VHD disk related VBI data extraction from VBI lines 17(280), 18 (281), and 20 ( 283).

[popolonx]

Current ld-process-vbi does not support VHD disk specific data extraction from VBI lines.
VHD specific data is encoded in VHD disks in VBI lines 17(280), 18 (281), and 20 ( 283).
Data is biphase mark encoded, and there is 29 bits. One bit (0 or 1) length is 12/(455fH), fH=15,734.26 Hz.
First bit (MSB) starts 412/(455*fH) from horizontal synch pulse trailing edge.

VHD disk and those bit encoding are described in the book https://archive.org/details/how-to-enjoy-vhd-pages-14-53 ,page 27. Also same page 27 as a separate picture here:
https://github.com/popolonx/MSX_STUFF/blob/main/vhd/vhd_disk_vbi_lines_encoding.png

[popolonx]

Basically this kind of function will decode the data from one line (need to call for lines 17,18,20, and maybe also for 21):

// Private method to read a 28-bit biphase coded signal (manchester code) from a field `
qint32 BiphaseCode::biphaseManchesterDecoder(
    const SourceVideo::Data &lineData, qint32 zcPoint,
    LdDecodeMetaData::VideoParameters videoParameters)
{
    qint32 result = 0;
    QVector<bool> manchesterData = getTransitionMap(lineData, zcPoint);

    // Get the number of samples for 2/3 of 1.68us
    double fJumpSamples = (videoParameters.sampleRate / 1000000) * 1.68 * 0.66;
    qint32 bmcBitDuration = static_cast<qint32>(fJumpSamples);
    // Keep track of the number of bits decoded
    qint32 decodeCount = 0;

    // Find the first transition
    qint32 x = videoParameters.activeVideoStart;

    while (x < manchesterData.size() && manchesterData[x] == false) {
        x++;
    }
    bool prev_value = true;
    qint32 prev_x = x; 
    int index = 0;

    while (x < manchesterData.size())
    {
        if (manchesterData[x] != prev_value)
        {
            prev_value = manchesterData[x];
            if ( ( x - prev_x ) > bmcBitDuration )
            {
                prev_x = x;
                result = (result << 1); // 0 bit
                index = 0;
                decodeCount++;
            }
            else
            {
                prev_x = x;
                if ( index == 1)
                {
                    result = (result << 1) + 1; // 1 bit
                    index = 0;
                    decodeCount++;
                }
                else
                {
                    // 1-bit half bit
                    index = 1;
                }
            }


        }
        else
        {
            x++;
        }
    }

    // We must have 28-bits if the decode was successful
    if (decodeCount <28) {

        result = 0;
        if (decodeCount != 0) {
            qDebug() << "BMC decode failed!  Got" << decodeCount << "bits, expected 28";
			// -1 is a good choice to indicate a parse error because it does not conflict with 0 (black line) or >0 (successfully parsed data)
			// differentiating between parse errors and black lines is useful because if parse errors are known, they can be autofixed by studying the surrounding picture number cadence
			result = -1;
        }
    }
    else
    {
        qDebug() << "BMC value 0b" << QString::number(result, 2);
    }
    
    return result;
}

Did not yet figure out how this can be smootly added to existing functions.