Investigating gameplay and gaming experience in an audio-only puzzle game using stressful versus calm ambience

experience in an audio-only puzzle game using stressful versus calm ambience

Catherine Crépault Wibe

Audio Technology, bachelor's level 2021

Luleå University of Technology

Department of Social Sciences, Technology and Arts

Abstract

This study investigates if gameplay in an audio-only puzzle game differs when the player is

feeling stressed or calm. This was done by conducting a test where 14 participants played an

audio-only, puzzle game with two levels. One level containing stressful ambience and the other

level containing calm ambience. The players were timed during the playtest and the results were

analyzed. The results showed a significant decrease of played time in the level containing

stressful ambience. The study has also investigated whether the participants preferred the level

with the stressful ambience or the level with the calm ambience. This was done by asking the

participants a series of questions. The participants' answers for these questions were analyzed by

coding their answers and extracting keywords. The results showed no significant difference in

player preference.

Acknowledgments

I would like to thank my supervisor, Jon Allan, for his valuable support when writing this paper.

I would also like to thank my classmates for their help, support and valuable input in the

pre-study and with writing the paper.

Table of contents

1. Introduction 4

2. Background 4

2.1. Audio-based games 4

2.2. Ambience in virtual games 5

2.3. Stressful Sounds 5

2.4. Calming Sounds 6

2.5. The use of auditory icons and earcons 7

2.6. Previous Work - Stress affecting gameplay 7

2.7. Aims and purpose 8

3. Method 10

3.1. Experiment 10

3.1.1. Building The Game 10

3.1.3. Pilot Study 13

3.1.4. Stimuli 14

3.1.5. Subjects 16

3.1.6. Listening test 17

3.1.7. Conditions 18

3.2. Data Collection 18

3.3. Data Analysis 19

4. Results and Analysis 20

4.1 Quantitative data 20

4.2 Qualitative data 23

5. Discussion 26

5.1 Quantitative data 26

5.2 Qualitative data 27

5.3 Critique of method 29

5.4 Conclusion 30

5.5 Future Research 31

References 32

Appendix A 35

Appendix B 37

1. Introduction

Previous research shows that stress can affect the way we behave, the way we act in a virtual game (Steinrücke, et al., 2019). This study will take a look at the hypothesis that gameplay differs in a virtual game when the player is feeling stressed, compared to when they are feeling calm. The definition of ‘Gameplay’ in this study refers to how the player is playing the game, her or his movements, decision making and choices of these movements. Is there a correlation between how fast the player finishes a game-task and the emotional state of the player themself?

The study will also take a look at how participants experience the game, being stressful or calm, will there be an overall preference for either of the two levels? If so, why is there a preference and does it correlate to the level being either stressful or calm?

2. Background

2.1. Audio-based games

The research is centered around sound design in audio-only games. Audio-based and audio-only games are essentially games with limited or no amount of visuals. According to Rovithis (2012), audio-based games are almost as old as video games; what makes them different from video games is the much more important role of sound in the gameplay, where the sense of hearing, the ability to listen, perceive and interpret sound is of highest importance. Moreover, audio-based games can fall into four different kinds; rhythm games, instrument games, art games and musical puzzles. Some examples of published, audio-based games such as ‘Hunter’, where players target and fire objects, the horror game ‘Kowa-Oto’ and ‘Papa-Sangre’ which combines puzzle

elements and interactive narration to create a sonically virtual world. Research on audio-based

games has led to classifying the different genres in another way; According to Rothvitis (2012)

an Audio-based puzzle game could be classified by how they use sound properties to create riddles for the player to solve, by exercising the player’s memory, recognition and perception.

Rothvitis (2012) also mentions “3D Positional Audio” as a type of audio-only genre where an object is placed somewhere on the horizontal plane, while the distance from the object to the player is represented by varying amplitude.

2.2. Ambience in virtual games

Lee (n.d.) writes that ambient sound can be defined as a background sound, an atmosphere present in a scene or location. The term ambient sound is more focused on the more physical and functional part, whilst the term ambience is more focused on the psychological and emotional part. Ambience and ambient sound can normally be seen in digital media productions, such as films and games to construct a specific atmosphere.

According to Lee (n.d.), the ability of the ambience in a virtual game to be either diegetic or non-diegetic. Meaning that diegetic sounds refer to the game world which can be heard by the game character, while non-diegetic means the opposite.

2.3. Stressful Sounds

McDermott (2012) writes about certain sounds that can evoke hedonic and aversive responses in humans. There are several factors that come into play when it comes to sounds that are seen as

“unpleasant”. Such aspects can be acoustic properties, learned associations between sounds and emotional situations, context, input from other senses, and the listeners personality and mood.

McDermott (2012) gives the example of two properties that have substantial influence;

“sharpness” and “roughness”. High frequencies in the range of 2-4kHz contributes to the most

annoyingness. The annoying effect might be rooted in how we perceive certain frequencies.

Since the ear perceives these frequencies by as much as 30dB higher than other frequency ranges, these frequencies are most likely to damage the ear, which could be an explanation to why we have an aversive response to them specifically. Associations between sounds and the events that cause them clearly plays a role in whether they are perceived as unpleasant or not, (McDermott 2012). A few examples of such associations will be discussed below.

According to Parsons, et al. (2012), there are few sounds that carry as much biological clout as the cries of a distressed infant. They talk about how a distressed infant's cry elicits autonomic arousal in the listener, measured by heart rate, blood pressure and skin conductance or hand grip force. This might reflect a ‘high-alert’ state which prepares an adult to react rapidly to the infants crying (Parsons, et al., 2012).

