Audio- Part 1
Updated: Jun 4, 2019
With development of Enchanted Kin, the game has very much become an audio centric experience which has led to many sound ideations. The change from AR to an Alexa driven game initially pulled focus on the story and its narration. Rubi focussed on this creation, by looking at book narratives and developing a story for the game. This was a descriptive storytelling model alongside tactile interaction, lead by a narrator.
During this initial stage Alex Ayling delivered a lecture to us about sound and how we could utilise this in our games. He explained the broad spectrum of game music cannot be specifically defined but that it typically leaned towards a cinematic sound. Of particular note, was the distinction between sound as a game informer such as a specific environment in video games or how board and cards games have tactile sounds and conversation as part of its soundscape. These both apply to Enchanted Kin and finding this harmony will need to be practiced. Additionally the environment of where the game is played will affect what sound we will produce for the game. For example the sound experience of a concert hall is very different from that of a forest.
In discussion with both Alex and Adam about the game we promptly decided that the standard Alexa voice was not appropriate for a storytelling experience. The voice needed more warmth and have more human quality to their voice. I found that this can be achieved with Alexa development but has a limit of 240 seconds per response, making this difficult with the linear, block narrative.
We knew that we wanted to assign a theme song to the game and have that lead through to the voice over; yet Alex showed us that there were further game enhancing opportunities through a soundscape. This lends itself to Rubi's narrative with location changes as well as player input, for example ambient noise and sound effects.
Since the sound session with Alex, the narrative of the game has adapted into a further interactive model. Inside the narrative the narrator shows through more personality, commenting on their surroundings. This has made the type of voice we're looking for change from a gentle story teller to a more characterful guide throughout the game. The narrator/character role implies that they are a family helper of the players, which brings a lot of sarcasm. When looking for a voice it's going to be important that we find someone that brings both instruction yet with a slight attitude.
To help further my understanding of working with voice actors, at EGX Rezzed I attended a Rezzed Session focusing on this topic, "Telling Stories Through Voice Acting". Dave Gilbert on the panel's role on the panel was to advise on designers role in directing voice actors, which I found particularly helpful. In terms of sourcing a voice, he explained that hiring from an audio reel often isn't best practice as they have specific or limited ways of working. Having face to face auditions with voice corrections is a better route and you should take note of their ability to act as oppose to just voice over.
Another great tip for voice, was to make sure that the voice isn't annoying in big dosses. Overly characterful voices become repetitive and can frustrate a player, it's best to tone it down. This is something that we will need to be careful with Enchanted Kin, making sure that the voice does not become overbearing.
When directing the voice actor it's import too not be too specific, as you can waste valuable recording time but context is most definitely needed. The actors need to know what their doing and why they're doing it in the game. A voice actor on the panel explained the best concise direction they received was "we need Maggie Smith in space". Actors develop with their character throughout the recording process, knowing them and what they would say. This can lead them to give script corrections for a more convincing and true voice.
Following this session I understand that the character's feelings need to show in the script. This will be very important for the sound of the game. The script and character over time is becoming quite sassy, yet supportive of the family throughout the narrative. We hope that this attitude will appeal to the target audience, bringing humour to the game. This adds more value to the game and will create a deeper engagement with the players. In order to achieve this I must ensure that I am developing the narrator and what the target audience prefers and enjoys.
This is showing a reoccurring overlap of my two roles, audio and interaction. Seasoned voice actor Pat Fraley, covers this crossover by defining character voice to form, as the sound of the voice, and content, as the thinking and feeling the psychology of the character. The audio acts as the form. This enables me to keep these developments parallel, hopefully learning to best create a voice model for Enchanted Kin. There are 6 areas of importance to develop this; pitch, pitch characteristic, tempo, rhythm, placement and mouth work. These are areas which I should identify with the voice that we decide to work with. Thinking about how the narrater would be leaning towards amongst them. I feel that placement will be really important for the narrator role and I think that the diaphragm, chest and front of the mouth will be key places to work. This has given me a more practical knowledge of voice and will aid me in conversing and practically directing, supported by the more psychological and feeling based direction I learnt from the EGX Rezzed session.
I've gone on to think about the characteristics of the narrator, which has helped me develop both sound and interaction, via the narrative. By assessing potential character traits of the narrator it has invoked ideas about how they should sound. The aiding purpose of the narrator is imperative and therefore the messages should be put across clearly. This means that the voice should have subtle hints of character. This could be achieved by small amounts of comedy in the script or an occasional snappy tone from the narrator, ensuring not to distract from the gameplay.
This led to me thinking specifically about the sound of the narrator. This led from the character traits from my initial ideas. The voice standard for Alexa is extremely flat and feels robotic, this is not what we want to achieve from the game and therefore voice-based on Voice User Interface would be inappropriate. However using the ideas from character saw me create a concept for witch-esque housekeeper. This would add further context for relaying information to the players as well as it informing the style of voice. I'd describe this voice as intriguing and could lend itself to sound effects such as cackles.
Interestingly during our initial playtest, the player in the target market preferred an older female voice for the role. They said that the original narrative suited the Mage type character from Warcraft and older-wiser voices such as Oogway from Kung Fu Panda or similar to Robert Carlyle. This demonstrated a need for voice that has warmth, for example the tapped R's in a Scottish accent. This could be achieved through several dialects and having a more mature voice, as suggested. A more mature voice, would act well as a companion creating player trust.
