FCC Proposes Authorizing Voluntary All-Digital AM Broadcasting!

Discussion in 'Amplitude Modulation' started by NA5B, Nov 25, 2019.

ad: L-HROutlet
ad: l-rl
ad: Left-2
ad: L-MFJ
ad: MessiPaoloni-1
ad: Subscribe
ad: Left-3
  1. N1BCG

    N1BCG Ham Member QRZ Page

    This is clearly not the thread to read before morning coffee...
     
  2. KA9Q

    KA9Q Ham Member QRZ Page

    Yes, that's the most popular format. But it's not the only one.
    Seems straightforward enough. Every possible combination of the 9 distinct amplitudes and 52 distinct phases of the rectangular 64-QAM constellation would yield 468 different states. But only 64 are valid.

    Optimum QAM constellation design is not as obvious as it may first seem. Rectangular is certainly widely used and easy to implement, but the best choice depends on an understanding of the channel model, the particular trellis FEC code in use, and even the properties of the transmitter hardware (e.g., nonlinearities).
     
  3. AC0OB

    AC0OB Subscriber QRZ Page

    Now this is not an indictment of your knowledge of digital communications or experience at Qualcom, but I have to assume you don't teach nor have taught in a university environment?

    In teaching undergrad and graduate student you have to ease them into the subject one step at a time only after they have had basic analog and digital courses. Analog and digital communications is usually taught after a clear understanding of the basics. Digital communications systems design and development subjects such as using convolutional encoding/decoding, OFDM mapping, scrambling, interleaving and systems control processing and the such like are usually taught in graduate level classes.


    However, there is one point of mutual discussion regarding the Ibiquty HD document SY_IDD_1012s, entitled HD Radio Air Interface Design Description Layer 1 AM.

    In Sections 6.4 and 11.2.4 there is an Analog Audio Bandwidth Indicator bit (AABI) that is defaulted to 0 for the All-Digital MA3 service mode of operation that limits the analog audio bandpass to 5 kHz.

    Now taken that together with Section 4.4, Logical Channels and Table 4.1, it appears we will have a limited 5kz audio bandpass, with constellation mapping to OFDM modulation subcarriers, a total bit-rate of 44.4 kbits/s, occupying an idealized channel bandwidth of 20kHz.

    For the All-Digital AM system proposed it appears we are occupying 20 kHz of RF bandwidth and getting 5 kHz audio.

    In contrast, today's analog AM can provide a 10 kHz audio bandpass at an RF bandwidth of 20khz. [To be exact, and according to previous NRSC literature, it is more like a 9.7 kHz audio bandpass with a 19.4 kHz RF bandwidth].


    Pheel





     
  4. W6RZ

    W6RZ Premium Subscriber QRZ Page

    The AABI bit is indicating which waveform (either figure 5.1 or figure 5.2) is being transmitted, not the frequency response of the digital audio. The typical frequency response of the HDC-SBR digital audio is 12 kHz. Your argument is invalid.
     
  5. AC0OB

    AC0OB Subscriber QRZ Page

    1) I am not making an argument. I am making an assessment of the data presented in the IBOC/HD documentation as written so it is not invalid. The documentation needs further elaboration and clarification with respect to the bandwidth of the analog signal that will be accepted in the IBOC exciter.


    2) "The typical frequency response of the HDC-SBR digital audio is 12 kHz."

    Figures 5.4 and 5.5 of Page 19 only tell me this: With the primary sidebands a digital bandwidth of ~ = 9.811.0 kHz will be transmitted. When the Primary AND Tertiary sidebands are transmitted the digital bandwidth will be ~ = 18.8954 kHz.

    So here is the main question: if I output an analog signal from my processing unit to the IBOC exciter, and this analog signal contains an audio passband of 15 kHz, then will the analog audio coming out of my IBOC decoder at the receiving end contain an audio passband of 12 kHz which contains frequency components up to 12 kHz???