2.4. Calming Sounds

According to McDermott (2012), emotional associations with relaxing circumstances plays a role in how we perceive environmental sounds as pleasant. Sounds of ocean, wind and rain usually have more energy at low frequencies and feature slow temporal modulation (McDermott, 2012).

A study conducted by Saadatmand, et al. (2013) shows to identify the effect of nature-based sounds on agitation, anxiety level and physiological stress responses in patients under

mechanical ventilation support. The results showed that the experimental group had significantly lower blood pressure, anxiety and agitation levels than the control group.

Ratcliffe, et al. (2013) writes about how natural environments usually are perceived as restorative

following stress and attention fatigue. They talk about semi-structured interviews with twenty

adult participants, where bird songs and calls were found to be the type of natural sound most commonly associated with stress recovery and attention restoration. Although, it is little known about restorative perceptions of bird sound on their own and how these may relate to existing theories of environmental restoration (Ratcliffe, et al., 2013).

2.5. The use of auditory icons and earcons

Dingler, et al. (2008) explain auditory icons as being brief sounds that represent objects, functions and actions by taking advantage of the users prior knowledge and natural auditory associations with sound sources and causes. As long as a sound evokes the associated sound of an object or action, it is classified as an auditory icon. Earcons, on the other hand, are described as abstract, synthetic and mostly musical tones or sound patterns, non-verbal audio messages.

Moreover, earcons usually consist of five parameters to vary; rhythm, pitch, timbre, register and dynamics (Dingler, et al., 2008).

2.6. Previous Work - Stress affecting gameplay

Previous work by Steinrücke, et al. (2019) shows that stress influences a decision-makers

behaviour and also on the learning experience in training scenarios. The ability to assess the level

of stress experienced would allow manipulation of a game to improve the learning experience. A

study was conducted with two conditions, one paced and one non-paced. In the paced condition

the test subjects were exposed to in-game changes that aimed to induce stress by uncertainty,

time pressure and information overload. The study showed differences between the conditions

with respect to in-game performance. Participants in the non-paced condition seemed to follow a

more intuition-guided, not thought-through, decision making strategy, which in turn was less

accurate (Steinrücke, et al., 2019).

Another study conducted by Parsons, et al. (2012) shows that reported physiological change in response to infant cries translates into measurable difference in adults ability to move in a concerted and accurate manner. In the research, they played a series of infant cries and other sounds to a group of adults and measured their performance when playing a miniature version of the classic arcade game “Whack-a-mole”. The game required participants to press down on a target button, with a predetermined amount of force, in order to score points. Scores of the game were significantly higher after listening to the infant cries compared to the other sounds, with similar scores for men and women. The minimum pressure that the participants applied to the buttons were significantly greater after listening to the sound of the infant crying. This might suggest that infant cries can motivate people to act. The participants demonstrated a positive consequence of hearing the crying infant, improvements in speed and accuracy and in intentional movements, which may reflect on an adaptive physiological response when hearing the sound (Parsons, et al., 2012).

Yu. (2016) writes about previous studies that have shown that when stressed, individuals tend to make more habitual responses than goal-directed choices, being less likely to adjust their initial judgement and to rely more on gut feelings in social situations. According to Yu. (2016), it might be possible that stress influences the arbitration between the emotion responses in certain parts of the brain, so that final decisions are based on unexamined innate responses.

2.7. Aims and purpose

The intended purpose of this study is to look further into the hypothesis that induced stress

affects how you play a virtual game. This paper will mainly focus on non-diegetic ambience,

since the intention is to look at how the ambience affects the player emotionally, without playing a part of the story of the game. The choice of using an audio-only, puzzle-game for this

experiment was made solely due to interests in the genre. By investigating if and how stress will affect gameplay, there could be a potential of using this knowledge when designing an

audio-only puzzle game, but the knowledge could also be applied to other virtual games with visuals.

One aim of this study is to look at if the gameplay will differ in an audio-only game when comparing stressful versus calm ambience. This will be done by timing how long it takes for each participant to finish two separate levels in a game, one with stressful ambience and one with calm ambience. The game that was made for this study puts the player's use of perception on test, by placing obstacles in the game that the player has to walk around to be able to get to an

objective that makes a sound, somewhere on the map. The hypothesis is that it might go either way. Either the participants will finish the stressful level faster, due to improvements in speed and accuracy or the participants will finish the stressful level slower, due to a lack in decision making and situational awareness. The other aim of this study is to see how participants

experience the gameplay, being stressful or calm, will there be an overall preference for either of

the two levels? If so, why is there a preference and does it correspond to the level being either

stressful or calm?

3. Method

3.1. Experiment

3.1.1. Building The Game

An audio-only, puzzle game was made for the experiment. The game was built using Unreal Engine 4 (2014) and the sound was implemented using Blueprints visual scripting system in Unreal Engine 4 (2014). The game was chosen to be made like a first-person shooter (FPS) game, using the industry-standard controls “W,A,S,D” to move forward, backward, left and right.

The mouse was used for the player to look around in the game. The choice was made to make the controls as simple as possible, therefore eliminating the ability to jump or sprint in the game, giving every participant the same walking speed.

The goal of each level is to move towards a sound-object and to collect it. The choice was made to put obstacles spread across the map of each level, to give a little bit of challenge to the

gameplay, as seen in Figure 1. When the player has collected the sound-object, it moves to a new

location and the player has to find it again. When the player finds the sound-object the 3rd time,

the level is completed. Three levels were made for the listening test. Two of the levels, X and Y,

were identical in their design, the sound object appearing in the same places for each level and

the obstacles standing in the same spots, their only difference being their ambience sound. There

was also a training level called ‘Trial level’ made for the players to get more comfortable with

the gameplay before playing Level X and Y. The goal for the trial level was to make the player

comfortable with the controls of the game and to give them an idea of how it was played,

therefore it was ideal to use a trial level without any ambience, where only the sound object and

auditory icons and earcons were being displayed (the auditory icons and earcons containing sounds like footsteps and the sound of colliding with an object). After the levels were built, they were blacked out, leaving only a black screen for the player when they enter each level, making it sound-only. A menu was implemented when opening up the game, giving the players an option to either play Level X, Level Y or the Trial Level, as seen in Figure 2. A timer was implemented in each level, showing the played time in seconds after finishing each level, while also reminding the participants to write down the number, as shown in Figure 3.

