loudorren
Well-known Member
Hello to Old Quad Guy, Q-Eight, Dylan Berichon, Lucanu,
Addressing the QSI5022, the performance of the CD-4 demodulator built with these chips was far superior to any of the CD-4 demodulators out there at that time. The reason mostly is because the manufactures all heard through the grapevine about my Phase Lock Loop technique of demodulation. They all figured that if they used a Signetics NE565 PLL chip as the FM detector, this would solve the breakup noise and limited number of plays of CD-4 records. Technics, Panasonic, Harmon Kardon, and Heath Kit were the only OEMs to employ the QSI5022.
The secret to high performance in the CD-4 demodulator was to put a high gain FM limiter before the PLL. I have read some of the other threads on this site, where CD-4 users are making precision tangential adjustments to the pickup cartridge to reduce the 30 KHz subcarrier to subcarrier crosstalk. This is only needed because of the FM processing system in the demodulator.
FM has a unique characteristic known as capture ratio. This is the ability of an FM detector to detect and demodulate the stronger of 2 FM signals on the same frequency. The typical CD-4 demod has a capture ratio of 7 to 12 dB. This means that the stronger signal has to be at least 7 to 12 dB greater than the weak signal to reject the interference from that weak signal.
The QSI5022 has a capture ratio of .75 dB. This ratio allows the demodulator to work well even in non optimum setups. It also had the circuitry to allow external subcarrier crosstalk cancellation for pickup cartridges with poor to medium left-right separation in the subcarrier band (15KHz to 45KHz).
I kept a couple of the FM demodulators and I have two complete Quadraplex broadcast generators.
With regard to IBOC "HD" radio. This is one of those non technology systems I hinted about in a previous posting. This system is as far as you can get from "High Definition" (HD actually stands for hybrid digital). These digital signals are piggybacked onto the current AM and FM analog signals.
First let me introduce the Nyquist theorem. It states "to sample an analog signal and accurately transmit and reproduce that signal, the sampling rate must be at least two times the highest analog frequency to be sampled".
Lets take the case of the Compact Disc. The highest frequency to be sampled is 20KHz so the CD Sample rate is 44.1KHz. That is 4 KHz greater than two times, fulfilling the Nyquist criteria. Assuming a 16 bit resolution, the data rate of this now digital sampled signal is 16 * 44100 * 2 or 1.42 megabits(Mbps) per second for 2 channels (705kbps per channel at 44.1KHz sample rate).
IBOC uses a similar sample rate 48KHz, so they also meet the Nyquist criteria here. But wait, the next step in the IBOC processing completely destroys the accuracy of the original analog signal. They use a throw away data compression Adaptive Predictive Encoder to digitally compress the data rate.
These devices ignore the Nyquist theorem completely and make real time guesses at what part of the data need be transmitted and what part is "redundant". Their data rate instead of 1.42Mbps for 16 bit resolution, is 99Kbps for the FM version and 66 Kbps for the AM version. This is a compression ratio of 15 to 1 or greater. This means for every bit transmitted, 15 are thrown away. The purpose of this is to reduce the transmit bandwidth of the digital signal and make it fit in piggyback with the current analog signals in the AM and FM bands. Definitely not Nyquist.
This digital approach works fine for cellphones and two way radio where only voice information is transmitted, but not for high fidelity sound reproduction. Music contains so many harmonics and Fourier products that must be sampled, transmitted and received, it is impossible to use lossy compression techniques without damaging the content.
What about lossless compression. Lossless compression does something different. It does not reduce the bandwidth because in music somethings (Brass, Strings, Percussion) cannot be compressed and require full bandwidth to be transmitted and received. Those are sent uncompressed. Things that can be losslessly compressed, are. The result is a reduction in the total number of bits to be sent making the total size of a given content smaller. This is why the maximum compression ratio is about 1.5 to1.
To even consider multichannel high fidelity with IBOC is ludicrous.
I have only discussed the fidelity damage of IBOC radio. The deleterious effects on both the current AM and FM signals are quite prevalent.
