Telemonitoring with the VLC HQM 16 Audio Codec

Abstract

In the field of the telemonitoring, the VLC HQ 16 codec with the multichannel support (VLC HQM 16) can be used to:
- Record or transmit the body sounds (very low frequencies).
- Record or transmit data such as the ECG (ElectroCardioGram), the EEG (ElectroEncephaloGram) and the EMG (ElectroMyoGram).
- Record or transmit the ABP (Arterial Blood Pressure) waveforms data and the PPG (PhotoPlethysmoGram) waveforms data (from the pulse oximetry).
- Record or transmit the blood glucose waveforms data, in Continuous Glucose Monitoring (CGM).
- Optionally, compute and display the heart rate variability (HRV) and the blood pressure variability (BPV).


Description

For the telemonitoring, several international formats exist for the ECG and the other vital signs recordings.
One can cite:

- The SCP-ECG format, Standard Communications Protocol for computer assisted ElectroCardioGraphy.
- The MFER format, Medical Waveform Format Encoding Rules.
- The FDA-XML format, Food and Drug Administration-Extensible Markup Language.
- The DICOM format, Digital Imaging and Communications in Medicine.
- The ISHNE format, standard output format for the Holter ECG data.
- The EDF+ format, European Data Format.

Many other proprietary formats exist such as the ECG format of Philips which uses XML and the base 64 encoding.
The still image files (GIF, JPEG, PNG, TIFF, and PDF) can also be used as support for the ECG and the other vital signs data.
The sampling frequencies are typically between 1000 and 2000 Hz or less than or equal to 500 Hz. The most used resolutions are: 8, 10, 11, 12 and 16 bits per sample.

- SCP-ECG:
The SCP-ECG format is a digital recording format for the electrocardiograms.
It comprises the ECG signal and the metadata about the patient and about the ECG recording device such as the manufacturer's name, etc.
SCP-ECG is an ANSI / AAMI and CEN norm.
By default, with the SCP-ECG format, the samples are not compressed.
The records are binary with signed 16-bit words.
Three compression options can be used:
- Lossless reduction of the redundancy (differences between samples).
- Lossless compression with the Huffman method.
- High level Lossy compression.
The latter type of compression cannot be easily used everywhere because of the necessary resources, including the processing time and the memory.
With the Holter monitoring system (long duration) or during the physical exercises, this method cannot be used.
This method is recommended for the common ECG recordings with 12 simultaneous channels.

- MFER:
By default the samples are not compressed. The records are made of 16-bit signed integer.
This is an open and flexible format where no method of compression is preferred. It can be used as a simple container.

- FDA-XML:
This format, called also HL7 aECG (HL7 Annotated ElectroCardioGram) is a standard to store and retrieve the ECG data of a patient. Like other HL7 formats (Health Level 7), it is based on XML.
The data are in ASCII and are structured in XML, so the files are usually large.

- DICOM:
The DICOM is a standard for the medical imaging data management.
This standard defines a file format as well as a data transmission protocol (based on TCP / IP). The data can be compressed or not.
The first goal is to store data from the medical imaging (X-rays, Magnetic Resonance Imaging or MRI, scan using ultrasounds, ...), but this format can be used for other data types such as the ECG data or the other vital signs data.

- ISHNE:
The ISHNE format is a single file composed of a header followed by a larger data block containing all the stored ECG digital samples.

- EDF+:
The EDF is a simple and flexible format for exchange and storage of multichannel biological and physical signals, especially the EEG.
The EDF+ file can also contain interrupted recordings, annotations, stimuli, and events.

- VLC HQM 16 codec:
Our goal is to offer the VLC HQ 16 codec with the multichannel support (VLC HQM 16) as a way to compress the ECG and the other vital signs data.

- It is based on FFT (Fast Fourier Transform), so it is very fast and can be further accelerated with the GPU support (Graphics Processing Unit).
- It is quasi-lossless in energy: the energy of an uncompressed frame is almost the same as the energy of the compressed frame.
- The compression ratios are very high (between 8 and 32 per FFT buffer, according to the desired quality. The compression ratio can be higher for some quasi-stationary data).
- The data are derived directly from FFT (frequencies, magnitudes and phases), all the frequency analysis using FFT may be performed directly without doing FFT again.

