The 'Why don't you use Famicaloid?' argument

General discussion about tracking, help and support.
User avatar
w7n
Posts: 237
Joined: Fri May 15, 2015 1:37 am
Location: Nanamori-Chuu
Contact:

The 'Why don't you use Famicaloid?' argument

Postby w7n » Tue Aug 29, 2017 1:50 am

Chii--su, w7n here.
Now, I'm gonna give a little commentary on cover 39 of this Famicompo Pico. This cover is not mine (none of the entries with vocal simulation leads as a main feature is mine), so it's not like propaganda of this cover 39.

First to note: This cover is good. In fact, REALLY good. But there are some basic mistakes in simulating the lead vocals. Had the lead vocals been cuter, I would have voted it among the top ten.
So you would guess, the next thing I'm gonna say is 'why don't you just use Famicaloid', right? Well, not really. Famicaloid is NOT your saviour. However, using Famicaloid would help avoid a few basic problems in vocal simulation.

What basic mistakes has this entry made?
1: The vocals seem to be sampled with Hertz's Lua sampler. As a result, even a single vowel's waveform change constantly and the vocals sound muddy. Well, you can try to explain that it's attempting to recreate the 'singing squids' feel, but the original ain't like this.
To demostrate this, let's experiment with a simple N163 sine wave
8 9 11 12 13 14 15 15 15 15 14 14 13 11 10 9 7 6 4 3 2 1 0 0 0 0 1 1 2 4 5 6 (Wave 0)
Now move the two leftmost values to the far right:
11 12 13 14 15 15 15 15 14 14 13 11 10 9 7 6 4 3 2 1 0 0 0 0 1 1 2 4 5 6 8 9 (Wave 1)
13 14 15 15 15 15 14 14 13 11 10 9 7 6 4 3 2 1 0 0 0 0 1 1 2 4 5 6 8 9 11 12 (Wave 2)
...
Make five such waveforms, and define a wave sequence | 0 1 2 3 4 3 2 1 . Now even this sine wave sounds muddy.
So the lesson: DON'T use Hertz's Lua sampler if you want to simulate vocals. Sample them manually, make the waveforms' phases look alike, and don't change waveforms too often esp. with FDS.

2: The same waveforms are used across a wide note range.
This is also one of the critical problems with Wiz.'s FaMiC set. Cover 39 didn't avoid this problem either, and as a result, the vocals in the lower note range sound like old grannies. In my Famicaloid modules, one vowel waveform is normally used across a maximum of only 4 semitones.

Now, why isn't Famicaloid your saviour? Let's listen to a brainless recreation of the vocals in the same song.
Calamari Inkantation_Vocals.ftm
(15.9 KiB) Downloaded 43 times

Yes, this Famicaloid simulation is LIFELESS. The nuances in the vocals require more effort like pitch bends, vibratos, tremolos, volume envelope, and more sampling if possible/necessary. (In this case more sampling is the key, some waveforms just don't fit.)
But still, this module avoids both basic problems that cover 39 made.

So anyway, my personal advice is to start with Famicaloid if you want to attempt vocals, and pay good attention to the nuances of the vocals. This way you can both avoid basic problems and have a good grasp of some mechanics of vocals.
nsf.nesbbs.com under reconstruction, better player might be available in the future.

VRC6 PWM is GAY because it has colours of the rainbow in the NSFPlay keyboard visualizer.

User avatar
w7n
Posts: 237
Joined: Fri May 15, 2015 1:37 am
Location: Nanamori-Chuu
Contact:

Re: The 'Why don't you use Famicaloid?' argument

Postby w7n » Tue Sep 19, 2017 1:51 pm

Yep, that mmr2003 thread totally drowned this. Posted this at the worst time possible.
nsf.nesbbs.com under reconstruction, better player might be available in the future.

VRC6 PWM is GAY because it has colours of the rainbow in the NSFPlay keyboard visualizer.


Return to “General Talk”

Who is online

Users browsing this forum: No registered users and 1 guest