Noise Removal from Audio Recording for an Animal Photographer's Promotional Video

The purpose of the diploma thesis is to demonstrate the process of improving sound by removing noise from an animal photographer's promotional video, which was created as a team project. The product contains field video recordings showing the photographer at work, an interview with the photographer, and a musical background. Through the diploma thesis, the following findings were reached. If we remove noise from an audio recording, it will become clearer or more distinct.

Problem

The purpose of the diploma thesis is to demonstrate the process of improving sound by removing noise from an animal photographer's promotional video, which was created as a team project.

Solution and Method

The product contains field video recordings showing the photographer at work, an interview with the photographer, and a musical background. In the interview, the photographer talks about herself, her experiences, and the challenges of photographing animals. The interview was recorded in video format.

Due to the poor quality of the audio portion of the video interview and the desire to perfect the text, the interview was recorded again. This time only in audio format. Due to the content of the video recording, the new audio recording could not be used throughout the entire product.

Therefore, their product consists of two different audio recordings of the speaker. This was also the basis for using two different noise removal methods. In the parts where the video interview remained, they used the graphical method for noise removal called *** Noise Reduction process***.

With this method, we erase noise directly from the audio recording. This is done using various tools in the Waveform view. This method is not virtual and affects the recording destructively.

This means that it saves changes directly to the audio recording. With this method, we must be careful not to delete parts of the audio recording that we wish to preserve during the noise removal process, as after saving, this information would be permanently lost. In the parts where the re-recorded audio recording was used, they employed the ** algorithmic method for noise removal**.

With this method, we remove noise using the Adaptive noise reduction effect. This effect can be added either to an individual audio clip or to the entire audio track. This is a ** virtual effect** that does not affect the recording destructively.

This means that the changes created by the effect are not transformed or saved directly to the audio recording.

Results

Through the diploma thesis, the following findings were reached. If we remove noise from an audio recording, it will become clearer or more distinct. Thus, the first hypothesis H1 can be confirmed. Noise in an audio recording, to a certain extent, ** obscures the information** that we wish to hear or communicate.

Therefore, after removing or reducing noise in an audio recording, the information becomes clearer or more distinct. The noise range in an audio recording is determined based on the noise recording in the room called Noise Print. In it, only ambient sound is recorded, that is, the sound of the environment without speech.

With the help of the Noise Print, we can hear to what extent noise is present in the audio recording and which information will be removed from it. The second hypothesis H2 is also confirmed. The method chosen for noise removal depends on the type of noise in the audio recording.

Noise in audio recordings can vary. Therefore, removing all types of noise using only one method is not as effective as adapting methods according to the type of noise. The Adaptive Noise Reduction effect, which was used on the re-recorded interview audio, quickly removes variable broadband noise and imperfections such as background sounds, rumble, and wind. The Noise Reduction Process Effect is the procedure that was used on the audio from the video interview.

With this procedure, noise is removed using the Noise Print, which is selected graphically from the audio recording. With this procedure, we can precisely determine which frequencies will be removed from the audio recording. The third hypothesis H3 is also confirmed: Audio processing will directly improve the quality of the final product, as a video recording is an inseparable whole of sound and image.

Sound in a video directly influences the viewer's experience when watching the video. Therefore, processed and quality-improved sound directly affects the quality of the final video product. With both methods of removing noise from audio recordings, they achieved the stated purpose of the diploma thesis – improving the sound in the animal photographer's promotional video.

With the help of this diploma thesis, the reader can learn about the procedures for removing noise from audio recordings in the Adobe Audition CC 2015 programme. In this way, we solve the problem of noise in video sound and obtain a higher-quality final product.