Interview with Hrishikesh Dani

Since Designing Sound is now the home of the Sound Design Challenge, I thought it only appropriate that the winner’s interviews be posted here as well. Here’s the interview with Hrishikesh Dani, winner of April’s challenge – SDC009: The Game Audio Challenge…

Designing Sound: So, why don’t you give us a little description of who you are, what you do and how you got into sound work.

Hrishikesh Dani: In the late 90s, the arrival of an acoustic guitar at home opened my ears for music. After having developed a taste for wide range of music genres, I started digging more on the engineering and technological aspects of music production. While pursuing a Bachelor’s in Electrical Engineering at the University of Mumbai, I focused on topics such as DSP and Electronic Circuit Design. This provided me an opportunity to do some circuit bending and building analogue processors for my electric guitar as my final year projects.

After graduation in 2007, I started working as an Assistant Sound Engineer at a post-production studio called Bandwagon Studios in Mumbai. Even though it was exhausting, I was lucky to assist two studio rooms at the same time as I got to learn Pro Tools on Mac and Nuendo on Windows. Later on, I worked as a Studio Audio Engineer for Globe Recording Studio, where I got an opportunity to supervise the studio construction and record/edit localization IVRs. After a few months, I got a splendid opportunity to work as a Sound Designer in the promo department at UTV Television Networks, Mumbai. I got good experience working in a steep deadlines environment as I was designing sound for five UTV television channels.

I enrolled for Masters in Sound Design programme at the University of Edinburgh in 2009. The programme at UoE provided a great opportunity to learn some amazing interactive and real-time programming tools such as Max/MSP, CLM, and I was lucky as they introduced Game Audio tools such as FMOD in the curriculum the same year. I was fascinated by FMOD and wanted to explore more into game audio interactive middle-wares. Having played games all my school and college years made it seem as a perfect career path where I can truly enjoy the challenge of game audio designing and implementation along with being a part of such an amazing industry.

Currently I am trying to find my place in the gaming industry world-wide, so job applications are on roll even though there are very few vacancies around. To keep myself busy I am working on couple of hobby indie games and mods, exploring field-recording and learning more game-audio tools and techniques.

DS: You mentioned on the site that this was your second time using Wwise (the first time replacing sounds for the sample game Audiokinetic provides, Cube). How comfortable were you with the software before the announcement of the challenge, and how did you use the month we gave everyone to prepare?

HD: Yes, a few months ago, I had a very brief encounter with Wwise. Cube, the sample game provided by Audiokinetic was a good starting point but I was disappointed with it because it only lets the Wwise learners to replace certain existing sounds with their own designed ones. Implementation wise I was still at level-zero.

Even though I had gone through every tutorial video of Wwise on their YouTube channel, I was not very comfortable with it. For me the results of actual implementation give confidence, and prior to the challenge, I was clueless on how to go about practicing design and implementation with Wwise. The sample project of 24-hour Forest Ambience provided by Richie Nieto for the challenge was a great boost for me and it helped me understand RTPC and Actor-Mixer Hierarchy structure in Wwise better.

During the prep time, I was thinking more on various possible background ambience concepts such as a 24-hour city traffic ambience, or a workplace/office ambience, or even a soundscape for a linear/non-linear story (for egample, Escapement by J.G. Ballard). I spent more time on ideating the layers of elements and creative implementation approaches for each concept instead of playing around with Wwise. This gave me a good head start in understanding what can be possible within a powerful and sophisticated interactive audio middle-ware tool such as Wwise.

DS: So did your previous experience with FMOD provide a rough path for you to develop as far as work-flow and implementation in Wwise? How would you compare these two middle-ware applications?

