Sunday, June 26, 2022
Saturday, April 30, 2022
We recorded the parts in two days and mixed it and posted it. Which leads me to the subject of today's post: Never do that!
Why? As it turned out, the mix levels were off. What sounded good that day turned out to be not so good. Why does this happen? Well, the problem with recording/mixing for an entire day is, after a while, your ears suffer from ear fatigue. Don't consider the song "finished" at the end of a long recording/mixing session. Put it aside. Come back to it after a day, or even two days. You want to listen to the mix with fresh ears.
The problem with our mix was that the lead vocals was too loud. Also, some of the other parts like the acoustic guitar and keyboards were too soft. It's funny how the levels sounded good to me at the end of the day, but the very next day, I realized that the levels were not correct. It's a rule I should always follow--always check the mix again the next day before you consider the mix to be finished--but I keep breaking this rule because I'm so impatient!
But don't be like me! Take your time! Let the mix "rest" and come back to it the next day.
Here is the song with the incorrect mix levels.
Can you hear it? So let that be a lesson for me. I don't have really good ears--I'm not one of those people who can really hear things like legendary mixers Bob Clearmountain or Chris Lord-Alge. I need to be pretty slow and methodical and not rush things, but it's hard for me because I'm impatient and just want to get things over with!
Thursday, January 20, 2022
You'll need to write a song every two days, so you can't get bogged down trying to make it "perfect." I'm planning to just construct simple songs--verse/chorus/verse/chorus. Nothing too complicated, not too many chord changes.
I know Megumi doesn't want to always have to set up her guitar to record a backing track, so she's planning on writing many of the songs just using the built-in virtual instruments (such as the piano) in the DAW. In fact, she downloaded a whole mess of pre-constructed midi chord files so she can just drag and drop the chord in. Drawing in the notes for the chords isn't hard, but it gets tedious so this will sure make things go faster.
You can download these chord files at:
We've already started practicing writing songs! Synthesizer V Solaria will be singing our demos for us! Check out our first "pre-FAWM" song:
Tuesday, December 28, 2021
In my last post, I posed the question: Is it time to ditch Vocaloid in favor Synthesizer V? Six months later, I can confidently answer that question with a resounding YES.
First, for someone starting from scratch, the cost is ZERO. Synthesizer V Studio Basic and all the lite voicebanks are free.
Second, moving up to Synthesizer V Studio Pro to take advantage of all the great features of the full version and buying the full version voicebanks is very reasonable in terms of cost. The editor is only 90 bucks. You also aren't forced to buy a bundle of voicebanks--you can just build your library with the ones you want.
Third--and this is a huge reason to throw your hat in with Synthesizer V--with the 1.5 update, through the magic of AI, the full version of the AI voicebanks can now sing in English, Japanese, and Mandarin Chinese without manually messing with the phonemes. That basically triples your voicebanks! (Note: this feature is only available with Studio Pro and the full versions of the AI voicebanks).
Of course, Eleanor Forte singing in Japanese or Chinese won't match up with a true native singer in those languages, but she is surprisingly good at it. Definitely more than good enough to use her as a backup singer in those cases. And the same is true for Koharu Rikka singing in English or Chinese and all the other voicebanks.
The case of Tsurumaki Maki AI (JPN) is interesting because I've read posts from people saying that her singing in English sounds exactly the same as Tsurumaki Maki AI English!
And if this isn't enough of a reason, in January, Solaria will be released. Solaria has a powerful voice. Have you heard her demos? She can really belt out a song. Think power singers like Linda Ronstadt or Kelly Clarkson. And, a male English voicebank is in the works.
There really isn't a reason NOT to move on to Synthesizer V. Although, honestly, there is a huge contingent in the vocal synthesis sphere that likes the older stuff--I'm talking about Vocaloid 2 or 3 voicebanks. I just think that's kind of weird. It's like using Windows 3.1 instead of Windows 10 (and soon, Windows 11). But hey, you like what you like, right?
Tuesday, June 22, 2021
I was impressed by the ease of use. I think the interface is easier to use than the Vocaloid 5 editor or the Piapro Studio VST. The only thing that made me hesitate to buy Synthesizer V Studio Pro was the lack of English voicebanks. They have Eleanor Forte lite, and initially, I thought it wasn't really suitable for most genres, and I have since come around--I think she does very well in specific situations. You just have to pick the right song for her.