    3) From IBOC/HD document SY_IDD_1012s, December 14th, 2016.
    "11.2.4 Analog Audio Bandwidth Indicator (AABI)
    The analog audio bandwidth indicator (AABI) is a one-bit flag used to indicate the maximum bandwidth of the analog audio signal when transmitting a Hybrid waveform. If the flag is cleared [AABI = bit 0], the maximum analog audio bandwidth is 5 kHz; if the flag is set [AABI = 1] , the maximum analog audio bandwidth is 8 kHz. The AABI flag is always 0 when an All Digital waveform is being transmitted."

    So one has to assume from the text that if the AABI flag is always 0 for the All-Digital waveform, then this implies the analog bandwidth that is allowed to be converted to digital is limited to 5 kHz.

    If Xperi is allowing an analog audio bandpass greater than 5kHz to be converted to digital when the AABI flag is set to 0, then shouldn't the documentation EXPLICITLY state that??

    Pheel
     
    Last edited: Dec 17, 2019
  6. AC0OB

    AC0OB Subscriber QRZ Page

    You are looking at the wrong waveforms. Those are for the Hybrid signals, Pages 14-18

    We're talking about this document: https://www.nrscstandards.org/stand...s/standards/nrsc-5-d/reference-docs/1012s.pdf

    The discussion here is focused on Page 52 - Section 11.2.4.

    The All-Digital signals in Figures 5.4 and 5.5 of Page 19 and tells us nothing about the audio bandpass that those digital waveforms represent.


    Pheel
     
    Last edited: Dec 17, 2019
  7. AC0OB

    AC0OB Subscriber QRZ Page

    1) I never said it was the only one.

    In the handout I posted I referenced a number of papers and texts. Are you familiar with any of them?

    2) Can you add anything to the questions I posed below?

    "However, there is one point of mutual discussion regarding the Ibiquty HD document SY_IDD_1012s, entitled HD Radio Air Interface Design Description Layer 1 AM.

    In Sections 6.4 and 11.2.4 there is an Analog Audio Bandwidth Indicator bit (AABI) that is defaulted to 0 for the All-Digital MA3 service mode of operation that limits the analog audio bandpass to 5 kHz.

    Now taken that together with Section 4.4, Logical Channels and Table 4.1, it appears we will have a limited 5kz audio bandpass, with constellation mapping to OFDM modulation subcarriers, a total bit-rate of 44.4 kbits/s, occupying an idealized channel bandwidth of 20kHz.

    For the All-Digital AM system proposed it appears we are occupying 20 kHz of RF bandwidth and getting 5 kHz audio.

    In contrast, today's analog AM can provide a 10 kHz audio bandpass at an RF bandwidth of 20khz. [To be exact, and according to previous NRSC literature, it is more like a 9.7 kHz audio bandpass with a 19.4 kHz RF bandwidth]."


    "So here is the main question: if I output an analog signal from my processing unit to the IBOC exciter, and this analog signal contains an audio passband of 15 kHz, then will the analog audio coming out of my IBOC decoder at the receiving end contain an audio passband of 12 kHz which contains frequency components up to 12 kHz???

    3) From IBOC/HD document SY_IDD_1012s, December 14th, 2016.
    "11.2.4 Analog Audio Bandwidth Indicator (AABI)
    The analog audio bandwidth indicator (AABI) is a one-bit flag used to indicate the maximum bandwidth of the analog audio signal when transmitting a Hybrid waveform. If the flag is cleared [AABI = bit 0], the maximum analog audio bandwidth is 5 kHz; if the flag is set [AABI = 1] , the maximum analog audio bandwidth is 8 kHz. The AABI flag is always 0 when an All Digital waveform is being transmitted."

    So one has to assume from the text that if the AABI flag is always 0 for the All-Digital waveform, then this implies the analog bandwidth that is allowed to be converted to digital is limited to 5 kHz.

    If Xperi is allowing an analog audio bandpass greater than 5kHz to be converted to digital when the AABI flag is set to 0, then shouldn't the documentation EXPLICITLY state that??"


