syllable and start-from-beginning, respectively). Now the start of our phrase is great. “I gave this” is all okay. But let’s listen to “book”. It sounds like “buk” like, “bubble” with a protruding jaw. We don’t want that. Let’s try and substitute a pure vowel. In Votox, spell it, “bOk”. Now the Votox O might look a little odd, but in the context, it sounds much better. What’s left? The only other rough word is what we looked at previously, “again”. Lets nix the schwa and sub in another pure vowel, Votox “a”. Ah, much better! Our final phrase is now spelt,


“aE give t!Es bO KTO maE oOld Frend. HEE SPoO KTO mEE, And na Ta-gen.”



Have a listen:
http://www.lextechmedia.com/tutorial/goodvow.mp3

[nlamartina]

2. Adjusting Velocities

Materials:
MIDI file: http://www.lextechmedia.com/tutorial/tut1.mid
Finished CHO file: http://www.lextechmedia.com/tutorial/Tut2.cho

This is where the real magic happens. I’d say that 40% of your work in WB is going to come from actually typing in the words. Another 40% is going to come from adjusting the velocities of the letters inside those words to make them sound more realistic, because the big issue with sounding realistic usually isn’t with the actual phonics. It has nearly everything to do with letter emphasis. The struggle comes from the fact that WB sets all letters to their default velocity after they’re typed. Naturally, we don’t speak with all of our letters as the same volume. There’s a difference between the beginning and ending consonants within each syllable, along with how each syllable is pronounced in relation to others. Even whole words have this relationship, so you can see why it’s important that we pay close attention to how we ask WB to pronounce things.

The ‘corrective process’ (as I’d like to call it) involves nothing more than manually playing into WB and listening closely to the results. Don’t worry about quick passages or even melody at this point. Simply play smooth and even, and listen for several different things, like letters that jump out at you, incorrect vowels, and poorly-timed diphthongs.

http://www.lextechmedia.com/tutorial/vel.JPG

Adjusting letter velocities can be tricky, but there are a handful of things that you’ll find yourself doing each time. Identifying these needs will help speed up your production. For instance, two letters that consistently need to be adjusted are ‘n’ and ‘r’. ‘R’ is almost always too loud, and ‘N’ will sometimes be too much or too little depending on the word (quite a concept, huh?). ‘T’ is frequently too quiet. When adjusting letters like this, I find it is most useful to use the multiplier or the divider in the velocity menu. I also keep the attack context sensitive.

http://www.lextechmedia.com/tutorial/letter.JPG

To me, this setup just evokes a more realistic performance out of the choir regardless of who is playing the keys (me or my sequencer). When choosing integers for the multipliers and dividers, I tend to use numbers close to fractions, like 1.11, 1.25, 1.33, and 1.5. If I ever have to adjust a velocity by more than 1.5, it means that something’s wrong, so I’ll usually double-check that I’m editing the right letter and that the mod wheel has been twiddled. For example then, *1.11 increases relative volume of a letter a bit. *1.33 increases it more noticeably. /1.11 subtly decreases volume, while /1.33 decreases it a lot.

http://www.lextechmedia.com/tutorial/high.JPG

It’s also a good idea to check the actual velocity of the note in the sequencer to make sure it’s not too high or low.

Now for our example, go back to manually playing WB. “I” is already pretty good. “Gave” is weak. Too little G and V. Let’s bump them up. Set the G to multiply by 1.33 (we’ll say *1.33 for short). Set the V to *1.11. Listen again. Still not perfect, but better (more will be done in the next section). The next chunk is good. “This book to my old friend. He…” Here comes “spoke to” again. It’s a bit weak on the T. Set it to *1.11. Now our phrase is tweaked.

[nlamartina]

3. Adjusting Timings

Materials:
MIDI file: http://www.lextechmedia.com/tutorial/tut1.mid
Velocity and mod tweaked MIDI file: http://www.lextechmedia.com/tutorial/tut2.mid
Finished CHO file (after speed learning): http://www.lextechmedia.com/tutorial/Tut3-1.cho
Finished CHO file (after speed tweaking): http://www.lextechmedia.com/tutorial/Tut3-2.cho

This is the last tricky part, and it can be a bugger. Letter timing is critical here because certain letters will sound louder or softer depending on how long they’re held. Now before we get excited and start trimming letter velocities, we have one critical step to complete before we get anywhere, and that is to speed learn our words. It’s worth mentioning here that if you haven’t absolutely finished your piece, now is the time to do so. Once we speed learn, the letter timings will be set in stone, and messing with our music after the fact will ruin our work, so make sure you’re happy now.

