| #!/usr/bin/env python |
| ## Copyright (c) 2019, Alliance for Open Media. All rights reserved |
| ## |
| ## This source code is subject to the terms of the BSD 2 Clause License and |
| ## the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License |
| ## was not distributed with this source code in the LICENSE file, you can |
| ## obtain it at www.aomedia.org/license/software. If the Alliance for Open |
| ## Media Patent License 1.0 was not distributed with this source code in the |
| ## PATENTS file, you can obtain it at www.aomedia.org/license/patent. |
| ## |
| __author__ = "maggie.sun@intel.com, ryan.lei@intel.com" |
| |
| import numpy as np |
| import math |
| import scipy.interpolate |
| import logging |
| from Config import LoggerName |
| from operator import itemgetter |
| |
| subloggername = "CalcBDRate" |
| loggername = LoggerName + '.' + '%s' % subloggername |
| logger = logging.getLogger(loggername) |
| |
| |
| def Interpolate(qp, logbr, qty): |
| ''' |
| generate interpolated RD sampling point based on QP values. |
| using bilinear interpolation. bit rate are in the log domain |
| incoming logbr and qty should be in the increasing order |
| incoming qp could be in any order |
| ''' |
| int_qp = []; int_logbr = []; int_qty = [] |
| for i in range(0, len(qp)): |
| int_qp.append(qp[i]) |
| int_logbr.append(logbr[i]) |
| int_qty.append(qty[i]) |
| #handle duplicated qp |
| if (i<len(qp)-1) and (qp[i] == qp[i+1]): |
| continue |
| #interpolate logbr and quality point between qp[i] and qp[i+1] |
| #qp order is non-determinstics, because it could come from different resolution |
| if (i<len(qp)-1) and ((qp[i] != qp[i + 1] + 1) or (qp[i] != qp[i + 1] - 1)): |
| qlist = [] |
| if (qp[i+1] > qp[i]): |
| qlist = range(qp[i] + 1, qp[i+1], 1) |
| else: |
| qlist = range(qp[i] - 1, qp[i+1], -1) |
| for q in qlist: |
| # bitrate(qp_target) = (bitrate(qp0) - bitrate(qp1)) / (qp0 - qp1) * (qp_target - qp0) + bitrate(qp0) |
| # quality(qp_target) = (quality(qp0) - quality(qp1)) / (qp0 - qp1) * (qp_target - qp0) + quality(qp0) |
| # result int_logbr and int_qty will always be in non-decreasing order. |
| # but order of int_qp is similar as input qp |
| br = (logbr[i] - logbr[i + 1]) / (qp[i] - qp[i + 1]) * (q - qp[i]) + logbr[i] |
| qt = (qty[i] - qty[i + 1]) / (qp[i] - qp[i + 1]) * (q - qp[i]) + qty[i] |
| int_qp.append(q) |
| int_logbr.append(br) |
| int_qty.append(qt) |
| ''' |
| print("before interpolation:") |
| for i in range(len(qp)): |
| print("%d, %f, %f"%(qp[i], logbr[i], qty[i])) |
| print("after interpolation:") |
| for i in range(len(int_qp)): |
| print("%d, %f, %f"%(int_qp[i], int_logbr[i], int_qty[i])) |
| ''' |
| return int_qp, int_logbr, int_qty |
| |
| # BJONTEGAARD Bjontegaard metric |
| # Calculation is adapted from Google implementation |
| # PCHIP method - Piecewise Cubic Hermite Interpolating Polynomial interpolation |
| def BD_RATE(br1, qtyMtrc1, br2, qtyMtrc2): |
| brqtypairs1 = []; brqtypairs2 = [] |
| for i in range(min(len(qtyMtrc1), len(br1))): |
| if (br1[i] != '' and qtyMtrc1[i] != ''): |
| brqtypairs1.append((br1[i], qtyMtrc1[i])) |
| for i in range(min(len(qtyMtrc2), len(br2))): |
| if (br2[i] != '' and qtyMtrc2[i] != ''): |
| brqtypairs2.append((br2[i], qtyMtrc2[i])) |
| |
| # sort the pair based on quality metric values in increasing order |
| # if quality metric values are the same, then sort the bit rate in increasing order |
| brqtypairs1.sort(key = itemgetter(1, 0)) |
| brqtypairs2.sort(key = itemgetter(1, 0)) |
| |
| logbr1 = [math.log(x[0]) for x in brqtypairs1] |
| qmetrics1 = [100.0 if x[1] == float('inf') else x[1] for x in brqtypairs1] |
| logbr2 = [math.log(x[0]) for x in brqtypairs2] |
| qmetrics2 = [100.0 if x[1] == float('inf') else x[1] for x in brqtypairs2] |
| |
| if not brqtypairs1 or not brqtypairs2: |
| logger.info("one of input lists is empty!") |
| return 0.0 |
| |
| # remove duplicated quality metric value, the RD point with higher bit rate is removed |
| dup_idx = [i for i in range(1, len(qmetrics1)) if qmetrics1[i - 1] == qmetrics1[i]] |
| for idx in sorted(dup_idx, reverse=True): |
| del qmetrics1[idx] |
| del logbr1[idx] |
| dup_idx = [i for i in range(1, len(qmetrics2)) if qmetrics2[i - 1] == qmetrics2[i]] |
| for idx in sorted(dup_idx, reverse=True): |
| del qmetrics2[idx] |
| del logbr2[idx] |
| |
| # find max and min of quality metrics |
| min_int = max(min(qmetrics1), min(qmetrics2)) |
| max_int = min(max(qmetrics1), max(qmetrics2)) |
| if min_int >= max_int: |
| logger.info("no overlap from input 2 lists of quality metrics!") |
| return 0.0 |
| |
| # generate samples between max and min of quality metrics |
| lin = np.linspace(min_int, max_int, num=100, retstep=True) |
| interval = lin[1] |
| samples = lin[0] |
| |
| # interpolation |
| v1 = scipy.interpolate.pchip_interpolate(qmetrics1, logbr1, samples) |
| v2 = scipy.interpolate.pchip_interpolate(qmetrics2, logbr2, samples) |
| |
| # Calculate the integral using the trapezoid method on the samples. |
| int1 = np.trapz(v1, dx=interval) |
| int2 = np.trapz(v2, dx=interval) |
| |
| # find avg diff |
| avg_exp_diff = (int2 - int1) / (max_int - min_int) |
| avg_diff = (math.exp(avg_exp_diff) - 1) * 100 |
| |
| return avg_diff |
| |
| ''' |
| if __name__ == "__main__": |
| brs1 = [64052.664, 6468.096, 4673.424, 3179.4, 2298.384, 1361.184] |
| qtys1 = [1, 1, 0.99999, 0.99998, 0.99996, 0.99992] |
| brs2 = [68461.896, 7554.96, 4827.432, 3294.024, 2380.128, 1401.744] |
| qtys2 = [1, 1, 0.99999, 0.99998, 0.99996, 0.99992] |
| |
| bdrate = BD_RATE(brs1, qtys1, brs2, qtys2) |
| print("bdrate calculated is %3.3f%%" % bdrate) |
| ''' |