This codec is optimized for 16 kHz or less sampling rates, with 512 points FFT buffers.
With a sampling rate of 500 Hz, one will issue about one frame per second. It is possible to reduce the transmission frequency of frames (e.g. one frame every two seconds) and increase the size of the FFT buffers (e.g. 1024 points).
Compression rates can be increased in this way especially with quasi-stationary signals as those of the electrocardiogram or those of the arterial blood pressure.
For the low frequency body sounds (heartbeats, lungs, arteries, ...), by default one will use a 16 kHz sampling rate, 512 points FFT buffers and one will issue 31.25 frames per second.

For the stationary signals such as the ECG, the ABP or the PPG, it is expected to see if we can do much better in term of compression ratio using the codebook version of VLC codec. A single integer can be used to represent a vector of the positions and / or the magnitudes of the local peaks, thus can serve as signature for a frame.
For more information, see at the following address:
Codebook Version

The compression is done in real time.
There is no mandatory pre-processing or post-processing.
There is also no problem of edge effects (with less than 10% frame overlapping).

We will write a codec integrated to the PJSIP product, creating compressed samples and being able to generate or to read a W64 file (Sony Pictures Digital Wave 64).
This format was developed by Sonic Foundry and is maintained by Sony. It is also known as the Sony Wave 64 Audio File.
This format supports an arbitrary number of channels, sampling rates and bits per sample. It supports files larger than 4 Go and one can add metadata.
Many audio players, analyzers and editors support this format (Sony Sound Forge Pro, VLC Media Player, ...) with uncompressed samples.
For more information on the PJSIP product, see the following addresses:
   - vlrPhone / vlrMemos
   - PJSIP

We will propose converters from the W64 compressed format (with our codec) to the six formats described above, and from these six formats to the W64 compressed format.
We will also propose a converter from the W64 compressed format (with our codec) to the uncompressed W64 format (or to the uncompressed normal WAV format for sizes less than or equal to 2 GB), and conversely.

The apps will be compatible with the Windows, Android and iOS operating systems, as well as with all the operating systems supported by PJSIP.
The apps will be able to be integrated (directly or with slight modifications) to services, software and hardware which need their features.

Notes

The multichannel can be used to record multiple homogeneous channels in parallel, for example:
- The ECG with 3, 6, 12 or 18 channels.
- The EEG with 25, 32, 64, 96, 128 or 256 channels.
Using the same sampling rate, it can also be used to record multiple heterogeneous channels, such as channels derived from one or more waveforms from calculations, templates or more or less complex algorithms. One can cite for example:
- The RR Intervals.
- The Heart Rate.
- The Systolic Blood Pressure.
- The Diastolic Blood Pressure.
- The Mean Blood Pressure.
- The Pulse Pressure.
- The Stroke Volume.
- The Cardiac Output.
- The Left Ventricular Ejection Time.
- The Total Peripheral Resistance.
- The Aortic Characteristic Impedance.
- The Arterial Compliance.
The compressed data of these channels are directly FFT data, so we can have quickly the values of the spectral density per segment (the variability per segment in the frequency domain).

Optionally, we will compute and display the heart rate variability (HRV) and the blood pressure variability (BPV).
More information on the HRV and the BPV at the following address:
   - Heart Rate Variability and Blood Pressure Variability


Subjects

ElectroCardioGram (ECG)
ElectroEncephaloGram (EEG)
ElectroMyoGram (EMG)
Arterial Blood Pressure (ABP)
PhotoPlethysmoGram (PPG)
Pulse Oximetry Plethysmographic Waveforms
Continuous Glucose Monitoring (CGM)
Blood Glucose Waveforms
Heart Rate Variability (HRV)
Blood Pressure Variability (BPV)
SCP-ECG
MFER
FDA-XML
DICOM
WAV
W64
ISHNE
EDF
EDF+