HD: My previous experience with FMOD is more varied compared to Wwise. I have worked on some interactive audio projects and a couple of game mods where FMOD is used as the audio middle-ware. For example, Crysis Wars uses a very, very old version of FMOD hooked up to CryEngine using Lua and XML scripting; and then there is Dragon Age: Origins which uses FMOD along with its own proprietary toolset. Work-flow and asset management wise, these two games have approached FMOD differently and it is always a great learning experience to observe and analyse how AAA Game Audio guys do it. It is a shame that I could not find any game SDK which uses Wwise to mod with, except for the Audiokinetic sample game, Cube.

I think my FMOD experience facilitated my understanding of how a game audio middle-ware works. The basics can be translated from one into another; whether you are working with FMOD or Wwise or XACT or on some game engine such as Unreal or Unity 3D. As far as work-flow and implementation techniques are concerned, FMOD and Wwise have their own unique features and approaches, so it is obvious that the work-flow will be different but the idea behind implementation and asset management remains somewhat similar in both. Software stability wise, I found Wwise to be sturdier as opposed to FMOD, or maybe the previous versions of FMOD were buggy and I feel the latest versions are more stable. Also both Firelight and Audiokinetic are upgrading their middle-wares with interesting stuff heading towards granular and procedural audio, something to look forward for.

DS: Let’s talk a bit more about your prep work. We gave out some hints as to what the challenge would be (an ambient background) early on. How did you use that information, and what happened to the ideas you came up with once the specifics of the challenge were revealed?

HD: To me SDC009 was a perfect reason and a great opportunity to explore Wwise more thoroughly, so I worked towards the challenge from a learner’s perspective and tried to push the limits within the realms of whatever I learnt. Prior to the challenge, it was more about what could be possible with my limited knowledge of Wwise. The concepts I mentioned earlier were helpful for me to explore the possibilities of different approaches for each idea but I didn’t put them to practice due to other commitments during that month.

Once the specifics of the challenge were revealed, I liked what the challenge was all about; creating a background ambience for a thunderous and stormy day at the beach was something I didn’t think of during prep time. The very first thing I did was to do some research on various thunderstorm models, about the various stages involved and the average time duration they last for. I realized Monsoon Thunderstorm model had some interesting stages of progression from calm to thunder and back to calm within a span of 3-4 hours. As I hail from Bombay, Monsoon Thunderstorm is something I am very much familiar with and I felt I should re-create my memory of a thunderous day for the challenge.

DS: Sonically, your entry was very distinctive and had a nice progression. So I think you were definitely able to connect it with that personal experience. You also had a lot going on under the hood. What are some of the techniques you incorporated that you’re particularly proud of?

HD: Thanks. Yes the personal experiences usually help in re-creating realistic settings.  Also, thanks to the fantastic BlastwaveFX sound effects pack provided, almost all the elements required to create the background ambience for a stormy day at the beach were there. Regarding things under the hood, there were indeed a lot of things going on but if I start explaining them it will take quite a bit of the text space here. So it will be good if the interested readers head over to the blog post which I will be writing soon on my blog website detailing the implementation techniques with pictures and videos.

Though some of the implementation tricks which I feel can be mentioned here are the ones I used for distance based volume & EQ changes in surrounding sounds from a listener’s perspective for a span of 50 metres from the shore. Also careful placement & positioning of appropriate randomized SFX on the timeline & 3D space, and volume/low pass filter automations for time & distance based parameters in individual blend containers and then into a master blend container resulted in a smooth progression of overall combination of Crowd, Ocean Waves, Rain, Thunder and Wind blend containers.  And somewhat unusual re-purposing of Silence Sound SFX into Information Slates for the main blend container to make it easier for the reviewers to understand the precise time based progression stages was something that clicked as a smart idea while figuring out how to fill in the blank space in the main blend container.

DS: You used two adjustable parameters in you entry; one for Distance and one for Time (if I’m remembering correctly at the moment). Since the audio engine is as important as the sounds that go into it, how do you imagine those parameters being utilized together in a game? What’s your ideal use of that particular implementation idea?

