Hey-
Here's another "first" piece done with Vocaloid. Unlike some of the others who have posted here, I like Vocaloid both for the sound and ease of use. Since I use a lot of different speech synths for The Bots, I have to say that Vocaloid is bringing it to another level. No, there's not the imminent danger that human performers will be totally replaced by these things, but that's not the point. For me the point is that these are expressive instruments, and as they keep getting more so, they will in fact start replacing human performers more and more. Well, that's what The Bots believe. Anyway, please feel free to download and critique "My Mind Is Free" which features Vocaloid Leon, from
www.thebots.net
-you might also want to check out "Can't Explain" which is probably the most advanced tune The Bots have done prior to Vocaloid.
Here are a couple of critical comments about Vocaloid:
1. As some others have noted-the copy protection really sucks. It's a challenge-response based on your network card. My audio pc is not on the internet, so I went thru the whole procedure of generating my challenge code, bringing it to an internet pc, getting the response back to the original pc, only to find in use that vocaloid really couldn't run alongside my sequencer. Because it is not real time, and the rendering will cause problems for Sonar, I realized it's best to run Vocaloid on a second machine. But I only realized this after I had wasted one of my precious authorizations on the wrong machine. What you want is vocaloid running on a second machine that is networked to your main audio machine, then you can render and audition in Vocaloid while you are running your main sequencer. The Vsti is not worth using.
2. a VERY BAD PROBLEM is that the word-to-phoneme conversion simply doesn't work. There appear to be some words that Leon ain't gonna say no matter what. I couldn't get the word "baby" to go thru no matter what I did -all it would say is the default "ooh." This is almost indescribably frustrating, and severely limits the usefulness of the product. I think for about 20 years now TTS systems have been successfully converting natural language text into phoneme tables. It is amazing that Vocaloid doesn't do this, and I hope Yamaha/Zero-G will fix this problem asap.

Anyway, there's my rant, please enjoy the tunes.
-bc

Thanks for posting an alternate use of Vocaloid.

I would agree that it is probably a bit better suited for the artificial/synthetic style of music that you're doing. Although I actually think the artificial voices in "Can't Explain" sound much better than Leon does in "My Mind Is Free".

I would also hazard a guess that you like Vocaloid's 'ease of use' because you're creating relatively small vocal snippets and you actually want them to sound synthetic. In this context, there is no ease of use issue because I do agree that it's very easy to use if you don't want it to sound real.

Thanks again for posting. BTW, I really like the arrangement of "Can't Explain". Nice job.

Ojuzu

Will check out your tune shortly (still listening to Fitches great stuff. ) But what you say about this 'Challenge/Response' sounds so crazy. I'm just a small guy on a small laptop and enjoy loading a CD-ROM and at worst - entering a big #. I would seriously hope Zero G or Yamaha could implement this at some point for a sap like me? And couldn't it be said that once one goes through all this 'challenge' crap to legitimise their Vocaloid - couldn't they just Zip it up and put it on Kazaa or somewhere then?
How effective is this 'security measure'?

And when you say about the phoneme structures and all, Vocaloid does not allow you to type in 'Bay Bee' to get 'Baby' then? That's how I've been fooling SAM and other software in pronounciation skills so far. It may take time for users to figure this out, but I would hope it's somewhat possible with Vocaloid?

Fitch's nice song is over now, and now time to check out your vocaloid song!

Hey Ojuzu-
I'm glad you like Can't Explain, I spent a lot of time on that. My Mind Is Free was just a quick experiment over the weekend, and I expect to do a lot more with Vocaloid despite its limitations. I try not to approach it from the expectation that it will sound exactly like a human, but rather, it's this thing that has very human qualities, and is capable of a lot of musical expression. I'll post here when I get some more vocaloid music done, and I expect I'll get better at it with time...

Hey JDub-
Yeah I tried alternate spellings, I can't get it to start a word with "b" at all. Very strange.
bc

thebots wrote:

The Vsti is not worth using.

as much as i hate vocaloid, i have to say, i't been confirmed that there are a multitude of probs with the vsti version which yamaha are apparently hard at work fixing. so you may not have wasted your auth after all. and also, why a separate machine? what about the rewire version?