Fortunately, this is the easy part. Make sure your sequencer is plugged into WB first. Now click the “Learn” button, and press “OK” (make sure that the mode is set for “Change Speed”).

http://www.lextechmedia.com/tutorial/learn.JPG

Now hit Play in your sequencer. The progress bar of WB will begin to fill as it adjusts your timings. When the piece concludes, hit OK (if the progress bar doesn’t fill completely up, that’s alright). Notice what happened. Prior to this, all syllables had two parts: one for the note-on message, and one for the note-off message (like the word “car”, where “ca” is note-on, and “r” is note off). Another way to put this is to say that each syllable has a separate attack and release (the sustain is part of the attack). While this gives you the flexibility to play the choir as fast or as slow as you wish, the ambiguity will give you issues when your notes get shorter, since the default note attacks may be too long for your quick passage, resulting in burping or dropped parts of speech (as evidenced in previous audio examples where the word “this” refused to play through). Now that you’ve speed learned, all letter velocities are set to the song’s actual note values. This, in short, is awesome. One other important thing this process has done is remove any designation between syllable parts. There is no longer an attack and a release. It’s just one attack. Even if you manually play it by tapping notes on your controller, each syllable will play at an exact length, and this is where the final bit of gloss is applied.

Here's what everything sounds like after learning:
http://www.lextechmedia.com/tutorial/learn.mp3

Let’s listen to it. Much better, isn’t it? Even that bothersome “this” is working. It all sounds great, but we need to lengthen and extend some letter timings to remove any lifts from the singing. A vocal coach would tell our choir, “Sing on the breath! Sing on the vowel!” Breaks are inappropriate, and here is how we avoid them. First we listen. The first part is good, except that there’s a lift between “book” and “to”. Fixing this is easy. Simply go to the offending letter’s timer (in this case, Votox “o”), and drag its length rightward (the red bar). See that red marking on the timeline above? That’s like the ‘danger zone’, where WB is afraid we might spill our letter length over to the next syllable. That’s exactly what we’re going to do to make sure this elision works. Take the bar and pull it over to the 1200ms mark.

http://www.lextechmedia.com/tutorial/red.JPG

Now listen again. Hear that? The lift is gone, because we let the “o” of “book” sound off longer, to the point where it came in contact with the “k”, completing our elision! Congratulations! You’ve learned ‘the big secret’ of WB, and you’re on your way to getting this finished!

Listen some more. The “to” in “to my old” is too short. Lengthen it like above. The end of the O timer (green this time) is at 400ms right now, but we can lengthen it to a little past 450ms. Now do the same for the “to” in “spoke to me”. You pick a value to extend it to this time. Play around with it. How about the “and” of “and not again”? For this one, it’s just slightly different. The end of this syllable ends with a consonant, so we don’t want to lengthen that. Instead, take the entire “d” timer, and scoot it to the right so it buts up against the 500ms point. Now scoot the “n” time rightward also so it overlaps the “d” a bit. Now extend the “A”. Good. Now you can do both! Finish the rest of it on your own and compare it with mine for guidance. Finally, we can even go back and tweak out MIDI file a bit to make it more expressive (see above MIDI file, tut2.mid).

Here's the final product:
http://www.lextechmedia.com/tutorial/final.mp3

[nlamartina]

I'll add this part later. I'm tired of typing for right now.

[EricRichmoind]

THANKS FOR THIS!!!!! AMAZING!!!!

[luima]

EXTRAORDINARY WORK. THANKS A LOT !!!

Why dont post a single PDF or HTML?

[nlamartina]

I can do that, if there's demand. I did it here so that discussion on the topic could be kept in a relevant place. It's also a little more exposed and intimate (if that's the word for it), rather than just slapping a link on the boards and saying, "Have a look!" :)

Sure though, I should be able to make an HTML copy without issue. I'll get it posted soon.

[V o n h ö g e n]

Thanks a lot for your EXCELLENT tutorial!!

I have quickly restyled your tutorial in iWork Pages, so you can save some time for the final chapter ;)

I have just mailed you the resulting PDF- and html-files. Please, use it if you like. If it's not what you want, just trash it. After all, you indicated that you prefer putting it online via our forum, because of the expected feedback.

