Academic Research Result Release (briefly)

About 1 years ago I started writing an academic essay about a relationship of video game sound and players culture. For this reason I developed a video game with my younger brother as an experiment tool, published here and let everyone to play. The dissertation is already completed and being graded. Those research result I hope I could share to someone who interested in game sound design, or even game developer as some kind of reference or further studies. 

A whole research is divided into two part: an observation part and an experiment part. The observation part aim at finding any characteristic of American and Japanese game sound respectively. And the experiment part aim at find out any preference on game sound are existed between native Chinese speakers and English speakers. Finally it would compare two data and conclude a relationship among them. 

"Dream Within a Dream" is developed to conduct an experiment part of this research. Below I just point out something that related to this part from the dissertation. If you feel interest with my whole research please feel free to contact me. I could give you the paper. And you may use my paper with my permission. 

Methodology

In the video game, players are told to control an avatar to defeat some monsters which front of them. There is a scroll bar on right hand side of the screen which could change attack style of the avatar. Each time subjects go to next level would be told to scroll the bar. Visual animation and actual damage point of its attack is no difference after scrolling but the mix of sound effect does. Subjects should consider which attack style could benefit their combat (with only sound mix different) and select his suitable option. 

The three options from the attack style scrolling bar, their sound mix are reference from the below observation samples:

These three attack style are representing the audio mix preference which found in the observation part. Slash A is representing the “Mid>High>Low” style (Mid tone loudest, high tone middle, low tone weakest), with 9/10 of chance giving an attack yelling sound; Slash B is representing the “Mid>Low>High” style, with 7/10 of chance giving an attack yelling sound; Slash C is representing the “High>Mid>Low” style, and it would only giving an attack sound without yelling voice. 

Players would be told to control two avatars with different art style. After they finished one level, the game would be changed to another art style with different avatar. After finished it would then go back to the first art style. There have totally eight levels, each avatar is going to play four times.

This game is having four background music, composed by two composers who don’t know each other. They are requested to composed two background music in fighting scenes, one in traditional orchestral style and one in rock style, with a same melodic line, to an specific avatar. Those music would then being put into a specific sequence that players couldn’t realise any relationship with plot or the monsters.

After playing the game, subjects would being requested to fill an online questionnaire. Those questions are as below:

The questionnaire is also collect respondents some personal information like age, gender, playing with earphone or not, and most importantly, their mother language.

The result of observation would measure a mean and a standard deviation of those data; both observation and experiment result would being adopted a Chi square test to test whether answers distributions of two categorical variables are independent from each other. Similar method is used by Xiao Hu and Jin Ha Lee in their research report that released in 13th International Society for Music Information Retrieval Conference (ISMIR 2012). 

Overview

There have totally 52 complete responses. 28 of them speaking Chinese and 19 of them speaking English, the others are speaking other language. So there are 47 responses in compare the cultural group. Among them there have 20 female and 32 male; 29 of them play with earphone and 23 of them play without earphone. The result review would compare the data in cultural group, gender group and with/ without earphone playing group.

Result  Discussion

The above analysing suggests that cultural background difference among players do not influence their perceiving on video game sound very much, but gender factor influencing more. When asking about the highest damage attack, both the first hero and the second hero their answers distributed were very different with a low p-value in gender group. The second hero question even have a p-value of 0.046, which cross the line of being significant different. In opposite none of the answers distribution in cultural group its p-value is lower than 0.05.

Results of the experiment shows that respondents have different preference of the game sound on the first hero and the second hero, especially on comparing the gender group and cultural group. Actually they look obviously different. The first hero is look like a 16 years old teenager while the second hero is look more childish and tiny (some respondents comment it look so weak). Putting two different art style avatars for experiment is aim at prove respondents making their decision only determined by the audio parameter whether what the visual style the avatar are. However after experiment it suggests that they do change their preference when visual style change. Review the result, when asking about the suitable sound of the first hero, the major selection of both gender group and cultural group are “Slash A”, while asking the second hero, however, their selection changed to “Slash B”.

In gender group comparison, male is more tend on selecting “None of them/ I don’t know”. This is the second major answer of the first hero question, and a major answer of the second hero question. For these, it may being concluded that male is more rationality and would more doubt about situation not only rely on game sound; For the last question asking about how game sound benefit on their performance of playing video games, more Male rate “4” while female rate “5” more. It could be interpreted that female are more rely on game sound than male.

In the earphone group, it is observed that the wearing earphone group would averagely rate more score than the non wearing group in the last question. It is no doubt that players wearing an earphone could listen more detail about game sound, that they could more enjoy what game sound give them and benefit their performance of game playing.

Because of the limitation of sample size, the age group would only be analysed the group of “13-17”, “18-22”, “23-27”. And here it would pick up some spacial points for discussion. Most of the respondents in“13-17” group are English speakers. Their answers distribution were generally match with the English group. However when asking them the last question, 50% of them rate “4” rather than “5”. It is difference from the English speakers who generally rate “5”; There are about 2/3 of “18-22” group’s respondents are Chinese. Their answers distribution normally follow the Chinese group. But in the third question when asking about a suitable music for the second hero, only 44% of respondents selected “the second one”, while Chinese group have 68% of them selected “the second one”. The “18-22” group is generally rate a higher score than Chinese group in the last question; There are nearly 2/3 of “23-27” group’s respondents are male. Again, it generally follow the male group’s selection. Beside, on the third question asking about a suitable music for the second hero, there are 77% of respondents selected “the second one”, while male group have only 56% of them selected “the second one”. For the above phenomenon I think I could not explain a reason very accurately. For the 13-17 age range respondents I think a most suitable explanation would be they more interest in visual element rather than sonic element when playing video games.

Review among those samples, it is found that “Slash A” is the most often selection when asking about the highest damage of the first hero (Hero A) . However the answers distribution of the second hero are more different without a major option. It is guessed that combination of attack yelling voice and sword slashing sound give human being more impact than just a slashing sound so they rather selecting “Slash A”. For the second hero (Hero B) question, as it looks so weak, different from what we expected a hero should be, visual image influence their judgement that they hard to find out which attack is “the strongest”.

For the question asking for a suitable background music of the second hero, samples’ answers were uncertain cause by a failure question design.  

Conclusion

After analysing it indicated that video games made in different countries might have their own audio mix preference. Players with different cultural background might have their own audio preference too. But these preference are not obvious. Thus game sound designers aren’t influenced by players’ preference and do their audio mix, they rather determine by how the character was and designed the sound effects. On the other hand, cultural background is not an only factor influence players perceive game audio, gender, and how they receive the game audio (wearing earphone or not) are also influence them in perceive game audio. Moreover, as it is found different cultural group might have their own preference in perceive game sound, it could try figure out if this phenomenon is already exist among American and Japanese pop songs. If the result is positive then it could be assumed that different cultural background might exist their own sonic preference.