AAC encoder: Extensive improvements

This finalizes merging of the work in the patches in ticket #2686.

Improvements to twoloop and RC logic are extensive.

The non-exhaustive list of twoloop improvments includes:
 - Tweaks to distortion limits on the RD optimization phase of twoloop
 - Deeper search in twoloop
 - PNS information marking to let twoloop decide when to use it
   (turned out having the decision made separately wasn't working)
 - Tonal band detection and priorization
 - Better band energy conservation rules
 - Strict hole avoidance

For rate control:
 - Use psymodel's bit allocation to allow proper use of the bit
   reservoir. Don't work against the bit reservoir by moving lambda
   in the opposite direction when psymodel decides to allocate more/less
   bits to a frame.
 - Retry the encode if the effective rate lies outside a reasonable
   margin of psymodel's allocation or the selected ABR.
 - Log average lambda at the end. Useful info for everyone, but especially
   for tuning of the various encoder constants that relate to lambda
   feedback.

Psy:
 - Do not apply lowpass with a FIR filter, instead just let the coder
   zero bands above the cutoff. The FIR filter induces group delay,
   and while zeroing bands causes ripple, it's lost in the quantization
   noise.
 - Experimental VBR bit allocation code
 - Tweak automatic lowpass filter threshold to maximize audio bandwidth
   at all bitrates while still providing acceptable, stable quality.

I/S:
 - Phase decision fixes. Unrelated to #2686, but the bugs only surfaced
   when the merge was finalized. Measure I/S band energy accounting for
   phase, and prevent I/S and M/S from being applied both.

PNS:
 - Avoid marking short bands with PNS when they're part of a window
   group in which there's a large variation of energy from one window
   to the next. PNS can't preserve those and the effect is extremely
   noticeable.

M/S:
 - Implement BMLD protection similar to the specified in
   ISO-IEC/13818:7-2003, Appendix C Section 6.1. Since M/S decision
   doesn't conform to section 6.1, a different method had to be
   implemented, but should provide equivalent protection.
 - Move the decision logic closer to the method specified in
   ISO-IEC/13818:7-2003, Appendix C Section 6.1. Specifically,
   make sure M/S needs less bits than dual stereo.
 - Don't apply M/S in bands that are using I/S

Now, this of course needed adjustments in the compare targets and
fuzz factors of the AAC encoder's fate tests, but if wondering why
the targets go up (more distortion), consider the previous coder
was using too many bits on LF content (far more than required by
psy), and thus those signals will now be more distorted, not less.

The extra distortion isn't audible though, I carried extensive
ABX testing to make sure.

A very similar patch was also extensively tested by Kamendo2 in
the context of #2686.

1 files changed, 56 insertions, 0 deletions

diff --git a/libavcodec/aacenc_utils.h b/libavcodec/aacenc_utils.h
index dbc9554379..b2ce22186b 100644
--- a/libavcodec/aacenc_utils.h
+++ b/libavcodec/aacenc_utils.h
@@ -96,6 +96,54 @@ static inline int find_min_book(float maxval, int sf)
     return cb;
 }
 
+static float find_form_factor(int group_len, int swb_size, float thresh, const float *scaled, float nzslope) {
+    const float iswb_size = 1.0f / swb_size;
+    const float iswb_sizem1 = 1.0f / (swb_size - 1);
+    const float ethresh = thresh;
+    float form = 0.0f, weight = 0.0f;
+    int w2, i;
+    for (w2 = 0; w2 < group_len; w2++) {
+        float e = 0.0f, e2 = 0.0f, var = 0.0f, maxval = 0.0f;
+        float nzl = 0;
+        for (i = 0; i < swb_size; i++) {
+            float s = fabsf(scaled[w2*128+i]);
+            maxval = FFMAX(maxval, s);
+            e += s;
+            e2 += s *= s;
+            /* We really don't want a hard non-zero-line count, since
+             * even below-threshold lines do add up towards band spectral power.
+             * So, fall steeply towards zero, but smoothly
+             */
+            if (s >= ethresh) {
+                nzl += 1.0f;
+            } else {
+                nzl += powf(s / ethresh, nzslope);
+            }
+        }
+        if (e2 > thresh) {
+            float frm;
+            e *= iswb_size;
+
+            /** compute variance */
+            for (i = 0; i < swb_size; i++) {
+                float d = fabsf(scaled[w2*128+i]) - e;
+                var += d*d;
+            }
+            var = sqrtf(var * iswb_sizem1);
+
+            e2 *= iswb_size;
+            frm = e / FFMIN(e+4*var,maxval);
+            form += e2 * sqrtf(frm) / FFMAX(0.5f,nzl);
+            weight += e2;
+        }
+    }
+    if (weight > 0) {
+        return form / weight;
+    } else {
+        return 1.0f;
+    }
+}
+
 /** Return the minimum scalefactor where the quantized coef does not clip. */
 static inline uint8_t coef2minsf(float coef)
 {
@@ -125,6 +173,14 @@ static inline int quant_array_idx(const float val, const float *arr, const int n
     return index;
 }
 
+/**
+ * approximates exp10f(-3.0f*(0.5f + 0.5f * cosf(FFMIN(b,15.5f) / 15.5f)))
+ */
+static av_always_inline float bval2bmax(float b)
+{
+    return 0.001f + 0.0035f * (b*b*b) / (15.5f*15.5f*15.5f);
+}
+
 /*
  * linear congruential pseudorandom number generator, copied from the decoder
  */