AUD175 Project 1 Visualizing sound

In this project I was tasked with recording a number of sound sources in different locations and analyzing them in different types of software. I have chosen to record three sounds, a single clap, hitting a bowl and ringing a call bell.

For this project I chose to use two microphones, an Audio Technica AT875R (2021) which is a small diaphragm condenser microphone with a line + gradient polar pattern, and a Shure SM57 (2021) which is a dynamic microphone with a cardioid polar pattern. The audio interface used was a Rode AI-1 (2021) along with a 10m XLR to XLR cable and a generic microphone stand.

I also chose to record the three sounds in two locations my home office and a bathroom. My home office has some acoustic treatment but isn't perfect. It has carpeted flooring and a number of thin sound absorption panels placed around the room but it also has 3 desktop monitors and two 2cm thick wooden desks which cause a number of reflections. The bathroom used is a generic small rental bathroom with tiled walls and flooring along with a bath tub and a glass half door. Lets just say its the definition of an echo chamber.

The audio software used for this project was Adobe Audition (2021) and REAPER (2021). I recorded with Adobe Audition and used its built in Spectral Frequency Display (2021). I used REAPER for its built in equalizer and to find the average time of each sound for the two locations. For the video I used Adobe Premiere Pro to sync and cut together the footage, screen recordings along with the recorded audio.

Step one: Record the Audio

Using Adobe Audition I recorded two files one for each location. Within those files I recorded multiple takes of each sound source for each of the four microphone configurations. Those four configurations were the AT875R and SM57 both with and without their foam wind screens.

I started by setting up the AT875R on a microphone stand with its wind screen and recorded multiple takes of each sound. I then removed the foam wind screen and again recorded multiple takes of each sounds. Then I swapped the AT875R for the SM57 and repeated this process. I then saved the audio file and moved the recording setup into the bathroom where repeated the same process for all four microphone configurations for all three of the sounds.

At the end I had two ".wav" files one for the office and one for the bathroom containing a total of 22 usable sounds. (I forgot to record the bowl in the bathroom with the SM57 without the foam wind screen so I didn't include the similar office recording, and I only noticed this the day after I had worked on everything).

Step two: Prepare the Audio

To prepare the audio for analysis I first imported the camera footage into Adobe Premiere Pro along with the audio I had just recorded. After synchronizing the audio to the footage I went through and chose the best matching takes for all of the configurations and then removed all the unused takes. To maintain consistency I chose to level all the audio clips so they all peaked at -6 dB within the UV Meter. I then exported all the selected clips as single audio file and imported that file into Adobe Audition and REAPER.

To aid in the analyses process I screen recorded Adobe Auditions built in spectrogram of the all the audio clips and imported that screen recording to below the camera footage within Premiere Pro. This is used to visually show the differences between the microphone configurations and the different rooms on a broad scale within the same image.

I then screen recorded REAPER's built in equalizer of all the audio clips. I then imported that screen recording into Premiere Pro where separated all the audio clips. I then layered these clips on top of each other with an opacity effect to create 6 gifs two gifs per sound source containing all four microphone configurations. This is used to demonstrate how the rooms natural reverb effects the a sounds sources release tale and to show the real time frequency differences of the chosen audio clips.

Clap

Bowl

Call Bell

Glossary of Terms

Equipment and software used

Audio Technica - AT875R

Shure - SM57

Rode AI-1

Adobe Audition

Reaper

Premiere Pro

Audio Recordings with Spectrogram.mp4

The video above is a montage of the three sound sources recorded with the four microphone configurations in the two rooms. The section at the bottom of the screen a Adobe Audition's spectrogram of the audio that was recorded and is synced to the footage. I decided to created this video to show the environment and microphone configurations that were used to record the three sound sources.

Step three: Analyse the Audio

Below is a collection of images used to aid in the analysis of the three sounds along with an analysis of the spacific sound. On a broad scale however I noticed how the location had the same effect regardless of the sound. Comparing the recordings of the office to the bathroom you can clearly see how the sounds recorded in the bathroom are much longer due to the sound reflecting off the hard tiles covering the small room. Compared to sounds recorded in the office which seems to have had very little effect on the length of the sound due to the room having some form of acoustic treatment and carpeted flooring.