Figure 1, an overview of the map of Level X and Y in the editor of Unreal Engine 4 (2014). The picture was edited in Gimp-2.10 (2021). The x represents the starting point for the player. The 1, 2, and 3 represents where the sound-object spawns. The grey boxes around the map are the obstacles.

Figure 2, an overview of the startup-menu. Made in Unreal Engine 4 (2014).

Figure 3, an overview of the timer-popup from the players perspective. Made in Unreal Engine 4 (2014).

3.1.2. Sound parameters in Unreal Engine 4

It was important to think about the settings of sound parameters in Unreal Engine 4 (2014) since the levels excluded anything visual. The sound-object being the most important one since it needed to be located fairly easily. For this it was chosen to give the sound-object a “Round”

attenuation-setting, a setting that reached across the whole map in a circular shape, making the player able to hear it anywhere on the map. The attenuation function was set to “Natural sound”, which resembles a more realistic falloff-attenuation, making the changes in volume greater at close distance and just audible at far distances. This to make it easier for the player to locate how far away the sound-object was in relation to themselves. “Spatialization” was enabled on the sound-object, boosting the sound-objects dB in either the left or right stereo channel when moving focus, while simultaneously lowering it in the other channel, making it possible for the player to hear if the object was to the left or to the right. “Attenuation listener focus” was enabled to boost the sound-objects dB when the sound-object was in range of listener focus, lowering it’s dB when turning ‘your back’ to it, making it possible to hear whether the object was in front or behind you. “Sound occlusion” was left out since there were a lot of obstacles in the maps, which would have made it too hard to locate the sound-object. The sound-object was the loudest sound in all of the levels, to make it stand out and easy to locate. The ambience was set to play in 2D but more quiet than the sound-object, the two different ambiences played at the same dB-level at -14LUFS. The auditory icons also played in 2D, being the less loud sounds, only to give the player indications on what was going on in the game.

3.1.3. Pilot Study

A pilot study was made before choosing the stimuli for the listening test. The intention for Level

X was to use a stimuli as the ambience that made the participants feel calm. The intention for

Level Y was to use a stimuli that made the participants feel stressed. For this, three people participated in the pre study where they got to listen to the stimuli chosen for the ambience in Level X and the stimuli chosen for the ambience in Level Y. The participants then had to write shortly about how the stimuli made them feel. The chosen stimuli worked as intended and the participants wrote that they felt stressed by the stimuli chosen for Level Y and calm by the stimuli chosen for Level X.

The pilot study included playtesting. The packaged game was sent via Google Drive to four different participants for playtesting. The feedback was useful since it turned out that there was a hole in the floor of every level, making the player get stuck for no reason, which later was attended to. There were also some complaints about the sound from the sound-object being too quiet, this too was attended to. After attending to it’s faults, the game was sent back to the participants and approved by the participants before the listening test.

In this pilot study, the participants had to read the instructions given before playing, and answer the questionnaire after they finished playing the three levels. This proved to be useful since the feedback showed that the instructions lacked in clarity. One example was from a participant who accidentally put their headphones on the wrong way, leaving the left headphone on the right ear and vice-versa. This made it practically impossible to play the game, the importance of

headphone placement was later added to the instructions.

3.1.4. Stimuli

The choice of stimuli for the ambience of Level X was a sample of birdsong and a sample of

calm rain, which was edited and mixed. The choice of stimuli for Level Y was a sample of a 10

months old baby crying and screaming and a sample of multiple clocks ticking fast, which, too, was edited and mixed. The samples were found in a soundbank and edited using Logic Pro X (2013). The samples were edited in such a way to make it loop seamlessly throughout gameplay.

The intention for both stimulus was to have it use similar frequency ranges for both Level X and Level Y. Therefore a High Pass Filter was added to both the stimulus, as seen in Figure 4. This was also done to minimize masking of the sound-object, which included some more lower frequencies. The stimuli used for ambience in Level X and Level Y were normalized using the plugin Izotope Ozone 9 Element to -14LUFS each to give them a similar loudness level.

Figure 4, A screenshot of the high pass-filter used for the atmosphere stimulus. Made using Logic Pro X’s (2013) Channel EQ-plugin.

The stimuli for the sound-object was a recording made by a professional voice actor in his own

home studio, and sent over via Google Drive. The instructors given to the voice actor was to

record several takes of him calling the player to come over, including sentences in swedish

saying “Here I am, come on over here..” and so on. The actor was also given the instructions to

sound as neutral as possible while calling the player over, not to sound too stressed or distressed.

All of the recordings were used for the stimuli, each sentence played 2,5 seconds after one another, repeatedly, throughout the gameplay.

There were three types of auditory icons used in the game to give the player a clearer indication of what was going on in the gameplay. The first one being the sound of footsteps from the player when walking, this to indicate that the player is actually moving. The second one being a

collision sound when colliding with the obstacles spread across the map, this to indicate that the player collides with them. The collision sound consists of two samples merged together, one being the sound of a person grunting and the second one being the sound of impact with stone.

The third auditory icon used was a confirmation sound when the player collects the sound-object, this to indicate that it is actually found. This sound consisted of two, high pitched, notes played in an octave in the key of D, on a sampled African Kalimba using the EXS24-plugin in Logic Pro X (2013).

