GPSYCHO - Variable Bit Rate
Suggested usage:
lame -v -V 2 -b 128 input.wav output.mp3
VBR mode automatically uses the highest quality option. So both"-v" and "-h" are not necessary when using -V. Options:
- -V n (where n=0..9)
- 0 = highest quality
9 = lowest quality - -b <minimum allowed bitrate>
- -B <maximum allowed bitrate>
Using -B with other than 320kbs is not recommended, since even a128kbs CBR stream will sometimes use frames as large as 320kbsvia the bitreservoir.
Variables used in VBR code description:
sfb- Scale factor band index.
thm[sfb]- Allowed masking.
thm[sfb] = How much noise is allowed in thesfb'th band, as computed by the GPSYCHO. gain[sfb]- MDCT coefficents are scaled by 2^(-.25*gain) before quantizing. Smallervalues of gain (more negative) mean that more bits are required to encode thecoefficients, but the quantization noise will be (usually) smaller.
desired_gain[sfb]- The amount of gain needed so that if gain[sfb] <= desired_gain[sfb],the quantization noise will be <= thm[sfb].
An MP3 can use the following variables to achieve a given gain[sfb]. Forlongblocks:
gain[sfb][i] = 2^ [ -.25 * ( global_gain -210 - ifqstep*scalefac[gr][ch].l[sfb] - ifqstep*pretab[sfb]) ]
For shortblocks (i=0..2 for the three short blocks):
gain[sfb][i] = 2^ [ -.25*( global_gain -210 - 8*subblock_gain[i] - ifqstep*scalefac.s[sfb][i]) ]
In both of the above cases, calculateifqstep:
ifqstep = scalefac_scale==0 ? 2 : 4;
Algorithm
The VBR algorithm is as follows.
- Step 1:
psymodel.c - Computes the allowed maskings, thm[sfb] thm[sfb] may be reduced by a fewdb depending on the quality setting. The smaller thm[sfb], the more bits willbe required to encode the frame.
- Step 2:
find_scalefac() invbrquantize.c - Compute
desired_gain[sfb] by iterating over the values ofsfb from 0 toSBMAX. At each value, computedesired_gain[sfb] using a divide and conquer iteration so thatquantization_noise[sfb] < thm[sfb] . This requires 7iterations ofcalc_sfb_noise() which computes quantization errorfor the specified gain. This is the only time VBR needs to do any (expensive)quantization! - Step 3:
VBR_noise_shaping() in vbrquantize.c - Find a combination of global_gain, subblock_gain, preflag, scalefac_scale,etc... so that:
gain[sfb] <= desired_gain[sfb] - Step 4:
VBR_quantize_granule() invbrquantize.c - Calculate the number of bits needed to encode the frame withthe values computed in step 3. Unlike CBR, VBR (usually) onlyhas to do this expensive huffman bit counting stuff once!
- Step 5:
VBR_noise_shaping() invbrquantize.c - if bits < minimum_bits: Repeat step 3, only with a larger value ofglobal_gain. (but allow bits < minimum_bits for analog silence)
if bits > maximum_bits: decrease global_gain, keeping all other scalefactorsthe same.
Usually step 5 is not necessary. - step 6: VBR_quantize() in vbrquantize.c
- After encoding both channels and granules, check to make sure that thetotal number of bits for the whole frame does not exceed the maximum allowed.If it does, lower the quality and repeat steps 2,3 and 4 for the granules thatwere using lots of bits.
Flow
The actual flow chart looks something like this:
- VBR_quantize()
- determine minbits, maxbits for each granuledetermine max_frame_bitsadjust global quality setting based on VBR_qdo frame_bits=0 loop over each channel, granule: compute thm[sfb] bits = VBR_noise_shaping(): Encodes each granule with minbits <= bits <= maxbits frame_bits += bits lower the global quality settingwhile (frame_bits > max_frame_bits)
- VBR_noise_shaping()
- find_scalefac() (computes desired_gain)Estimate largest possible value of global_gaindo compute_scalefac_long/short() scalefacts, etc. so that gain <= desired_gain) bits = VBR_quantize_granule() if (bits < minbits && analog silence) break; if (bits >= minbits) break; decrease global_gain (which increases number of bits used)while 1if bits > maxbits do increase global_gain bits = VBR_quantize_granule() while (bits > maxbits)
find_scalefac()- Simple divide and conquer iteration which repeatidly calls calc_sfb_noise() with different values of desired_gain until it finds the largest desired_gain such that the quantization_noise < allowed masking Requires 7 iterations.