Hey ttoz-
I think the problem is how to get close to a realtime "performance" in terms of auditioning Vocaloid. It seems like the rendering is somewhat processor intensive and sonar is not happy to remain running with tons of plugins while vocaloid renders. Having it on a second machine makes it much easier to audition new vocaloid lines on top of a sequence without having to stop the sequencer or jump between windows. Hey, you're right, maybe there is hope for the vsti!!
bc

ok, well i listened to the track.

nothing against you, but as usual, vocaloid.

the two or 3 words that were spoken a couple of times, as usual, were almost impossible to understand.

this sounded no better than analogx's freebie "say it". sorry, but to me, it's looking less and less that there is any hope for vocaloid. yet we know people are buying it.

Well done thebots, I love your work!!! I am a fan

Like ttoz, I couldn't understand what Leon was singing/saying, but I don't think there was any deliberate effort to make Leon sing anything comprehensively anyway. If I can get Leon to sing 'baby' I'll email you the VOCALOID midi file.

BTW thebots, your MP3s do not play in my WMP "The input media format is invalid" is the error! I had to open them in Audition to play them.

FAYder wrote:

BTW thebots, your MP3s do not play in my WMP "The input media format is invalid" is the error! I had to open them in Audition to play them.

oh,ok, i thought there was something wrong with my system. mine gave the same eroor message. however they worked fine in the quicktime player

Good tune Bots! But Leon reminded me of an old vocoder sound or something? But impressive how you squeezed it in there.

Look foward to hearing more music from you - with or without Vocaloid.

J Dub wrote:

Look foward to hearing more music from you - with or without Vocaloid.

me too, i liked the other tune alot better. so preferably without vocaloid

Hey You Guys-
I'm glad you like Can't Explain, but don't be so hard on Vocaloid, I think that it's an instrument that takes time to learn, and I hope to do something more impressive with it soon. I'll definitely post on KVR and let you know! FAYder you're right it's not like I was trying for articulation of words, I don't think that's what Leon's good for (but a mysterious AI entity just emailed me a midi file for making Leon say "baby" and suggested I look more closely at the User Dictionary). One other problem is that the phoneme table in the user dictionary uses a font where a small "L" and capital "I" look exactly the same and this/these character is used extensively. The mp3s on www.thebots.net are all born in an old version of MusicMatch Jukebox, version 6 I believe.

Quote:

The mp3s on www.thebots.net are all born in an old version of MusicMatch Jukebox, version 6 I believe.

thebots, get yourself something else to make MP3 files - I can't listen to this either!

Baby should be simple - you need 2 notes and one will be "b @" or "bh @" (I think that's correct but check in the manual for the right phonemes) and the second note will be "b I".

I'll let you know if I can listen to the mp3s at some point....

What are you using for the voices on Can't Explain?

Sounds like a mix of vocoder, "found" snippets of dialog and some speech generators.

Complements for the song, it's really cool.

thebots wrote:

The mp3s on www.thebots.net are all born in an old version of MusicMatch Jukebox, version 6 I believe.

Please get LAME, which seems to be one of the best, and a GUI for it like RazorLame. Works like a charm and it's FREE!

hey guys,

i'm wondering wouldnt it rock to abuse vocaloid for doing funky vibes/basslines? i could imagine it would sound great to do some voweled bassline! someone should try it out (i dont have vocaloid)

tOm

jbraner wrote:

.. but check in the manual for the right phonemes.

NO!

Forget what the manual says a phoneme should be, use your ears, this is audio software not writing software. There are no absolutely correct phonemes, there are too many possible variations on the spelling of pronunciations. You have to go by what words 'sound' like. Then even if the word doesn't exist, you must create it and add it to the user dictionary. You must click the auto button before you enter the word, and add it to the user dictionary. Then manually change the phoneme spelling under the note.

So, if Vocaloid doesn't sing 'baby' to your satisfaction, try spelling it 'paby', you might find that the 'p' sounds more like the 'b' you want. You might be surprised what you get if you think laterally with Vocaloid. It's just another tool, that you have to find ways of manipulating to suit your needs. It's not bloody Speak And Spell.

Just dawned on me where the voice in My Mind Is Free sounds similar to! 'The Voice Of Energy' off Kraftwerks Radio Activity album.

Not finding fault with it or anything, just that I think the two sound somewhat similar.

Hey jbraner-
You're right, I'm an idiot. I forgot to put the space in after the "b" and vocaloid just wouldn't render it. Note that this is in the "phonemes" window. Vocaloid will not translate the word "baby" using its "Phoneme Transformation" function when you type it in the normal text window, which is why I originally thought this didn't work at all.
thanks!