3.1.5. Subjects

The participants for the listening test were people who were acquainted with first person games since earlier, where the controls used were similar to the ones that were used in this listening test.

There was a conscious decision to not look for people who define themself with the term

“gamer”, since it was considered too much of an open, and unclear, label. The participants were

picked out at random with the one criteria that they had been playing first-person shooter games

before. There was no criteria that the participants had to be experienced listeners. There were 14

participants in total. They were Swedish citizens, between the age of 19-38 years old, 12 men

and 2 women.

3.1.6. Listening test

The listening test was made online. The finished game was packaged and uploaded on Google Drive along with a link to instructions and questionnaire. This was sent to the participant. The instructions and questionnaire was made in Google Forms. There were two different

questionnaires made, being identical but one telling the participant to begin with the Level X

after the Trial Level (Marked “Form X”), and the other one telling the participant to start with

Level Y after the Trial Level (Marked “Form Y”). The researcher sent “Form X” to the first

participant, then “Form Y” to the next participant, alternating the order every other time, while

keeping track of the participants in a list. This was to create a random aspect on which level the

participants get to play first, lowering the risk of getting faulty data. By randomizing the order, it

lowered the risk that the data was a result of being accustomed with the levels and in how to play

the game. The instructions told the participants in what order to play each level and to write

down the number appearing on the screen after each level. The number in question was the time

it took for the player to finish the level, written in seconds. This was not mentioned in the

instructions, to not give away the intention behind the research. The instructions told the

participants the controls of the game and to play the game in a calm setting where they could

focus and play each level uninterrupted. The instructions also told the participants how to play

the game, explaining that the participants would hear one type of noise when colliding with the

obstacles, and another noise that confirms that you have “collected” the person shouting at you,

i.e. the sound-object. The researcher was available on chat throughout the listening test to help

with any questions. When the participants were done playing the three levels, they proceeded

with answering the questionnaire.

3.1.7. Conditions

The choice of stimuli was an experimental factor. There were two levels of factors when it came to stimuli, version X and version Y. The two different versions of the game, the two different stimuli were independent variables. A dependent variable was how each participant rated the different questions and the time it took for the participant to finish each version of the game. One of the conditions of the listening test was that it was made online, which made it harder to

control. The instructions stated that the participants needed to use headphones for the test,

although the quality of the headphones was not regulated. The participants also needed Windows 64-bit, a keyboard and a mouse for the listening test. The instructions encouraged them to sit down in a calm environment when doing the listening test and to play the levels without interruption, although this was also hard to regulate completely.

3.2. Data Collection

Both qualitative and quantitative data was collected in the listening test. Qualitative data was collected from the questionnaire, answered by the participants. The survey included an AB 5-scale,where the participants graded either version X or Y to be higher rated, or no difference.

This meant that it was important for the participants to keep track on which version was which.

A simple solution for this was to be clear on which level the participant played first. The

participants were also able to motivate their answer with their own words, after each question,

maximum two sentences. The questions looked as shown in Table 1.

Table 1

A representation of the question asked in the survey.

1 = Level X

2 = Somewhat Level X 3 = No difference 4 = Somewhat Level Y 5 = Level Y

Did you find any of the levels (X and Y) to be more hard than the other?

Motivate your answer with a maximum of two sentences.

Ans: 1,2,3,4 or 5.

Did you find any of the levels (X and Y) to be more stressful than the other?

Motivate your answer with a maximum of two sentences.

Ans: 1,2,3,4 or 5.

Did you find any of the levels (X and Y) to be more fun to play than the other? Motivate your answer with a maximum of two sentences.

Ans: 1,2,3,4 or 5.

Is there anything else that you want to say about the experience and/or differences of Level X and Level Y?

To the left of the table are the questions that the participants got asked in the survey represented. To the right are the answers that the participants got to choose from represented.

Quantitative data was collected by timing how long it took for the player to finish each level, from start to finish. This was done by implementing a timer in the game and letting the

participants write down the time, in seconds, on a piece of paper, to then write it down again in the questionnaire.

3.3. Data Analysis

The quantitative data has been analyzed by coding the answers of the AB 5 scale to numbers,

such as -2, -1, 0, 1, 2. The average from all the answers, for each question has been calculated, to

see if they differ from the hypothesis = 0. A one sample t-test has been conducted where the

reference value is = 0. The number of total participants was 14. The alpha level for this analysis

was 0.05. The average time it took for the participants to finish Level X and Level Y respectively

have been collected and a paired sample t-test has been conducted to see if there is a significant difference in the average in time for each level.

The qualitative data has been analyzed according to Merriam, S. B. (2014), in a table. There is one table per question. Every participant is marked with a number 1-14. Every participant's answer has been coded to a + or a - on either Level X or Level Y, or both. What the + and - represents for that specific question is written above the table itself under section 3.2. Every answer has been sorted into different themes, depending on what kind of subject the answer touched on. Every answer got one keyword of importance, written in english with the original answer in swedish next to it, in brackets.

4. Results and Analysis

4.1 Quantitative data

The participants were asked which one of the two levels (X and Y) were the hardest one (M = -0.143, SD = 1.350). The answers showed no significant differences in opinion, t(13) = -0.396, p

= .699. A column chart has been made to display the distributed answers, see Figure 5.

Figure 5, A column chart showing the distributed answers for the stated question.

The participants were asked which one of the two levels (X and Y) were the most stressful one

(M = -1.5, SD = 0.650). The answers showed a significant difference where Level Y appeared

more stressful, t(13) = -8.629, p = .00. A column chart has been made to display the distributed

answers, see Figure 6.

