Dry Vox Authoring
- This topic is empty.
-
Posts
-
I’d like to start this topic to discuss various aspects of Dry Vox (as it applies when vocals are not separated.)
1 – Are Dry Vox necessary to release through C3?
2 – How does it affect the scoring? Does the game analyse the sounds made by the dry vox and compare that to what you are singing in order to determine how close to the words you get? As it pertains to note tubes, I had always thought the game only scored pitch and length and that you are allowed to sing anything you want or even hmmm. What else is there to score?
3 – What is the best approach for creating them? I’ve read that some people just sing and record the song. Others talk the parts. Either way, is it important to sustain the entire note? Or is it just a matter of providing some sort of sound at the start of each lyric event?
4 – Continuation of question #3… How does the game know to keep the character’s mouth open for the whole note? I assume it interprets the dry vox at the start of each lyric event to determine lip sync but then use the midi note tube to determine the length of the note. Is that how it is done?
I’d like to start this topic to discuss various aspects of Dry Vox (as it applies when vocals are not separated.)1 – Are Dry Vox necessary to release through C3?
2 – How does it affect the scoring? Does the game analyse the sounds made by the dry vox and compare that to what you are singing in order to determine how close to the words you get? As it pertains to note tubes, I had always thought the game only scored pitch and length and that you are allowed to sing anything you want or even hmmm. What else is there to score?
3 – What is the best approach for creating them? I’ve read that some people just sing and record the song. Others talk the parts. Either way, is it important to sustain the entire note? Or is it just a matter of providing some sort of sound at the start of each lyric event?
4 – Continuation of question #3… How does the game know to keep the character’s mouth open for the whole note? I assume it interprets the dry vox at the start of each lyric event to determine lip sync but then use the midi note tube to determine the length of the note. Is that how it is done?
1. Yes.
2. If the dryvox pitch doesn’t agree with the MIDI pitch, RB3 widens the scoring window for pitch. The scoring is only based on pitch. The vowel/consonant sounds are used only to move the singer’s mouth.
3. Either singing or talking is okay. Talking will have the side-effect of making scoring easier, which is fine. The important thing is to line up with the actual singer, so the RB3 engine moves the in-game singer’s mouth at the right time.
4. The MIDI notes aren’t used for animations.
I’d like to start this topic to discuss various aspects of Dry Vox (as it applies when vocals are not separated.)4 – Continuation of question #3… How does the game know to keep the character’s mouth open for the whole note? I assume it interprets the dry vox at the start of each lyric event to determine lip sync but then use the midi note tube to determine the length of the note. Is that how it is done?
The vocals chart has nothing to do with the lip animation. Analysis software produces an animation file for lips from the sound of the dryvox.
You can abuse this in funny ways by recording an “eee” sound where you want the singer to clench his teeth or a “mmmm” sound if you want to make him frown. These sounds don’t get played back in-game but drive the character’s facial animation.
Thanks for the info. One thing that is still not clear though is what makes the animation hold the note for a long time? Does the dry vox have to also match the length of the note?
In an ideal world, the dry vox is the original audio recording from the vocalist without any processing or effects.
In a C3 world, the dry vox is a recording of you signing along (or speaking along) to the song.
In both cases, the length of the note in the dry vox should match the length of the note being charted (or be longer, if say you’re clipping a fast drop at the end or soemthing). The in-game mouth animations will match the recording length — the only time anything in the .mid matters is when designating HARM2/3 singalongs iirc, and even then it might not with a separate dryvox (I’ve never tried without it).
I’m adding a 5th question:
5 – When using LeFluffie, shouldn’t I be able to get a dry vox file or track out of an rb3con file? Is it contained in the mogg file? I tried opening “Nightswimming” and opened the mogg but it only contains the 44khz song. Enter Sandman contains all the stems but not any 16khz dry vox track. So can the dry vox track be obtained using LeFluffie?