So for the longest time, I was sitting on the fence. Then AH-Soft announced the Tsurumaki Maki voicebank--in both Japanese and English! That was the game-changer for me. I went ahead and purchased Synthesizer V Studio Pro and then bought Tsurumaki Maki AI English and also the Koharu Rikka AI voicebank. It was a tough choice for me between Rikka and Saki because both sound really good. I felt I should have one full version Japanese voicebank. And you know what? That not only gives me the option of creating songs in Japanese but also in English because the Japanese voicebanks can sing in English pretty decently!
What I like about Synthesizer V is the ability to draw parameter curves with much more control. Vocaloid 5 has the Attack/Release effects--which are great, but sometimes you want to manually draw in a parameter curve, and although Vocaloid 5 kept the parameters feature, drawing in the curves is really clunky.
Synthesizer V, however, does it very smoothly, and in the case of Pitch, you can draw the curve right on the piano scroll over the notes. That is a huge advantage.
Synthesizer V does not have a note preview though. I hope they can add that in in the future. Because what that means is that I constantly have to click on the keyboard to check the note rather than just hearing the note as I draw it in.
Most of the voicebanks have two options--Standard and AI. I prefer the AI version. The Standard version sounds more like the traditional Vocaloid voicebank. The AI version is inching ever more closely to a real human voice. The AutoPitch feature available in the AI version is KILLER. Let me say that again. The AutoPitch feature available in the AI version is KILLER.
So should you buy Studio Pro and the full voicebanks? The free versions are all very usable. If you are a hobbyist, it's fine to use the free versions. In the context of a mix, you probably won't really notice it. Well, unless you make a direct comparison to the full, paid version. There is a difference. I can hear it easily. The full voicebanks are sampled at multiple pitches whereas the lite version is sampled at only one. This means the voices can get pretty thin and raspy when you go outside the range.
Listening to Tsurumaki Maki, she sounds a lot warmer than her lite version. There is a richness in the tone of her voice that is lacking in the lite version. Again, this is not to say that the lite version is bad--on the contrary, it's very, very good. But if you want your song to sound its best (and also do commercial stuff with it), you should buy the full versions of the voicebanks.
I still have all these Vocaloid voicebanks, so I'm still going to be using them, but...it doesn't make sense to me to invest in any more Vocaloid voicebanks. Unless Yamaha wakes up and makes Vocaloid 6 AI, that is. We'll see if that happens.
For now, the future is extremely bright for Synthesizer V.
Here is a cover of We Are the People (Martin Garrix) sung by Tsurumaki Maki AI English!
Friday, February 19, 2021
You know, adjusting and manipulating pitch is probably the single most important thing you can do when tuning Vocaloids in order to make them sound realistic.
The problem, though, is that doing it by hand (drawing pitch curves) is extremely tedious, not to mention, very inaccurate. It involves a lot of trial and error until you get it to sound the way you want it to.
This is why I love the Vocaloid 5 Editor. With Attack and Release effects, you can save a lot of time and cut out a lot of that trial and error. And, if you don't find an attack or release effect that is quite what you're looking for, you still have to old tools available to you to manually draw the pitch curve.
Unfortunately, Yamaha does not sell the Vocaloid 5 Editor by itself. You have to buy a package, either Standard or Premium. Buying a package saves you a whole lot of money in the long run because--in the case of the Standard package, you get four Vocaloids with it. But it's a pretty steep initial investment.
Whereas, if you buy a Crypton Future Vocaloid (like Miku V4 or Rin), they give you the Piapro Studio VST plugin. It's not as full-featured as Vocaloid 5, but you can still pretty much do everything you need to do with it.
Let's take a look at the Enka attack in Vocaloid 5. It's one of my favorites! Here is what the pitch curve looks like.
|Enka attack in Vocaloid 5|
Let's see if we can replicate it by drawing the pitch in Piapro Studio!
|Replicated Enka attack in Piapro|
Not quite the same, but kind of similar! That's what you get when you're trying to draw a curve with a computer mouse!
Friday, December 4, 2020
Megumi met this girl--a kid, really--I mean, she's only 16 years old. Her name is Hatsune Miku. She loves to sing, so Megumi invited her over to one of our band practices. She's a nice kid, has kind of a quirky voice, and she has very long blue hair tied up in pigtails. My first thought was, "What the...??"
Well, long story short, we wrote a song and Miku is singing it. What a crazy world, eh?