FreeDV plus Video

FreeDV plus Video: Technical Details

The theory and principles behind this new mode

FDMA: The Technology Behind FreeDV plus Video
Frequency Division Multiple Access (FDMA) is a technology with wide usage in satellite and cellular communications. It allows multiple users (FreeDV and ZL2AFP's FFT-TV, for example) to share a common signal. A problem of FDMA systems is intermodulation (IM) products, both between users and outside the signal bandwidth. Joe Montana has prepared an excellent presentation on FDMA, part of a lecture on multiple access in satellite systems. Slides 1 through 17 cover FDMA. The book "Satellite Communications" by Pratt, Bostian and Allnutt provides a good discussion of FDMA on pp. 223-233 (see the "Intermodulation Example" on p. 228).
Montana Link - - Book Link

OFDM Tutorials with Application to FreeDV plus Video
Orthogonal Frequency Division Multiplex (OFDM) signals are used for digital television, WiFi, and even the FreeDV component of the FreeDV plus Video signal! (the sub-carriers for the video component are orthogonally spaced, but the video signal is analog with luminance and chroma information transmitted as FM deviation on the sub-carriers). "Understanding OFDM" is a 4-part tutorial from Wireless Training Solutions (WiTS), with an emphasis on WiFi signals. Part 3 covers forward error correction (FEC), with a tutorial showing how convolution coding works (FEC is used with the FreeDV "1600" mode). The Decision Tree links use animation to show how FEC works, using the example in the tutorial (a WiTS "Silver" membership -- free of charge -- is required to view the animation links).
FEC Tutorial Link - - TX Decision Tree Link - - RX Decision Tree Link

Another OFDM tutorial is part of Charan Langton's series on digital communications engineering. The discussion of orthogonality (spacing of sub-carriers) applies to both the FreeDV and video components of the FreeDV plus Video signal.
Langton OFDM Link

Measuring FreeDV plus Video System Linearity
"Shoulder Attenuation", viewed on the FreeDV spectrum display for the received signal, is a useful measure of linearity for the FreeDV plus Video system. See pp. 62 (para. 9.10) and 115-116 (Annex E, para. E.10, E.10.1-E.10.3) of ETSI standard TR 101 290. The "UHF Channel 47" example on p. 116 is similar to the upper shoulder of the FreeDV plus Video signal (of course, the lower shoulder should also be examined!).
ETSI Shoulder Attenuation Link

Three Problems Affecting the FreeDV plus Video Transmitted Signal
The IMD Problem: Intermodulation Distortion (IMD) causes interference to adjacent channels on either side of the FreeDV plus Video signal. It is caused by interaction between the FreeDV plus Video sub-carriers due to overdrive to the transceiver audio chain from the sound card. Click the link below for details on IMD, its causes and solutions, by W8JI.

Adaptive predistortion, a technology new to amateur radio, can reduce overall intermodulation distortion (IMD) by 20 dB or more. Reduced IMD will greatly improve voice and image quality with FreeDV plus Video. Adaptive predistortion has been developed for the OpenHPSDR project, and is incorporated into the PowerSDR_mRX software used with the Apache Labs ANAN-10E/10/100/200 series of SDR transceivers. A paper on adaptive predistortion, by Dr. Warren Pratt, NRØV, can be downloaded at the link below. Or see Dr. Pratt's 2014 presentation (PDF or video) at Friedrichshafen (NRØV developed the adaptive predistortion software for PowerSDR_mRX).
W8JI IMD Link - - NRØV Link - - NRØV Presentation PDF Link - - NRØV Presentation Video Link

The Power Sharing Problem: When the amplitude of the FreeDV sub-carriers is too high (with respect to the video sub-carrier amplitudes), the result is IM products and noisy video. So a separate adjustment of the FreeDV sub-carrier amplitude is required, and the adjustment procedure is described on the "Operating" page. This issue is called "power sharing" in the FDMA literature. Discussions of power sharing are found in the book "Satellite Communications" by Pratt, Bostian and Allnut, pp. 231-232 (book cited above) and in the book "Satellite Communications" by Dharma Raj Cheruku, p. 205 (link below).
Book Link

The PAPR Problem: Multi-carrier modes have a high peak-to-average-power-ratio (PAPR), typically 6:1 for FreeDV plus Video. Thus, the transmitted average power must be kept low enough ("back-off") so signal peaks are not clipped. Click the link below for details on PAPR, its causes and solutions, from IJCCTS.