HD: For SDC009, The combination of Distance and Time parameters worked well to convey the idea of free listener movement on a path perpendicular to the sea-shore, while the thunder storm progresses on positive timeline. However, in a 3D game situation, I don’t see much significant use of Time parameter as compared to Distance parameter (possibly the most important and widely utilized parameter).

In a First Person or Third Person perspective game, it makes sense to use Distance parameter to spawn/trigger sound items placed at various positions in the map level. As the player character moves in a level, whenever he/she reaches inside the scope of the sound item/event, the corresponding sound will be spawned and gradually increase in volume as the player gets closer to the sound item along with other DSP effects applied on it and vice versa. Other parameters such as Speed, Mass/Weight, Intensity of the Event, etc may work better paired with Distance parameter for events such as footsteps, body falls, etc.  On the other hand, Time parameter is somewhat restricted to linear sequencing of events on a positive timeline and hence is less interactive compared to Distance parameter.

Using both these parameters together in a game is possible in some special events.  Imagine a situation with Day-time/Night-time differences in the behaviour of a sound item; for example a Tree. The daytime sounds can be of breezy leaves rustling layered with occasional cheerful bird twitter, while the night-time sounds can be of ominous rustling layered with occasional nocturnal bird sounds such as an Owl’s hoots. As the player character enters inside the scope of the tree sound event, corresponding sounds will be spawned depending on the day time or night time inside the game. One might think why not any outdoor ambience sound could have the same implementation? The answer is simple, because there is no need for a Distance parameter for ambience, so a 3D Spread parameter along with Time parameter will do a better job for ambience sounds.

DS: What is something that you’ll take away from this challenge, and how will you apply that to your work in the future?

HD: This challenge provided me a perfect reason to indulge more into Wwise and explore its functionalities. Working with restrictions or limited knowledge to create more creative and unusual results is something that I have been trying to do for past few months whenever I attempt to learn a new game audio tool. This challenge provided me a good platform to learn Wwise and push the limits when it came to scheming out the overall result, figuring out the Actor-Hierarchy structure and work flow, selecting appropriate sound assets to blend into each other smoothly and re-purposing the units and functionalities of Wwise to realize the challenge effectively.

In future work, the problem solving aspect will definitely be the most helpful one. I believe every problem has a most viable solution; one just has to make an extra effort to find it. Also working with restrictions to carve out new possibilities is something which can be very fruitful in learning new techniques.

DS: I know you looked through the other entries when I made them available for download. Who and what are some of the things that stood out to you when exploring them?

HD: Yes indeed, I learned something new from each entry. I am amazed by how each entrant has a unique approach and perspective on implementation techniques. Some entries which stood out for me were by LucaF, JackM, OlaS and ChristianF. LucaF’s stereo widening implementation technique was quite interesting; and also, the detailing under the hood was impressive. JackM’s entry was quite outstanding. He took a slight deviation from the brief of the challenge, but he did manage to do it creatively. I liked the use of reverb and RTPC with distance parameter in his entry. OlaS’s entry was impressive as, well with some interesting positioning of elements. And finally, ChristianF’s FMOD entry had a good use of Time and Distance parameters in it. Overall I would say, all entries combine as a great Wwise learning resource for anyone who wants to get started with Wwise.

DS: Anything else you’d like to share with the rest of the community before we wrap this up?

HD: Thanks a lot, Shaun. It was fun answering the questions. I think we have covered a lot of things. I wanted to share a blog post on the detailed breakdown on my entry for anyone interested to know about it. Unfortunately, I have been caught up with lot of other stuff lately, but would like to let the community know that I will be posting it sooner or later on my blog http://auralscope.tumblr.com … so stay tuned!

For any who may be interested, the compiled entries are still available over on DynamicInterference.com. A download link can be found at the bottom of this post. The sounds that were used in the challenge are no longer available, but you can still explore the projects without them. They’re a great resource, as they allow you to look at different implementation approaches. Congratulations again to Hrishikesh Dani.