Figure 6, A column chart showing the distributed answers for the stated question.

The participants were asked which one of the two levels (X and Y) were the most fun to play (M

= 0.214, SD = 1.477). The answers showed no significant differences of opinion, t(13) = 0.543, p

= .596. A column chart has been made to display the distributed answers, see Figure 7.

Figure 7, A column chart showing the distributed answers for the stated question.

The time it took for each of the 14 participants to finish level X (M = 159.3, SD = 122.5) compared to the time it took for each of the 14 participants to finish Level Y (M = 118.9, SD = 69.9), shows a significant difference in Level X taking longer to finish, t(13) = -2.299, p = .003.

4.2 Qualitative data

Table 2

Question: “Did you find any of the levels (X and Y) to be harder than the other?”

+ = Harder - = Not harder

Participant Level X Level Y Theme Keyword

1 + Sound React (Reagera)

2 + Sound Hard (Jobbigt)

3 + Sound Stressful (Stressande)

4 + Sound Distracting (Störande)

5 + Sound Distracting (Distraherande)

6 + - Sound Motivating (Motiverande)

7 + Game mechanics Wall (Vägg)

8 + Game mechanics Harder (Svårare)

9 + + Game mechanics Alike (Lika)

10 + Sound Hear (Höra)

11 + Sound Distracting (Distraherande)

12 + - Game mechanics Walls (Väggar)

13 + + Game mechanics Objects (Objekten)

14 - - Sound None (Ingen)

A table showing the results of qualitative data. The keywords are translated to english with the original language, Swedish, written in brackets. Made in Google docs.

Table 3

Question: “Did you find any of the levels (X and Y) to be more stressful than the other?”

+ = Stressful - = Not stressful

Participant Level X Level Y Theme Keyword

1 + Sound The baby (Bäbisen)

2 + Sound The clocks (Klockorna)

3 + Sound Screaming child (Barnaskrik)

4 + Sound Timer (Räknaren)

5 + Sound Screaming baby (Bäbisskriket)

6 + Sound Ticking clocks (Klockticket)

7 + Sound Screaming baby (Barnskriket)

8 + Sound Child (Barn)

9 + Sound Clocks (Klockorna)

10 + Sound Ears (Öronen)

11 + Sound Pulse (Puls)

12 + Sound Stressful (Stressande)

13 + Sound Clock (Klocka)

14 + Sound Baby (Bäbis)

A table showing the results of qualitative data. The keywords are translated to english with the original language, Swedish, written in brackets. Made in Google docs.

Table 4

Question: “Did you find any of the levels (X and Y) to be more fun to play than the other?”

+ = More fun - = Not more fun

Participant Level X Level Y Theme Keyword

1 + Sound Peaceful (Fridfull)

2 + + Genre Fun (Roligt)

3 + - Sound Screaming (Skrikljudet)

4 + - Sound Pleasant (Behagligare)

5 + - Sound Babies (Bäbisar)

6 + - Sound Enviroment (Miljön)

7 + Game mechanics Easier (Lättare)

8 + Feeling Exciting (Spännande)

9 + Feeling Funnier (Roligare)

10 + - Feeling Calm (Lugnt)

11 + + Feeling Motivating (Motiverande)

12 - + Game mechanics Easier (Lättare)

13 + Feeling Perform (Prestera)

14 - - Feeling None (Ingen)

A table showing the results of qualitative data. The keywords are translated to english with the original language, Swedish, written in brackets. Made in Google docs.

5. Discussion

5.1 Quantitative data

As one can see under section 3.1, there was a significant difference in the average time it took for each of the participants to finish Level X and Level Y respectively. Level Y shows an average of 118.9 seconds while Level X shows an average of 159.3 seconds. This means that Level X took more time to finish. This, in turn, could correlate to what Parsons, et al. (2012) writes about previous experiments where the participants demonstrated a positive consequence of hearing a crying infant, showing improvements in speed and accuracy and in intentional movements. If the previous studies correlate with the result of time difference in this study, it could mean that the participants in this study got motivated to finish the level faster due to the sound of a crying baby. This, in turn, is interesting since Table 3 clearly shows a correlation between keywords like

“screaming child” and “clocks” to the feeling of stress in Level Y. This could mean that the participants showed an increasing feeling of stress correlated with the sound of a screaming baby, while still finishing Level Y faster than Level X. This correlation is backed up with the significant difference that Level Y appeared more stressful in the analysis under section 3.1. This could back up the hypothesis that stress affects the way that a virtual game is played.

The participants were asked which one of the two levels were the hardest one, the answers showed no significant difference in opinion. The participants were also asked if any of the two levels were more fun than the other, the answers showed no significant difference in opinion.

This makes it hard to draw a conclusion between any of the levels stimuli, the level of stress, and

the player experience.

5.2 Qualitative data

As one can see in Table 2, under section 4.2, when asked which level appeared to be harder, the participants motivated their replies with keywords that were coded two fall into two different themes in particular. The themes in question are ‘Sound’ and ‘Game mechanics’. The chosen keywords that fell under theme ‘Sound’ were “react”, “hard”, “stressful”, “distracting”,

“motivating”, “hear” and “none”. “Distracting” being a keyword that appeared three times from different participants. An example from one of these answers containing the word “distracting”

was from participant 5 saying: “The baby screams were irregular and more distracting”. Another example is from participant 11 saying: “Level Y contained more transients in the sound, which was distracting”. Both of these answers can be found highlighted under Appendix B. This could give an indication that the aspect of distraction played a bigger role in how hard the levels were perceived by the players. Regarding the keyword “hear”, this keyword could fall under the category of masking, since the full sentence from this participant regarding Level Y was: “It was harder to orient oneself and to hear the direction of the man's voice.”.