How HF Propagation Affects FreeDV plus Video
Tony Bombardiere, K2MO, has prepared an excellent video demonstrating the performance of FreeDV digital voice under various conditions of HF propagation. The older "1400" and "1600" modes were tested using AE4JY's PathSim propagation simulation software. Click the link below to view the K2MO video.
K2MO FreeDV Propagation Link

WA6NUT has prepared a PDF document showing propagation effects on the FreeDV plus Video signal. AE4JY's PathSim software was used to demonstrate the performance of FreeDV+ ("1400" mode) under 15 propagation conditions.
WA6NUT FreeDV+ Propagation Link

An Alternate Approach to Video Conferencing over HF
Researchers at two telecommunications institutes in Portugal have developed another approach to video conferencing over HF. This approach makes use of MPEG-4 audio and video compression. The MPEG-4 standard incorporates a method for animating a neutral face model (a still face image), providing facial movement in sync with the speech being transmitted.

The scheme can provide high video resolution and frame rate, but is limited to the transmission of a facial image with computer-generated facial movement. In contrast, FreeDV plus Video offers low video resolution and frame rate, but with no limitation on subject matter (using a video camera or AVI video clips).
HF Video Conferencing Paper: Link 1 - - HF Video Conferencing Paper: Link 2

Design Features in the FreeDV plus Video Software
Digital Voice (FreeDV) Component by Dr. David Rowe, VK5DGR
Digital voice uses a COder-DECoder (CODEC) to efficiently compress and de-compress the analog voice signal. FreeDV uses the CODEC2 software developed by Dr. Rowe. Click the links below for a clearly-written description of the CODEC2 design (link 3 is a video presentation by Dr. Rowe).
Rowe CODEC2 Link 1 - - Rowe CODEC2 Link 2 - - Rowe CODEC2 Link 3

Video (FFT-TV) Component by Dr. Con Wassilieff, ZL2AFP
The original ZL2AFP FFT-TV software has been modified for compatibility with the FreeDV signal. More information on FFT-TV (now known as Analog OFDM NBTV) is found at the link below.

A compression technique, "chrominance subsampling", is used in ZL2AFP's TX and RX software. Without chrominance subsampling, 108 sub-carriers (instead of 49) would be required just for the video alone! The ZL2AFP scheme uses a 36:4:9 subsampling format for the YUV video. Click the link below for a diagram showing subsampling for FreeDV plus Video. Click the links below for details on chrominance subsampling by Douglas Kerr and Charles Poynton.
Diagram Link - - Kerr Link - - Poynton Link

In the ZL2AFP's RX software, the 49 FM'ed video sub-carriers are processed by a bank of bandpass filters, using a "transmultiplexer" scheme. The transmultiplexer also converts the frequency division multiplex (FDM) signal to a time division multiplex (TDM) format. The transmultiplexer is based on the Fast Fourier Transform (FFT) algorithm. Three separate 256-bin transmultiplexers are used: a transmultiplexer at the front end of each of three separate signal processing channels. The channels are almost identical -- except for the last stage, where different motion filters are used. The outputs of the motion filters are used to build the image for each frame, pixel-by-pixel. Click the link below for details on how transmultiplexers are implemented in software, by John Treichler.
Treichler Link

The ZL2AFP "FFT-TV" video system uses an analog scheme. The Y values are represented by the FM deviation of each of the 36 Y sub-carriers (one sub-carrier for each line in the displayed image). The U values are represented by the FM deviation of each of the 4 U sub-carriers. The V values are represented by the FM deviation of each of the 9 V sub-carriers. In ZL2AFP's RX software, the output of each transmultiplexer is applied to a separate FM demodulator. Click the link below for details of how an FM demodulator is implemented in software, by Eric Wicklund, KR7A.
KR7A Demod Link

- 5 July 2013 -

  Back to WA6NUT home page  

(Search the entire site from the home page)