ISSS608 2017-18 T3 Assign NEVIL BRUNO Q2

OVERVIEW

DATA PREPARATION

Q1: PATTERNS

Q2: CHIRP-CHIRP!

Q3: WHAT NOW?

REFERENCES

BACK TO DROPBOX

Question 2

Kasios now claims that last year's analysis was flawed and biased. To combat these conclusions, Kasios has launched their own “investigation” into the Pipit situation, and they are now reporting that there are plenty of Rose-crested Blue Pipits happily living and nesting in the Preserve. To back up this claim, they have provided a set of Pipit bird calls, recently recorded across the Preserve, with locations of where they were recorded. Clearly, they claim, the Pipits are a thriving population. Using R, we will be taking a closer look at the audio recordings provided by Kasios and validating their claims.

Kasios claims all the recordings taken across the preserve are of Rose-crested Blue Pipits. They have provided 15 recordings taken across different parts of the preserve. Each recording has a unique File ID from 1 to 15. The locations of each recording are given below:

For our analysis, we will be using R to plot the audio files and do a visual analysis to determine various characteristics.

Determining Vocalization Type and Best Audio Quality

Song or Call

Fortunately, we have recordings of every bird species. For our analysis, we will be comparing the Kasios audio files to the Bird audio files to validate Kasios’ claims. As discovered in our initial EDA, there are 2 main vocalization types and 5 levels of audio quality. It is essential that the same vocalization type and best quality audio files be used for the analysis. We first take a sample for call and song for the same bird. In this case, we will use the Bombadil (chosen at random). A call and song of the same quality level (Quality A, chosen at Random) is taken and compared.

Visually inspecting both plots, for Song type audio files, we obtain distinct patterns (better amplitude and better spaced audio tones) which will assist us in our analysis.

Determining Quality

An analysis on the 5 different audio qualities will be done to determine which grade represents the best audio quality. To perform this, we would require the same bird. We will be using 5 different audio quality files of the Bombadil Song files (chosen at random).

The audio plots for each type are given below:

From a visual and audio inspection, D and E are more distorted, and have lower SNR compared to the other 3. B has low amplitude values, which means the sound volume level will be low. Comparing A and C, we get more distinct and clearer audio waveform patterns for A. This is essential for our analysis. Hence, we can rule our B, C, D and E from the analysis and take audio quality A for our analysis.

Obtaining Plots for Audio Files

Bird Species Files

Audio for the 19 different bird species are taken and plotted. Vocalization type Song and Quality level A is taken for each bird. Each file was selected randomly. These same plots will be used for our visual comparison analysis

1. Bent-beak Riffraff

2. Blue-collared Zipper

3. Bombadil

4. Broad-winged Jojo

5. Canadian Cootamum

6. Carries Champagne Pipit

7. Darkwing Sparrow

8. Eastern Corn Skeet

9. Green-tipped Scarlet Pipit

10. Lesser Birchbeere

11. Orange Pine Plover

12. Ordinary Snape

13. Pinkfinch

14. Purple Tooting Tout

15. Qax

16. Queenscoat

17. Rose-crested Blue Pipit

18. Scrawny Jay

19. Vermillion Trillian

Kasios Files

Audio for the 15 audio samples provided by Kasios are taken and visualized.

Sample 1

Sample 2

Sample 3

Sample 4

Sample 5

Sample 6

Sample 7

Sample 8

Sample 9

Sample 10

Sample 11

Sample 12

Sample 13

Sample 14

Sample 15

Comparision of Audio Files

For the analysis, 15 Kasios audio plots are compared to the 19 bird species. Distinct patterns in the audio plots will help us identify which bird species it belongs to. The two plots are placed one on top of the other. The common patterns have been circled. Each colour oval represents the similar pattern i.e. if there are two colours, each colour represents a particular pattern. By identifying these significant and distinct visual patterns, we can map each audio file to the respective bird song pattern.

Comparison Sample 1