Subject: 'Clear audio for people with hard of hearing’

Dear Audacity development team

I greatly appreciate your open-source software and am grateful for its many years of existence.

However, I miss one important function: the option to clean up audio files specifically for people with hard of hearing. This should be labelled and understood as ‘Clear audio for the hard of hearing’. In many audio files, especially in the audio tracks of videos, disturbing background noises can often be heard. For example, spoken dialogue is often accompanied by background music and background noises such as hissing, humming, scratching, rumbling or traffic noise. Despite modern hearing devices such as hearing aids, it is often difficult for people with hard of hearing to understand spoken dialogue. It is therefore very important for many people with hard of hearing that the audio is clear and distinct. For cleaning, all available tools such as noise cancellation, removal of music and traffic noise, etc. should be compiled offline at the beginning. If possible, these tools should be downloaded as an ‘extension’ and applied automatically at the click of a mouse. The tools concentrate exclusively on the spoken voice and eliminate the rest as superfluous, distracting background.

About the implementation: You should not expect too much for the first step, as there is no perfect solution, but it is a good start. It is important to reduce or eliminate all unnecessary background noise as much as possible. Later, further developments can be made, and the newer version of the ‘extension’ can be optimised. In the case of an “extension” with a version number, it should be noted which version is current. Older versions should also be kept so that everyone can be accessed if necessary.

Figures: According to the WHO (World Health Organisation), a rough estimate of 430 million people worldwide live with hard of hearing. The WHO estimates that around 700 million people will be affected by hard of hearing by 2050 - this means around 1 in 10 people. In Europe, the Hard of Hearing Association estimates that 52 million people are affected by hard of hearing. According to a study, there are around 9 million people with hard of hearing in Germany. The number of unreported cases is probably even higher. More and more people are becoming hard of hearing due to hearing damage or age-related hearing loss.

A quick note: Please do not confuse hard of hearing people with Deaf people. With a good hearing aid, people who are hard of hearing can talk on the phone to certain people around them or have normal conversations in a quiet room. People who are Deaf are dependent on written communication as they cannot hear. They can, for example, communicate via text chat on the Internet or converse in sign language, in which case a qualified sign language interpreter is required if the person they are talking to cannot sign. The “Clear audio” option is aimed at people with hard of hearing.

I would be very grateful for the implementation of a fully automatic cleansing of audio for people with hard of hearing, also known as ‘Clear audio for people with hard of hearing’, at once click of a mouse. Many thanks in advance for the implementation if this is carried out. :slightly_smiling_face:

1 Like

Um, sorry @Slow_jamz, but you’re way off on the frequency range of human voices. :slight_smile: Telephone systems traditionally use 300 Hz to 3,400 Hz (or 3.4 kHz if you prefer). However, filtering out everything outside those frequencies may make some consonants difficult to distinguish, so trying an upper frequency limit of (say) 8 kHz may be better.

@No_background_noise Really the best way to ‘clarify’ speech is to use a plug-in designed for that purpose, such as Klevgrand’s Brusfri, OR an online AI-based service like Cleanvoice. Cleanvoice have a list of voice ‘cleaners’ you can try: obviously their own service is #1 in their list! It’s a case of ‘try them all, then see which one works best for your application.’ If you do that, it would be interesting and helpful if you post your findings here afterwards.

I’ve tried some of these myself, with varying success, but my use case is very different from yours. Good luck!

Thank you for your feedback and recommendations. :slight_smile:

I learnt from Audacity’s news that the add-on plug-in tool ‘openVINO’ is new to me.

I do not think the ‘openVINO’ add-on plug-in tool is bad. However, I miss automatic recognition of spoken language as well as automatic suppression or removal of background noise in order to concentrate exclusively on spoken language. After speaking, natural sounds such as instrumental music or ambient noise can reappear.

Simplified alternative explanation: While the spoken voice can be heard, disturbing background noises are suppressed or removed. As soon as the spoken voice stops, the background noises return, such as instrumental music. When the voice starts again, the background noises are suppressed or removed again. This process continues…

If no solution is available, separating the different acoustic sound tracks after using the ‘openVINO’ add-on plug-in tool requires a lot of manual editing work. Although there is a workaround, it is not ideal.

As already explained, not only people with hard of hearing benefit from clear audio, but also everyone else in order to better understand the content of spoken dialogue.

This topic was automatically closed after 30 days. New replies are no longer allowed.