The chosen keywords that fell under theme ‘Game mechanics’, seen in Table 2, were “walls”,

“harder”, “alike”, “objects”. Walls being a keyword that appeared two times from different participants. The answers showed that more than one participants regarded Level Y as having a harder map, with more walls and obstacles, which is interesting since Level X and Level Y were identical in their design. The answers can be seen under Appendix B.

As one can see in Table 3, under section 4.2, when asked if any of the levels appeared to be more

stressful, the participants motivated their replies with keywords that were coded to fall into the

theme ‘Sound’. The chosen keywords that fell under this theme were “the baby”, “the clocks”,

“screaming child”, “timer”, “ears”, “pulse” and “stressful”. “baby” being a keyword that appeared 6 times from different participants and “clocks” appearing 4 times. This shows an indication of how the choice of stimuli for Level Y was one of the main reasons for why the participants rated Level Y to be a more stressful level.

As one can see in Table 4, under section 4.2, when asked if any of the levels were more fun to play than the other, the participants motivated their replies with keywords that were coded two fall into four themes in particular; ‘Sound’, ‘Game mechanics’, ‘Feelings’ and ‘Genre’. The chosen keywords that fell under the category of ‘Sound’ were “peaceful”, “screaming”,

“pleasant”, “babies” and “environment”. The answers show that there was a general disliking to the stimuli used for Level Y, and a general liking to the stimuli used for Level X. This can be backed by the previously mentioned research by McDermott (2012), which talks about how environmental sounds generally are perceived as pleasant sounds and how aspects such as

‘sharpness’, ‘roughness’ and association can add to how a sound is perceived as ‘unpleasant’.

Some of the participants' answers show that the pleasantness of the stimuli in Level Y added to a

fun experience. For example, participant 6’s motivation regarding Level Y stated “I rather

wanted to stay there, it was more fun to look for the man in that environment.”. Although, there

were also a few answers regarding the stimuli used for level X that stated how it added to a fun

experience, such as participant 9’s motivation: “Even if the experience of Y is more stressful, it

makes it more fun to perform when being a bit tense.”.

Chosen keywords for the theme ‘Game mechanics’, seen in Table 4, were “easier”, mentioned 2 times. These answers stated that Level Y had an easier map, although, as mentioned before, both levels had two identical maps. Interestingly enough it was the same participants stating that Level Y had an easier map in Table 3, that also stated that Level X had an easier map in Table 2.

Chosen keywords for the theme ‘Feeling’ were “exciting”, “funnier”, “calm”, “motivating”,

“perform” and “none”. These answers show a correlation between how the stress was

motivating, exciting and/or that the calmer level, level X, were more fun because of less stress.

The participants' opinions were divided. All the original answers can be read under Appendix B.

5.3 Critique of method

There were a couple of incidents in the listening tests where participants chose to play the levels in the wrong intended order. It was stated in the instructions, as seen under Appendix A, that the order that was written in the instructions were important since it alternated for each and every participant. Even still, some participants chose to play the levels in the order of Trial level first, then Level X and then Level Y, regarding what was said in the instructions. One solution to this could have been to change the name of the levels to something different, and not to two

consecutive letters in the alphabet.

Another thing that should've been taken into consideration when choosing the stimuli for the

sound object, was the importance of masking, especially in a sound-only game. The choice of a

voice calling as the sound object and the choice of a baby screaming as the atmosphere, could

have affected the player experience since the two stimuli move in a similar frequency area of

speech.

5.4 Conclusion

This study has investigated if gameplay in an audio-only puzzle game differs when the player is feeling stressed or calm. The players were timed during the playtest and the results were

analyzed. The results showed a significant decrease of played time in the level containing stressful ambience. The players were asked which of the levels appeared to be more stressful.

The results showed a significant inclination towards Level Y, the level with the stressful ambience. The study has also investigated whether the participants preferred the level with the stressful ambience or the level with the calm ambience. The participants' answers for these questions were analyzed by extracting keywords. The results showed no significant difference in player preference.

The results imply that there is a correlation between the emotional state of the players and how fast the players finished the task of the game. Looking at previously mentioned research by Parsons, et al. (2012), this might be due to a positive consequence of hearing the crying infant, giving the player improvements in speed and accuracy and in intentional movements. This could in turn mean that the results are useful when sound designing for an audio-only puzzle game, knowing that the choice of ambience might affect the gameplay itself. Due to previous research by Parsons, et al. (2012), one can also conclude that the same goes for virtual games with visuals. The ability of manipulating the player in a virtual game via sound might be a powerful way of finding different approaches when designing the game itself.

The results show no significant preference of any of the levels. Respectively, there was a

significant inclination towards the level with the stressful ambience appearing to be more

stressful. This might imply that there is no correlation between player preference and the emotional state of the player. Contrary to studies by McDermott (2012), that implies that the choice of stimuli could be regarded as “annoying”, there was not a clear indication on this being the case. Moreover, there were several answers in the questionnaire saying that the stressful ambience motivated them and made the game more fun to play. This knowledge might, too, be useful when applied to sound design, knowing that a stressful ambience does not necessarily need to correlate with the preference of the player.

5.5 Future Research

Future research could take a closer look at how the player is moving on the map when

conducting the listening test. In this way, there might be patterns of movement that can be taken

notice of, leading up to other conclusions regarding gameplay. Another thing to look further into

is the aspect of distraction. “Distraction” was several times mentioned amongst the keywords and

might be a possible factor of why the players got affected by the stressful ambience. This could

be researched by asking more questions regarding which level appeared more distracting.

References

Dingler, T., Lindsay, J., & Walker, B. N. (2008). Learnabiltiy of sound cues for environmental features: Auditory icons, earcons, spearcons, and speech. International Community for Auditory Display.