I'm pretty sure that if Doug and Nick would organize a monthly tutorial-competition, you'd win the cigar easily. Keep up the good work!

Jerome

p.s. Can't wait for Chapter IV...

[mee]

This is such a great tutorial! But the files that are linked as examples are no longer on the server.

Does nlamartina still write here? Can these links be updated?

Was a PDF version ever made?

[oleJoergensen]

Yes, it will be great to have this amazing tutorial.....

[lumcas]

Yes, it will be great to have this amazing tutorial.....

+1

[PBCliberal]

Me too. This is the best thing I've ever read for WB and Choirs. The necessity of twiddling with the modwheel just answered a whole host of questions I had about how sometimes it just seemed to go crazy.

[likecmposr]

This thread has just made my dayo.

[JCL]

None of your links are working for me! Looks like a very good tutorial, but we need to hear the examples. Thanks.

[Jeggis]

Hey all!

I've just seen this, after countless hours of frustration in WB. Most of my wowels is unbalanced, and everything sounds unreal. You mentioned how I should "twiddle" my mod wheel in WB before opening SC, but I can't honestly say I know what that means. I tried, and after typing a few words in WB it sounded good, but quickly turned bad.

I hope someone can explain this mod wheel twiddle process for me.
Thanks.

[Peterkjones]

nlamartina, ura saint! By trial and error and a vast expense of time, I've done a lot of what you so carefully describe and am getting really satisfying results. But you've expanded my range and set it out all so clearly! This should be an obligatory part of the WB bundle! html or pdf, please, as soon as you can manage it - we should all be incredibly grateful to you!

[cyril]

Hello

I am trying to get your files I get :

-------------------------
Download DriverMagic Now.

Update & Repair your Drivers, For Any Windows Computer, Quick & Easy
www.SymplisIT.com
Lex Persona

Certifiez et signez vos documents électroniques à valeur probante
www.lex-persona.com
Impellimax

Switch Drivers, Hybrid Circuits, Obsolete Replacements & More !
www.Impellimax.com
Driver

Compare Prices & Save - Shop Smart! Top Deals on Popular Video Cards.
PCWorld.com/GeForce
UniPrint. A PDF-based

Universal Printer Driver Enabling Users to Print Any Documents.
www.UniPrint.net
LEX-20

LEX-20 with warranty All laptop parts. Buy Now !
www.pchub.com
Check agency driver cards

Handheld device with rules built in activities, summary, infringements
digifob.com
4120b driver

Quickly Update & Fix All Drivers With The Top Rated Driver Software.
Update-Your-Drivers.Com
Driver Training made easy

D.I.Y. Driver Training Kit For Parents with Teen Drivers
www.coacharookie.com
Drivers

Super sélection et affaires pour les articles Drivers.
eBay.fr

--------------------

Can you do something

Best

Cyril

[JCL]

nlamartina, Can you get your links to work???

[cloudswimmer]

nlamartina, Can you get your links to work???


X2 :D

[Alec]

X2 :D

x3 :D

This tutorial is the holy grail for SC!

I really would like to listen to the audio examples!

[Jarrek2002]

Very well done tutorial!

[dezzo]

I can do that, if there's demand. I did it here so that discussion on the topic could be kept in a relevant place. It's also a little more exposed and intimate (if that's the word for it), rather than just slapping a link on the boards and saying, "Have a look!" :)

Sure though, I should be able to make an HTML copy without issue. I'll get it posted soon.

Can you make these Files you´ve posted available ? http://www.lextechmedia.com is not working :confused:

Thx,

D

[JCL]

The links seem to be working for some people , but not for others. Why is that?

[jloving]

it probably has something to do with the fact that this was posted circa 3 years ago ;)

if you take a peek at the OP's profile you'll see they haven't been active for nearly a year as well, so I wouldn't get your hopes up too high.

nonetheless it is an excellent tutorial!

Jon

[JCL]

Thanks, Jon.

[Peterg]

Hi

I did a search on nlamartina on google:

He is currently working for Stargate, the game company, as their sound designer, I am sure if someone contacted him at the Stargate forum he might just be able to post the missing files for the tutorial.

I am sure this is the guy that wrote the excellent tutorial for EWQLSC

http://www.youtube.com/watch?v=HTBS9hUjXCs

Regards
Peter