When comparing the difference in sound between the SM57 and the AT875R I noticed the noise floor of the two microphones were very similar even though I had to increase the gain of the SM57 a lot more then the AT875 to achieve a similar recording level. This was do to the fact that the SM57 is a dynamic microphone and AT875R is a small diaphragm condenser which uses phantom power to boost its signal meaning its more sensitive then the dynamic SM57.

For each sound I recorded with and without the foam wind screen for each microphone. I noticed that in this instance the foam had no effect on the sound recorded. I believe this is due to the sound sources being very short meaning there isnt much to really compare. If I had recorded vocals for instance you would have seen a larger difference between having and not having a foam wind screen.

Sound Source: Clap

Adobe Audition. Spectrual Frequency Disaply. (2021).

The image above is the spectrogram of the sound of a clap being recorded with a AT875R (1,2,3,4) and SM57 (5,6,7,8). This was recorded with two takes per microphone one with a foam wind screen (1,2,5,6) and one without (3,4,7,8). Sounds (1,3,5,7) were recorded in an office and sounds (2,4,6,8) were recorded in a bathroom.

I found that the average length of a clap in the office and the bathroom was 0.873 seconds and 1.394 seconds respectively. This means the natural reverb generated by the bathroom is adding an additional 0.521 second release tail to this sound source.

Because my hands aren't objects that can resonate like a bowl or a bell all the sound you see in the spectrogram is the sound of the clap reverberation around the room and not from the clap itself. Which is why you don't see any resonating frequencies.

Looking at this images and gifs I found that my claps are composed of mostly low mid frequencies from about 100hz to 2khz then it drops by about -6dB at 2.5khz and tapers off till around 10khz.

This gif is of all four takes of a clap being recorded in an office.

Office Clap.gif

This gif is of all four takes of a clap being recorded in a bathroom.

Bathroom Clap.gif

REAPER. ReaFIR. (2021).

Sound Source: Bowl

Adobe Audition. Spectrual Frequency Disaply. (2021).

The image above is the spectrogram of the sound of a bowl being hit recorded with a AT875R(9,10,11,12) and SM7B (13,14). This was recorded in two takes per microphone one with a foam wind screen (9,10,13,14) and one without (11,12). Sounds (9,11,13) were recorded in an office and sounds (10,12,14) were recorded in a bathroom.

I found the average time of hitting a bowl the office and the bathroom was 0.894 second and1.488 seconds respectively. This means the natural reverb generated by the bathroom is adding an additional 0.594 second release tail to the sound of a bowl.

Because a bowl is an object it will have some frequencies that resonate with it

Looking at this image and gifs I noticed that the bowl I was using has a number of resonant frequencies, mainly around 700hz, 1.2khz, 1.7khz and 2.9khz. I believe this is due to the rim of the bowl vibrating and amplifying the sound due to the concave shape of the bowl acting like a type of megaphone.

This gif is of all three takes of the bowl sound source recorded in an office.

Office Bowl.gif

This gif is of all three takes of the bowl sound source recorded in an bathroom.

Bathroom Bowl.gif

REAPER. ReaFIR. (2021).

Sound Source: Bell

Adobe Audition. Spectrual Frequency Disaply. (2021).

The image above is the spectrogram of the sound of a call bell being recorded with a AT875R (15,16,17,18) and SM7B (19,20,21,22). This was recorded in two takes per microphone one with a foam wind screen (15,16,19,20) and one without (17,18,21,22). Sounds (15,17,19,21) were recorded in an office and sounds (16,18,20,22) were recorded in a bathroom.

I found the average time of a call bell in the office and the bathroom was 7.467 seconds 10.712 seconds respectively. This means the natural reverb generated by the bathroom is adding an additional 3.245 second release tail to the sound source.

This sound is a little more interesting then the other ones because you can see the defining frequencies that make up the bell, These being 1khz, 2.65khz, 4.94khz and 7.8khz. I also find it interesting that the tone at 1khz and 2.65khz oscillates which is visible on the spectrogram and is audible as a sort of "wobble" in the tone.

This gif is of all four takes of the bell sound source recorded in an office.

Office Bell.gif

This gif is of all four takes of the bell sound source recorded in an bathroom.

Bathroom Bell.gif

REAPER. ReaFIR. (2021).