I’m adding a 5th question:5 – When using LeFluffie, shouldn’t I be able to get a dry vox file or track out of an rb3con file? Is it contained in the mogg file? I tried opening “Nightswimming” and opened the mogg but it only contains the 44khz song. Enter Sandman contains all the stems but not any 16khz dry vox track. So can the dry vox track be obtained using LeFluffie?
Nope. Magma bakes the dryvox down into the milo_xbox, which I don’t think anybody’s ever bothered to figure out. It’s likely just a series of pitch window values and animation cues (mouth open, different consonants, etc).
If the dryvox pitch doesn’t agree with the MIDI pitch, RB3 widens the scoring window for pitch. The scoring is only based on pitch. The vowel/consonant sounds are used only to move the singer’s mouth.
So suppose someone just talks through the whole dryvox.
Since the scoring window will be spread wider than my ex’s legs, are we, in a sense, inflating singers’ scores as opposed to creating a dryvox close to original pitch? (not that anyone cares about vocal scoring more than singing along to their favorite custom).
It doesn’t make that huge of a difference for pitch detection. From what I’ve seen it hardly makes any difference at all, or might actually be broken since I’ve never gotten the “your dryvox is too different from the midi” error that HMX claims exists.
If it does anything at all, all I think it does it let a tube give a little bit more phrase progress if you hit it, like talkies flagged with “^”. The actual window for how off you can be doesn’t appear to change (otherwise tuning_offset_cents would be redundant and songs recorded with a talking dryvox would be hittable at any pitch).
Nope. Magma bakes the dryvox down into the milo_xbox, which I don’t think anybody’s ever bothered to figure out. It’s likely just a series of pitch window values and animation cues (mouth open, different consonants, etc).Well that’s good to know. I can stop worrying about my awful singing abilities being extracted and published for all the world to hear.
Well that’s good to know. I can stop worrying about my awful singing abilities being extracted and published for all the world to hear.
Leaving secret messages?
lolI hope it’s okay for me to bump this old thread but I had a couple of questions about recording a dryvox track myself.
First, since I’m quite a bad singer I wonder if anybody knows if it’s better to still record a approximately sung part or just go with talkies in order to keep the scoring as close as possible to the midi notes? I know that if I record the sung part, it will at the very least have a lot of wavering on any long notes and probably some notes missed at the beginning so I wonder if this would actually mess up the “scoring comparision” worse than just talking?
Second, the song has an octave harmony, which I cant sing but since Rock Band automatically adjusts to your voice range, does it matter at all, if I just copy paste the harmony tracks?
First, since I’m quite a bad singer I wonder if anybody knows if it’s better to still record a approximately sung part or just go with talkies in order to keep the scoring as close as possible to the midi notes? I know that if I record the sung part, it will at the very least have a lot of wavering on any long notes and probably some notes missed at the beginning so I wonder if this would actually mess up the “scoring comparision” worse than just talking?Keep it simple and go with talkies. The result will be that it’s easier than it ought to be, but who cares?
Second, the song has an octave harmony, which I cant sing but since Rock Band automatically adjusts to your voice range, does it matter at all, if I just copy paste the harmony tracks?You should probably shift the harmony tracks down an octave, so it appears distinct on the screen. In practice mode, the guide pitch that RB3 plays matches the MIDI note octave, so it’s nice to have that be correct too.
Second, the song has an octave harmony, which I cant sing but since Rock Band automatically adjusts to your voice range, does it matter at all, if I just copy paste the harmony tracks?You should probably shift the harmony tracks down an octave, so it appears distinct on the screen. In practice mode, the guide pitch that RB3 plays matches the MIDI note octave, so it’s nice to have that be correct too.
Oh, I did that in regards to charting of course. I meant copypasting the dryvox takes, but since I’ll do it as talkies as recommended, I assume it shouldn’t have any effect.
Edit: Or now that I went and did the above, it occurred to me that I could probably just use the same file for both harmonies.
- You must be logged in to reply to this topic.