Speech Enhancement for Android

"Speech Enhancement for Android (SEA) is an off-line spectrum analyser with the ability to also conduct spectral modifications such as filtering for speech applications."
Join other users on the discussion forum to discuss your ideas and sort out any bugs you have come across:

https://groups.google.com/forum/#!forum/speech-enhancement-for-android

Or join the Facebook community:http://www.facebook.com/#!/enhancementapp

#### Tested Devices (Better suited for dual core devices) ####

Motorola Xoom with Android 4.0.4 (Good)

Samsung Galaxy S III (Quad core, AU release) with Android 4.0.4 (Good)

Sony Xperia with Android 4.0.4 (Good)

Motorola Atrix 2 with Android 2.3.6 (Good, must have sdcard installed)

HTC Desire HD with Android 2.3.5 (A bit slow)

Motorola Atrix with Android 2.3.4 (Good)

HTC Incredible S with Android 2.3.5 (A bit slow)

#### BUGS AND REPORTS ####

Apologies for incompatibilities on devices other than the ones listed above. If you find any bugs or issues please e-mail me ([email protected]) and include what you were doing at the time the bug occurred so I can fix it. If it happens I can't support your device I WILL REFUND YOUR MONEY.

#### Description #####

SEA frequency analysis has a selectable resolution from 128 to 4096 bands, between 0-4 kHz. This tool is great for audio engineers, anyone with a background in linguistics (phonology or phonetics), cognitive psychology (psycholinguistics), speech and music analysis or Digital Signal Processing (DSP). It is a useful tool for off-line frequency analysis of acoustic signals and a powerful educational aid for students in those fields of study.

You can record, distort, enhance and modify your own voice, giving you the ability to hear and see each step you take.

One of the frequency analysis tools available is the spectrogram. The spectrograms generated in SEA are very high resolution, allowing the user to observe harmonics and the fine structure of their recordings.

Initially it was designed for students and researches in Electronic/Electrical Engineering, however, anyone can try their hand at using it. It incorporates some past and present state-of-the-art speech enhancement algorithms which allow you to filter out noisy signals and see and hear the result of these algorithms first hand. Some of these algorithms are currently used in some telecommunications systems.

You can quickly make changes and share them with colleagues or teaching staff via e-mail, Blue-tooth and any other supporting image or sound transfer services. All spectrograms and frequency plots are zoom able via an image viewer.

Complete list of features as of V2.0.1:
- Record audio
- complete spectrogram analysis
- complete short-time magnitude analysis
- Linear predictive analysis:+ LPC Spectrograms+ LPC Magnitude analysis
- Saving of Spectrograms and audio files to desired location
- Distort recorded audio with different noise types from -30dB to +30dB
- Additive noise types include:+ White noise+ Jet+ Babble (human voice)+ car+ Street+ Restaurant+ Airport
- Matlabs JET color scheme
- Multi color image generation
- Exporting of AMS configuration as txt file
- Short-Time Fourier Analysis on all generated audio
- Real time amplitude recording
- dBFS scale on all spectrograms
- Saving and sharing of all images and audio
- The following Statistical based filters:+ spectral subtraction+ Frequency Wiener filter+ MMSE spectral amplitude+ MMSE with SPU+ LogMMSE+ Spectral energy
- The following acoustic modification methods:+ Unit Magnitude+ Random Phase+ Zero Phase
- Peak frequency tracking
- Generate spectrogram png files
- Generate .wav files
- comprehensive help and documentation
- Video tutorials