Kimball, S., Mattis, P. (2021). Gimp (Version 2.10) [Computer Software]. GIMP Team.

https://www.gimp.org/

Lee, M. T. (n.d). The Contribution of Ambient Sounds in Open World Video Games. [MA Music Technology Dissertation]. Academia.edu.

https://www.academia.edu/download/61250556/The_Contribution_of_Ambient_Sounds_in_Ope n_World_Games20191118-60892-1vjmrnt.pdf

Apple Inc. (2013). Logic Pro X (Version 10.0.0). [Computer Software]. Mac App Store.

https://www.apple.com/logic-pro/

McDermott, J. H. (2012). Auditory preferences and aesthetics: Music, voices, and everyday sounds. In R. Dolan, T. Sharot (Ed.), Neuroscience of preference and choice (pp. 227-256).

Academic Press. https://doi.org/10.1016/B978-0-12-381431-9.00020-6

Merriam, S. B. (2014). Qualitative research: A guide to design and implementation. San

Franscisco: Jossey-Bass. eLibrary https://libris-kb-se.proxy.lib.ltu.se/bib/16879902

Parsons, C. E., Young, K. S., Parsons, E., Stein, A., & Kringelbach, M. L. (2012). Listening to infant distress vocalizations enhances effortful motor performance. Acta paediatrica, 101(4), e189. https://doi.org/10.1111/j.1651-2227.2011.02554.x

Rovithis, E. (2012). A classification of audio-based games in terms of sonic gameplay and the introduction of the audio-role-playing-game: Kronos. In Proceedings of the 7th Audio Mostly Conference: A Conference on Interaction with Sound (pp. 160-164).

Ratcliffe, E., Gatersleben, B., & Sowden, P. T. (2013). Bird sounds and their contributions to perceived attention restoration and stress recovery. Journal of Environmental Psychology, 36, 221-228. https://doi.org/10.1016/j.jenvp.2013.08.004

Saadatmand, V., Rejeh, N., Heravi-Karimooi, M., Tadrisi, S. D., Zayeri, F., Vaismoradi, M., &

Jasper, M. (2013). Effect of nature-based sounds’ intervention on agitation, anxiety, and stress in patients under mechanical ventilator support: A randomised controlled trial. International Journal of Nursing Studies, 50(7), 895-904. https://doi.org/10.1016/j.ijnurstu.2012.11.018

Steinrücke, J., Veldkamp, B. P., & De Jong, T. (2019). Determining the effect of stress on analytical skills performance in digital decision games towards an unobtrusive measure of experienced stress in gameplay scenarios. Computers in human behavior, 99, 144-155.

https://doi.org/10.1016/j.chb.2019.05.014

Sweeney, T. (2012). Unreal Engine 4 (version 4.22). [Game Engine]. Epic Games.

https://www.unrealengine.com/en-US/

Yu, R. (2016). Stress potentiates decision biases: A stress induced deliberation-to-intuition

(SIDI) model. Neurobiology of stress, 3, 83-95. https://doi.org/10.1016/j.ynstr.2015.12.006

Appendix A

Instructions used for the listening test, in original language (Swedish)

“Du kommer att få spela tre banor i detta lyssningstest.

Spelet är för Windows 64-bit och du behöver hörlurar för att göra testet. Du behöver även t.ex papper och penna för anteckningar. Du kommer att få svara på några korta frågor efter att du har spelat. Allt som allt tar testet ca 15-20 minuter.

Spelet är ett ljudbaserat spel, alltså kommer det inte vara något visuellt i själva banorna. Spelet använder sig av typiska WASD-kontroller för att gå framåt, bakåt, höger och vänster. Du använder musen för att titta dig omkring. Viktigt är att ha höger hörlur på höger öra och vänster hörlur på vänster öra för att kunna lokalisera vart ljudkällan befinner sig i spelet.

Målet för respektive bana är att gå emot rösten som ropar på dig, när du har hittat mannen som ropar så hörs ett godkännande ljud, detta gör du tre gånger för varje bana. I varje bana finns hinder, objekt som står i vägen, du kommer att höra ifall att du går in i dom, gå då bara runt objekten.

Efter varje bana ska du skriva ner siffran som kommer upp på skärmen. Du ska spela varje bana endast en gång, helst utan avbrott, så det är bra om du kan sitta i lugn och ro.

Spelordningen för banorna ser ut som följer:

1. Trial 2. Level Y 3. Level X

"Trial" är endast en testbana för att du ska bekanta dig med kontrollerna. Viktigt är att du spelar just ovanstående ordning, då spelordningen alterneras för varje testperson.

Efter att du har spelat klart de tre banorna och skrivit ner dess respektive siffra, så kan du klicka

dig vidare i formuläret för att svara på frågorna.”

Appendix B

Full answers from the survey, in original language (Swedish)

Question: “Did you find any of the levels (X and Y) to be more hard than the other?”

