[Mp4-tech] Mpeg4 AAC +Linear interpolation in inverse quantization

[Shreya Pathak]

An HTML attachment was scrubbed...
URL: /pipermail/mp4-tech/attachments/20040903/b821785f/attachment.html
-------------- next part --------------
Hi,
1)   I wanted to know about linear interpolation used in the public source code FAAD for Mpeg4 AAC.
    x1 = invquant_table[q>>3];
    x2 = invquant_table[(q>>3) + 1];
    return sgn * 16 * ((errcorr[q&7] * (x2-x1)) + x1);
I didn't get why it is multiplied by 16.
    I did inverse quantization using 5th order polynomial.But the problem was heavy computations due to 5th order polynomial.
2) I wanted to know about how fixed point square root approximation has been dine in this code.
#define step(shift) \
    if ((0x40000000l >> shift) + root <= value)       \
    {                                                 \
        value -= (0x40000000l >> shift) + root;       \
        root = (root >> 1) | (0x40000000l >> shift);  \
    } else {                                          \
        root = root >> 1;                             \
    }
/* fixed point square root approximation */
/* !!!! ONLY WORKS FOR EVEN %REAL_BITS% !!!! */
int fixed_sqrt(int value)
{
    int root = 0;
    step( 0); step( 2); step( 4); step( 6);
    step( 8); step(10); step(12); step(14);
    step(16); step(18); step(20); step(22);
    step(24); step(26); step(28); step(30);
    if (root < value)
        ++root;
    root <<= (14/2);
    return root;
}    
Thanks and Regards
Shreya