After discussing with Rubi, we felt that the sassy dialogue was working well for the game. To understand further how we could apply this through voice, I looked at fictitious examples of sassiness. I specifically looked at Edna Mode (The Incredibles), Squidward (SpongeBob SquarePants) and Mary Poppins, who I enjoyed replaying through online soundboards. They're all had quite an authoritative voice, sounding confident in what they know/observe. This could be a quality that lends quite nicely to the instructive companion narrator, which we attempt to replicate. This could be achieved by lengthly dialogue and patronisation. Interestingly, more dynamic range emphasised sarcasm rather than typical thoughts of monotone sarcastic comments. In particular, the character Squidward, dynamic range formed rise and falls that were quite comedic. Providing comedic moments in the script could prove vital to attracting our target market of 8-12 years olds. Demonstrating the further importance of the dialogue we create and the way in which it is portrayed through vocals.
I have thought a lot about how sound can be used to enhance the player experience throughout Enchanted Kin. The voice is the core experience but needs to be elevated to ensure that families take on this role of sorcerers and feel emotionally connected to the story. With Alex's talk and development of the type of narrative interaction, I have identified that the game requires 3 core sounds: voice, sound effects and ambient sound. The voice will guide the narrative, further enhanced by sound effects.
Compared to our narrative heavy initial script, development has seen the script become more as an aid to the plater. It puts the experience more into their hands, with the interaction feeling more conversational and less one sided and sound effects enhancing the narrative plot line. Additionally to this I have thought about how sound effects could signify interaction for the player. As this is voice led experience with tactile input, key tones that assign to each one can become notable to player, for example a pick up card tone. The ambient sound, will imply the change of environment with each level having a specific sound.
With these decisions, I needed to have some practice with audio recording and editing in order to compose the audio clips right.
Generally the idea is that Enchanted Kin would be played in home environment such as a living room or kitchen, but for the purposes of presenting the game, we will be showing at degree show. This very much changes the volume expectations of the game. When editing it is advised that you edit with headphones and with an average volume of -12dB. This lies in the centre of expected volume in order to submit audio for Alexa games to be certified and authorised by Amazon.
I tested out these editing settings (with a mix of vocal and ambient sound), through headphones this sounded good, as well as sounding good when played through an Alexa. However I'm concerned about the volume when in a crowded room, particularly for the private view shows. I conducted some further research and found that their is a batch editing setting I can apply to the files, via Adobe Audition, to create a standard volume. This has a preset of "ITU-R bs. 1770-2 Loudness", which is typically used for broadcast. However testing this out proved that the volume was just far too loud. The key is to go for a lower standard volume such as "perceived loudness". This setting is particularly helpful as it is designed to the human ear and our lack of hearing range. I found that this with a low laptop volume gives desired effect.
// I'll warn you turn your volume down extremely low, this is a volume test and has been proven to be far too loud! DANGEROUS //
As well as the batch volume control, I also found out about a batch process that can be created like a recipe with your audio preferences. It is controlled by a recording of your editing, or one of their presets. This is really good for work flow and enables you to experiment faster and test out across a wide variety of specific edits i.e. a change from -15 to -19m dB. I used this to apply effects across files to see if it would work as a collective body. It also ensured consistency and is good for workflow. This will be helpful for time efficiency when editing the final recordings for Enchanted Kin.
I knew that despite our best abilities we wouldn't be able to record in perfect circumstances, so I believed that background removal would be a good place to start editing a sample file. In terms of research on editing, I jumped straight into practicing. I edited together a first piece of interaction flow of the initial opening, along with substitute sound effects in Audacity. I very much focussed on the background removal and cropping consistency, to ensure that the audio flowed like conversation in real time. After I completed this tutorial, I tried to export it, however I somehow seem to have an error and am unable to export a variety of normal audio file types from audacity. This has led me to change up my workflow and try some new software. As I'm familiar with the Adobe creative suite, I opted to use Adobe Audition. Fortunately for me the noise reduction process is extremely simliar .
I learnt this from youtube tutorials for both platforms (i.e. https://www.youtube.com/watch?v=10FFKl_0GSA for Audacity and https://www.youtube.com/watch?v=54VsPQBlOpY for Adobe Audition. You simply select a small section of disturbed background moments and capture a noise print. When you select all and apply noise reduction, you see a particle graph of the selected noise, and it simply removes it and does the same to all area of its file.
Unfortunately I lost an afternoons worth of work from the Audacity occurrence, however that afternoon I received correspondence from our new musician/composer for Enchanted Kin which makes up for that lost time with complied and intentionally produced audio files. From this, I learnt that it is always good to export early to ensure that you are using the right service, to avoid disappointment of losing a file and time.
As well as the background removal and collective volume assignment, I discovered more about the figures and ranges that are appropriate for voice work. I found a lot of videos on youtube but I was frustrated as looking to edit voice work, as I received an influx of videos demonstrating editing for voiceovers, specifically for youtube commentary. I felt that these were supplementing a visual experience compared to our audio centric game. For example in this tutorial , which helped me understand how to create a clean and professional voice file; I felt that the limiting of dramatic dynamics made the voice more volatile or even sometimes the opposite and robotic. We established that we want our narrator to have warmth to her voice, so I decided not to include that step along with a few other ones relating to video-focussed audio editing techniques. Going through tutorials I found that equalisation (EQ) and normalisation were the must have editing techniques. I used the parametric equaliser and normalisation at 93.6 to ensure that the audio was clear. This resulted in the difference below. I applied this with noise reduction after to avoid static like sounds and created a recipe which can be used on the final audio files when mixing with music.
With voice being the core interaction between the players and the game, having audio that is well put together is of upmost importance, by applying these techniques it will further their engagement and bring the magic further to life.