Participant Motivation

1 Alla ljuden gjorde det svårare för mig att märka och reagera på att jag gått in i ett hinder.

2 Jobbigt med ljudet av klockor och bebis. Sedan ett hinder precis innan man kom fram till rösten

3 Jag upplevde det som att Y hade fler väggar, men kan också varit stressande skrik som gav den känslan.

4 Lite svårare pga fler störande ljud, men rösten är så tydlig så funkade bra ändå.

5 The baby screams were irregular and more distracting

(

6 Jag blev nog mer "förförd" av fågelkvittret, tog längre tid på mig i världen, och mer motiverad av klocktickandet och barnskriken

7 Jag fastnade ett par gånger mot vad jag tror var en vägg, så jag kunde inte hitta den andra personen.

8 Upplevdes som en svårare labyrint

9 Jag upplevde dem ungefär lika svåra, dock lättare att navigera runt hindren ju längre jag spelat.

10 It was harder to orient one self and to hear the direction of the mans voice. (det var svårare att orentera sig och höra riktningen till mannens röst på bana Y)

11 Level Y contained more transients in the sound, which was distracting (Bana y hade mer transienter i ljudet vilket distraherade)

12 X Kändes mer av en labyrint med fler väggar och var mer kringelkrokig. Y var stressade men banan kändes lättare.

13 Glömde att man kunde gå runt objekten som stod i vägen. Det var svårt att veta hur jag skulle ta mig runt dem.

14 Bakgrunds ljudet spelade ingen större roll

Question: “Did you find any of the levels (X and Y) to be more stressful than the other?”

Participant Motivation

1 Jobbiga ljud. Speciellt den gråtande bäbisen.

2 Kändes som att klockorna och bebisens ljud blev snabbare och högre ju närmre jag kom rösten 3 Barnaskrik är lite mer stressande än regn.

4 Räknaren fick det att kännas som det gick på tid och bebisskrik är alltid lite stressande, men gick att zoona ut pga kort bana.

5 bäbisskriket är sjukt stressande och i kombo med tickandet känns det som om nåt kommer att hända om jag inte skyndar mig

6 Klockticket och barnskriken gjorde att det upplevdes mer "stressigt"

7 Det var lite mer stressigt med barnskriken i level Y men jag blev mer stressad av att fastna i väggen på level x.

8 Barn skrik är stressande och läskig ton i backgrunden också.

9 Klockorna och barnskriket var väldigt stressande

10 Blev mycket mer stressad på Y och behövde nästan ta en paus för att vila öronen 11 Vi får automatiskt puls av barnaskrik

12 Y var väldigt mycket stressande men gick snabbare att genomföra då jag bara ville bli klar med den.

13 Att ha både en klocka och ett barn som gråter gjorde att jag kände mig mest stressad då det kändes som att jag hade en tid att passa.

14 Den gråtande bäbisen var delvis mer stressande

Question: “Did you find any of the levels (X and Y) to be more fun to play than the other?”

Participant Motivation

1 Nivå X kändes mer fridfull.

2 Roligt med ett adiovisuellt spel. Är själv van vid att kunna se var jag ska nu förlitade jag allt på hörsel.

3 Även om det var roligt med utmaningen av fler ljud i Y så gjorde skrikljudet mer att man längtade tills den var över.

4 Behagligare att slippa ticket och bebisskriket, i övrigt kändes de identiska så det blev det avgörande.

5 STRESS, bäbisar och tickande är jobbigare än fåglar!!

6 I rather wanted to stay there, it was more fun to look for the man in that environment. (Ville hellre stanna kvar där, det var rätt roligt att hitta mannen i den miljön)

7 Bara för att jag klarade den lättare och snabbare.

8 Mer spännande och det kändes som det var mer bråttom.

9 Even if the experience of Y is more stressful, it makes it more fun to perform when being a bit tense. (Även om upplevelsen av Y är stressigare så är det roligare att prestera ifall man är lite spänd.)

10 Det var roligare att ta det lugnt och lösa pusslet på bana X än att känna sig stressad när jag gjorde det på bana Y

11 Bana x var mer meditativ i upplevelsen och bana y mer motiverande. Ingen av dem var mer eller mindre roliga.

12 Konstigt nog uppskattade jag Y mer då banan i sig kändes lättare och inte lika frustrerande när det kändes som att ibland hade X väggar runt omkring mig.

13 Jag gillar att ha lite press över mig. Det gör att jag vill prestera bättre 14 Det var ingen större skillnad i upplevelsen

Question: “Is there anything else that you want to say about the experience and/or differences of Level X and Level Y?”

Participant Answer

1 Naj

2 Kände att X var lugnare med skogsljuden och inte lika stressigt som Y med bebis och klockor

3 Ljuden i X gav mer stämning än att öka svårigheten av banan, medan Y försvårade det en aning, om än på ett lite obekvämt sätt.

4 Inget direkt i banorna, bara en allmän högre säkerhet i hur spelet skulle spelas.

5 gillade fågelsångsbanan mer bror

6 Jag tror mig uppleva att miljön med barnskrik och tickande klocka gjorde det lättare att

"fokusera" på vart mannen var i världen. Det kändes som att jag fick ett mer lokaliserat hörande med den ambiensen

7 Jag tror att den kan ha varit slumpen men fastnade som sagt mot en vägg i level X när jag var precis vid personen, så det var som att jag bara behövde gå runt den men när jag inte såg väggen var det svårt. Det hände någon gång i lever Y också men inte lika ofta så det

blev lättare att hitta personerna. Annars var det ju egentligen lättare att höra ropen i level x då det distraherande ljudet bara var naturljud.

8 Banna Y blev kanske lättare för att man tränat mer.

11 Jag funderar på ifall bana y var svårare för att barnaskriket och klockan var en uppenbar komposit av ljud som inte liknade en ambiens i utförandet. Jag tror att om man hade mixat ljudet som om i ett rum så hade det inte varit lika svårt jämtemot den inspelade ambiensen.

Jag tror eventuellt också att fletcher munson kurvan gör bana y mer svårmanöverlig.

12 På bana X upplevde jag ibland det som att jag varken hörde fotsteg eller att jag åkte mot en väg. Men att jag samtidigt inte kom någonstans. Annars har jag förmedlat det mesta.

13 Level X kändes lite svårare då, som jag nämde tidigare, jag glömde att en kunde gå runt objekterna som stod i vägen. Båda var lika roliga att spela :D