top of page
  • Writer's pictureSound Reformer / yamakaWA

Clear voice memos whose echoes are so badly untranslatable (lecture recording)

Important academic societies and lectures that require transliteration. I think there are many people who record the contents with a voice recorder. The other day, I received a consultation that the microphone in the venue was so loud that I could hardly hear the recorded content.


It seems that clients always use voice recorders when attending conferences. In order to avoid the risk of recording failure, we also recorded voice memos on the iPhone as well as a dedicated voice recorder at the same time. He asked us to present two types of files to this service and clear one of them.

The problem that the performer's voice echoes too much in a large space

The recorded sound source was certainly hard to hear as it was. The voice of the performer echoed in the large venue. These improvements are not the ones with a high probability of being adopted through trials, but are difficult projects.

The point of improving such a sound source is

① Suppress ambient noise

② Reduce only reverberation

③ Emphasize the voice of the performer

④ Overall sound pressure increase

The above is the main subject, but if the reverberation component is extremely removed, the balance of the sound will be significantly lost, and it will often be difficult to hear, especially in the case of a speaking human voice.

In addition, the sound that is difficult to hear is that the performer's voice echoes in the venue like a "warn", so even if the reverberation component is reduced, the echo will still be emphasized when raising the performer's voice. It's easy to go around in a dignified manner.

Therefore, it is difficult to present an improved sound source that nods the client unless (1) to (4) are processed in an exquisite balance.

Why this improvement worked

The sound source this time had a low-frequency "goo" sound mixed in with the "(1) ambient noise" component for a long time, so the biggest thing was that removing these sound improved the sound of the performer's voice. It's a success factor. We couldn't reduce the reverberation component so much, but we were able to capture what we were talking about, so we decided to use it safely.

Since this is a relatively difficult project, there is no guarantee that it will succeed, but if you have a problem with an "inaudible sound source" even though you must translate it, please feel free to contact us. You can try it for free with the sound source you actually want to improve.

It was Yamakawa, a sound former of hybrid sound reform that operates a service that clears voice memos.


bottom of page