This is an image of the different sound sources that were recorded in the different rooms compiled into REAPER. Here I have placed the each sound for each room on its own layer. Doing this I am able to take the total length of the clips and divide it but the number to clips toe find the average length of the sound in that location.

REAPER. (2021).

Average Time is calculated by dividing the total time by 4 (3 for the bowl). The difference in time is calculated by subtracting the office average time by the bathroom average time.

Conclusion

Over all I really enjoyed this project and the recording process. It was obvious from the beginning that the bathroom was going to have a huge impact on the sounds but though this project I was able to visually show and quantify its effect and what caused it. It also showed me that although my office has acoustic treatment its not perfect and is something I would like to work on. The spectrogram If by far my favorite way to analyze audio due to it displaying a large section of sound at once. In my personal experience I have used it to identify noise sources in my office and rectify them and to measure the ambient noise floor of my recordings.

Reflection on DATA, Process and Feedback

I've been analyzing my own audio for the past 4-5 years and tweaking settings on my equipment to better my voice online and on discord. The spectrogram has always been a favorite of mine especially since its built directly into the timeline on Adobe Audition (Adobe, 2021). However I only use it to identify the noise floor and detect consistent hums like PC fans and crickets outside. This project was interesting because I wasn’t looking at the at those I was instead looking at the sound itself. Identifying it's composition and makeup along with looking into how the room effects the sound being recorded.

This project wasn't without its hiccup's though. For instance it was meant to be a group project but my partner had left the campus and it was too late for me to team up with anyone else. Leaving me to fly solo and record, analyze and complete the project myself. I do feel the analysis could have gone more in depth however with my limited knowledge on sound, music and harmonics I wasn't able to dig any deeper. Looking at the data due I'm unable to see much of a difference between many of the clips. I had recorded with two different types of microphones and configuration hoping it was going to be an obvious difference but looking at the data I can only see a difference between the rooms not the microphones.

For feedback on this project I was told that my process and data collecting was great however my first attempt at a reflection was not on track to be a passing grade and initially I had no intext references or references at all. Luckily David extended the due date for the whole class due to some obstructions which allowed me with this feedback to rewrite my reflection and add all the relevant references. Referencing is new to me and the way I initially wrote this project was from personal experience. I tried to add intext references thought out the project however I am un able too due to not knowing how and where to place them.

References

Adobe. (2021). Adobe Audition Audio recording and editing software. Adobe. Retrieved 8 November 2021, from https://www.adobe.com/au/products/audition.html.

Adobe. (2021). Adobe Audition Spectral Frequency Display. Adobe. retrieved 8 November 2021, from https://helpx.adobe.com/au/audition/how-to/audition-spectral-frequency-display-cc.html

Adobe. (2021). Adobe Photoshop Official Adobe Photoshop. Adobe. Retrieved 8 November 2021, from https://www.adobe.com/au/products/photoshop.html.

Adobe. (2021). Adobe Premiere Pro Professional video editing software. Adobe. Retrieved 8 November 2021, from https://www.adobe.com/au/products/premiere.html.

Amatriain, X., Chew, E., & Foote, J. (2006). Proceedings of the 1st ACM Workshop on Audio and Music Computing Multimedia. ACM Press.

Audio-Technica. (2021). AT875R Line + Gradient Condenser Microphone. Audio-technica. Retrieved 8 November 2021, from https://audio-technica.com.au/products/at875r/.

Lukin, A. & Todd, J. (2006). Adaptive Time-Frequency Resolution for Analysis and Processing of Audio. Tka4.org. Retrieved 9 November 2021, from http://tka4.org/materials/lib/Articles-

Books/DSP/A.Lukin/LukinTodd-MultiRes.pdf.

REAPER. (2021). Audio Production Without Limits. Reaper.fm. Retrieved 8 November 2021, from https://www.reaper.fm.

Rode. (2021). Ai-1 Audio Interface. Rode. Retrieved 9 November 2021, from https://www.rode.com/interfaces/ai1.

Shure. (2021). SM57 Dynamic Instrument Microphone. Shure.com. Retrieved 8 November 2021, from https://www.shure.com/en-ASIA/products/microphones/sm57.

Page updated

Google Sites

Report abuse