The easiest way to say what we had to do to make this work is by putting it down in a step-by-step.
1. Record audio voices of all characters conversing with each other at the same time to avoid audio prints that don’t match-up.
2. After importing audio files into the computer, each individual voice must be saved as a separate audio file for every single time that character speaks. For example:
Vin: I have to do what? (audio file 1.1)
Nat: I need a file for every single part of the conversation. (audio file 1.2)
Vin: State won’t let you just run the whole wav file? (audio file 1.3)
Nat: No. (audio file 1.4)
Vin: So I need to make a new audio file for every part of the conversation. (audio file 1.5)
Nat: Yes. And then I have to upload each individual file every time a character begins to speak. (audio file 1.6)
Vin: That sucks. (audio file 1.7)
Nat: I know. (audio file 1.8)
That short conversation is now 8 different audio files.
3. So then each individual audio file must be imported into State whenever a character begins to speak. This process will then remove all the continuous room noise that was present in the background and create a horrid empty space of sound between one character having finished speaking and a new character beginning to speak.
4. Export the file from State.
5. In Premiere I exported the audio file to fix.
- Using Soundbooth I lowered the room noise a bit from where the characters are speaking. I did this using the previously blogged about method.
- To help with the cavernous sound of the recording I added an EQ Graphic filter and selected Presence.
- I also had to export some of the room noise to fill in the spaces between the characters talking. I took the few seconds of exported room noise and pulled them into Garage Band to create a clip of room noise that was several minutes long. I probably could have done this in Soundbooth, but since I already know how to do it in Garage Band, and I’m really not in the mood to learn a new trick right now, I made it in Garage Band.
6. I pulled the new audio into Premiere, including the room noise audio as a constant background hum, and then proceeded to edit as normal.
Our video tutorials have an interaction with two people. The first thing we did was record the actors directly into State separately and then output the file. This is the wrong way to do this. The final product had microphone noise. The room noise from one actor to the other didn’t match. Whenever the actors weren’t talking there was no room noise and instead a distinctive complete lack of sound. One audio was higher than the other.
It was a disaster.
Yesterday we re-recorded with both voice actors in the same room recording to the same device. We left off at least 15 seconds at the beginning and the end of the recording for just some room noise. I edited the audio wav files to the best clips and have sent them out to be imported into State. We’re hoping this will be an easy transition. I downloaded a wav file offline yesterday and State couldn’t seem to pick up the voice. When I did the same thing with a wav file taken from State it worked fine.
We used an M-Audio MicroTrack II Mobile Digital Recorder for the wav files.
First trials with the Xtranormal animation tutorials went reasonably okay. The biggest problem seemed to be the robotic voice provided by the program. There are very few options in each language and what tends to happen is that some words are hard to understand in the robotic voice the first time around. We want the videos to be understood by all and if the native English speakers are having trouble with robotic enunciation then… well.
So for the second draft we are going to be trying out a recording of live human voices. The xtranormal desktop software, State, allows a user to record a voice into the program and the animated person follows along to that recording instead of the robotic voice. This sounds like a good alternative. I haven’t finished putting together the new drafts, but here are some issues that have been raised with the use of the human voice:
1. Any editing would require the person who voiced it to come in. Otherwise you might potentially have to edit every single video with the new voice.
2. the robotic voice is something that whoever is doing the animation can work with on their own instead of having to wait for a schedule fit with the voice actor.
3. The librarians doing the voices are talking considerably faster than the robotic voice. This might give rise to a different problem. As the robotic voice was difficult to understand, a fast human voice might be difficult for our target audience to follow. As in, it’s hard to absorb information that you are hearing for the first time when it’s given too fast. My opinion.
So yesterday I downloaded the State software from xtranormal so I could make animated videos from my computer directly instead of from the website. Here are some findings followed by the video.
- State appears to offer a wider range of control over the animation than the online website. I say “appears” because both options are so difficult to navigate sometimes if you want to personalize the video that it’s hard to tell exactly what my options are.
- The downloaded software will only allow you to add assets that you have previously purchased. Different animations come in different package sets (like robots, stick figures, homeboys and so on). If you haven’t used anything from a previous set before those characters won’t show up in options. The same goes with languages. In order to reach the full potential of assets, you have to log into xtranormal through State and purchase the assets.
- I can’t figure out a way to publish my video from State into my xtranormal account. There has to be a way, right? But the website is so difficult to navigate I might have given up on what seemed like a lost cause. State allows you to publish to Facebook or Youtube. I exported to an mov file (a very low file size mov file) and uploaded to vimeo.
- The help forum is not very helpful
- When trying to add extra movements from the camera or characters, I have a hard time understanding the timings and nothing seems to really work the way I ant it to except by accident.
Final Thoughts: xtranormal and State are not for control freaks. There’s a lot of stuff you’re just going to have to let go. Also, despite the mostly good enunciation of the voices, perhaps a voice-over would be best.
Password protected video on vimeo