Good news! Today sees the launch of the project’s first ever app – The Good Recorder. Absolutely free and available now via the iTunes store, or click here.
What is The Good Recorder?
The Good Recorder is a sound recording app (currently only for iOS 7 devices) designed to help users achieve high quality audio recordings by monitoring for common recording errors and providing feedback about them. Currently the app incorporates findings and algorithms from our previous work with wind noise. The plan is to further develop the app with auto-detection of handling noise and distortion as our research in these areas progresses.
I already have an app to record audio, why do I need The Good Recorder?
Good question. Many readers of this blog will likely already use their phones and similar devices to make recordings. What makes our app different is a range of useful and convenient features designed to help improve the quality of your recordings. Uniquely, the app also alerts the user to the presence of wind noise – a valuable tool when making recordings outdoors.
In summary, The Good Recording app;
- alerts the user to the presence of unwanted microphone wind noise.
- provides real-time monitoring of the A-Weighted level in dBA.
- allows for user-calibration of the system, to an approximately absolute level in dBA.
- allows users to adjust sensitivity of the wind noise detector.
- provides an in-built player with visualisation of audio sections containing wind noise.
- allows users to disable the device’s default Automatic Gain Control (AGC) to capture the true dynamics of a sound event.
- stores audio as uncompressed WAV files.
- allows users-assigned metadata (description/tags/location) for recordings.
- automatically stores location for each recording.
- makes it easy for users to share files via or Email (either as WAV or m4a format).
- alerts users with a clip light when the device is likely to have been overloaded.
Tags and other information can be added to each of the audio files. The location of the recording is stored automatically. In playback audio scrubbing is possible by dragging the locator left or right.
SoundCloud users can upload/share their recordings easily, and tags and description will be included when uploaded.
Sound files can also be emailed as attachments (including the option to compress the audio from WAV to m4a first).
What is wind noise and how does the app detect it?
Wind noise is a common problem when recording outdoors, we’ve discussed it often in previous posts. Microphones are very sensitive to changes in air pressure, so even gentle breezes can often be heard as loud and annoying noise in recordings. This app detects the presence of this kind of noise and alerts the user in realtime.
The wind detector has been trained, using a machine learning method, to recognise sequences of audio frames which contain wind noise. Every 23ms the audio is analysed and a decision is made as to whether there is wind noise present. These decisions are aggregated over a short window of time and this forms the basis of the alert.
To get the best performance from the app it needs to be calibrated to the particular device recordings are being made with. This requires a short and simple procedure where the user finds a quiet location and speaks a sentence repeatedly for 5seconds.
How will I know if wind noise is a problem in my recording?
If wind is detected the user is alerted by the on-screen wind gauge turning red. The user then knows to alter their position or location to obtain a better, wind-free, recording. The user has the option to adjust the sensitivity of the detector (and the level of wind noise alerts are provided at).
The player shows a waveform view of the sound. If wind noise is present in the audio the relevant sections are highlighted in red. A slider is also provided which allows for post-hoc adjustment of the wind detection sensitivity.
Why did you make it?
Visitors to our site will already know we think audio quality is important. One way to help others get better quality audio in their recordings is to help to prevent or remove common errors people can make when recording. We are working on a number of experiments which help us to understand the effect of problems such as wind noise, distortion, and handling noise on audio quality. The results of these experiments can be put to good use in projects like The Good Recorder, which we hope will help both experts and amateurs make better recordings.
So then. It’s available now and it’s free. Give it a go and let us know what you think!
[…] The Good Recorder iPhone app includes a meter that uses decision trees to detect wind noise on the microphone signal being recorded and issues a warning to the recordist. The team involved in this research is hoping to publish the machine learning method as an open source toolbox for use by others. […]