    Pheel
     
  8. W6RZ

    W6RZ Premium Subscriber QRZ Page

    I'm looking at the correct waveforms. Maybe a truth table will help you.
    Code:
     PSM        AAB        RDB            Waveform
    00001        0          0            Figure 5.1
    00001        1          0            Figure 5.2
    00001        X          1            Figure 5.3
    00010        X          0            Figure 5.4
    00010        X          1            Figure 5.5
    
    X means "don't care" as specified in paragraph 6.4.

    In the standard configuration (RDB = 0) of the hybrid waveform, the bandwidth of the analog audio is one of two selectable values: 5 kHz or 8 kHz. Analog audio bandwidth control (AAB) specifies which bandwidth is to be employed. When AAB is a logical 0, the bandwidth indicated is 5 kHz. When AAB is a logical 1, the bandwidth indicated is 8 kHz. When transmitting the All Digital waveform, it is ignored. Analog audio bandwidth control is received from the Configuration Administrator at the L1 frame rate, Rf, and any change can be effected directly (at an L1 frame boundary) upon receipt without interrupting service. Digital coverage of a Hybrid station (primary subcarriers only) is adversely impacted by a second adjacent Hybrid transmission with 8kHz audio bandwidth.

    In the reduced digital bandwidth configuration (RDB = 1) of the hybrid waveform, the state of AAB is not applicable and shall be set to a value of zero.


    Whenever you see "analog audio bandwidth" or "bandwidth of the analog audio" in that document, they're talking about the bandwidth of the legacy AM signal.

    The digital audio frequency response is not defined in NRSC-5. See paragraph 3.4.1 of https://www.nrscstandards.org/standards-and-guidelines/documents/standards/nrsc-5-d/nrsc-5-d.pdf

    Source coding and compression of the main program service (MPS) and supplemental program service (SPS) audio must be performed before the audio information is fed into the audio transport subsystems. Each audio service (main program service and each individual supplemental program service) has its own source coding, compression and transport subsystem. NRSC-5 does not include specifications for audio source coding and compression. Suitable audio source coding and compression systems will use appropriate technologies (e.g., perceptual audio coding) to reduce the bit rate required for description of audio signals.
     
  9. AC0OB

    AC0OB Subscriber QRZ Page

    I know what the documents said. It is what they don't say that seems to cause questions to arise.

    Q1: Now, assuming I have the proper audio source coding and compression system (CODEC) of the appropriate technology, what is the highest analog audio frequency that I can both encode and decode with this All-Digital system within the specified 20 klHz bandwidth with reference to Figures 5.4 and 5.5 of Page 19?

    Q2: What CODEC system was used in the WWFD and the other eight testing stations?


    Pheel
     
  10. W6RZ

    W6RZ Premium Subscriber QRZ Page

    HD Radio uses a proprietary codec called HDC (High Definition Coding). But it was cracked in 2017 and found to be High Efficiency AAC version 1 with with just a few bits changed around to make it proprietary.

    https://github.com/theori-io/nrsc5/

    Since it's proprietary, nobody knows exactly how they're setting the low-pass filter. However, we can guess the parameters from HE-AAC. Here's a link to the Fraunhofer AAC implementation (which is considered best in class).

    http://wiki.hydrogenaud.io/index.php?title=Fraunhofer_FDK_AAC

    In the CBR bitrate table, the low-pass filter is set to 9640 Hz for 20 to 28 kbps, 13050 Hz for 28 to 40 kbps and 14260 Hz for 40 to 56 kbps.

    Figure 5.4 provides 40 kbps in 20 kHz of RF bandwidth and Figure 5.5 provides 20 kbps in 10 kHz of RF bandwidth. For a station sending the full 20 kHz of RF bandwidth, it's possible that the digital audio frequency response is 14.26 kHz. For a station sending the reduced bandwidth 10 kHz, it's possible that the digital audio frequency response is 9.64 kHz.

    After the HDC codec was cracked, my friend VE3IRR coded an FM NRSC-5 transmitter for GNU Radio.

    https://github.com/argilo/gr-nrsc5

    In this waterfall, he's modified the OFDM modulation to include some funny amplitude modulation along with the normal QPSK. It still decodes fine on an FM HD radio.

    [​IMG]
     
    Last edited: Dec 17, 2019
    KX4O likes this.

Share This Page