If digital radio is to be, it must be a great improvement over todays analog technologies. IBOC is not even in the same ball park. When you see or hear advertisements such as "CD quality sound" for IBOC, you now know it can not possibly be. SNAKE OIL SALES AND VAPORWARE
Lou Dorren
Addressing the QSI5022, the performance of the CD-4 demodulator built with these chips was far superior to any of the CD-4 demodulators out there at that time. The reason mostly is because the manufactures all heard through the grapevine about my Phase Lock Loop technique of demodulation. They all figured that if they used a Signetics NE565 PLL chip as the FM detector, this would solve the breakup noise and limited number of plays of CD-4 records. Technics, Panasonic, Harmon Kardon, and Heath Kit were the only OEMs to employ the QSI5022.
The secret to high performance in the CD-4 demodulator was to put a high gain FM limiter before the PLL. I have read some of the other threads on this site, where CD-4 users are making precision tangential adjustments to the pickup cartridge to reduce the 30 KHz subcarrier to subcarrier crosstalk. This is only needed because of the FM processing system in the demodulator.
FM has a unique characteristic known as capture ratio. This is the ability of an FM detector to detect and demodulate the stronger of 2 FM signals on the same frequency. The typical CD-4 demod has a capture ratio of 7 to 12 dB. This means that the stronger signal has to be at least 7 to 12 dB greater than the weak signal to reject the interference from that weak signal.
The QSI5022 has a capture ratio of .75 dB. This ratio allows the demodulator to work well even in non optimum setups. It also had the circuitry to allow external subcarrier crosstalk cancellation for pickup cartridges with poor to medium left-right separation in the subcarrier band (15KHz to 45KHz).
I kept a couple of the FM demodulators and I have two complete Quadraplex broadcast generators.
With regard to IBOC "HD" radio. This is one of those non technology systems I hinted about in a previous posting. This system is as far as you can get from "High Definition" (HD actually stands for hybrid digital). These digital signals are piggybacked onto the current AM and FM analog signals.
First let me introduce the Nyquist theorem. It states "to sample an analog signal and accurately transmit and reproduce that signal, the sampling rate must be at least two times the highest analog frequency to be sampled".
Lets take the case of the Compact Disc. The highest frequency to be sampled is 20KHz so the CD Sample rate is 44.1KHz. That is 4 KHz greater than two times, fulfilling the Nyquist criteria. Assuming a 16 bit resolution, the data rate of this now digital sampled signal is 16 * 44100 * 2 or 1.42 megabits(Mbps) per second for 2 channels (705kbps per channel at 44.1KHz sample rate).
IBOC uses a similar sample rate 48KHz, so they also meet the Nyquist criteria here. But wait, the next step in the IBOC processing completely destroys the accuracy of the original analog signal. They use a throw away data compression Adaptive Predictive Encoder to digitally compress the data rate.
These devices ignore the Nyquist theorem completely and make real time guesses at what part of the data need be transmitted and what part is "redundant". Their data rate instead of 1.42Mbps for 16 bit resolution, is 99Kbps for the FM version and 66 Kbps for the AM version. This is a compression ratio of 15 to 1 or greater. This means for every bit transmitted, 15 are thrown away. The purpose of this is to reduce the transmit bandwidth of the digital signal and make it fit in piggyback with the current analog signals in the AM and FM bands. Definitely not Nyquist.
This digital approach works fine for cellphones and two way radio where only voice information is transmitted, but not for high fidelity sound reproduction. Music contains so many harmonics and Fourier products that must be sampled, transmitted and received, it is impossible to use lossy compression techniques without damaging the content.
What about lossless compression. Lossless compression does something different. It does not reduce the bandwidth because in music somethings (Brass, Strings, Percussion) cannot be compressed and require full bandwidth to be transmitted and received. Those are sent uncompressed. Things that can be losslessly compressed, are. The result is a reduction in the total number of bits to be sent making the total size of a given content smaller. This is why the maximum compression ratio is about 1.5 to1.
To even consider multichannel high fidelity with IBOC is ludicrous.
I have only discussed the fidelity damage of IBOC radio. The deleterious effects on both the current AM and FM signals are quite prevalent.
If digital radio is to be, it must be a great improvement over todays analog technologies. IBOC is not even in the same ball park. When you see or hear advertisements such as "CD quality sound" for IBOC, you now know it can not possibly be. SNAKE OIL SALES AND VAPORWARE
Lou Dorren
Last edited: