Timothy B. Terriberry authored 14 years ago and Jean-Marc Valin committed 14 years ago

9b34bd83 caused serious regressions for 240-sample frame stereo,
 because the previous qb limit was _always_ hit for two-phase
 stereo.
Two-phase stereo really does operate with a different model (for
 example, the single bit allocated to the side should really
 probably be thought of as a sign bit for qtheta, but we don't
 count it as part of qtheta's allocation).
The old code was equivalent to a separate two-phase offset of 12,
 however Greg Maxwell's testing demonstrates that 16 performs
 best.

Timothy B. Terriberry authored 14 years ago and Jean-Marc Valin committed 14 years ago

Previously, we would only split a band if it was allocated more than
 32 bits.
However, the N=4 codebook can only produce about 22.5 bits, and two
 N=2 bands combined can only produce 26 bits, including 8 bits for
 qtheta, so if we wait until we allocate 32, we're guaranteed to fall
 short.
Several of the larger bands come pretty far from filling 32 bits as
 well, though their split versions will.

Greg Maxwell also suggested adding an offset to the threshold to
 account for the inefficiency of using qtheta compared to another
 VQ dimension.
This patch uses 1 bit as a placeholder, as it's a clear
 improvement, but we may adjust this later after collecting data on
 more possibilities over more files.

The first version of the mono decoder with stereo output collapsed
 the historic energy values stored for anti-collapse down to one
 channel (by taking the max).
This means that a subsequent switch back would continue on using
 the the maximum of the two values instead of the original history,
 which would make anti-collapse produce louder noise (and
 potentially more pre-echo than otherwise).

This patch moves the max into the anti_collapse function itself,
 and does not store the values back into the source array, so the
 full stereo history is maintained if subsequent frames switch
 back.
It also fixes an encoder mismatch, which never took the max
 (assuming, apparently, that the output channel count would never
 change).

Instead of just dumping excess bits into the first band after
 allocation, use them to initialize the rebalancing loop in
 quant_all_bands().
This allows these bits to be redistributed over several bands, like
 normal.

Timothy B. Terriberry authored 14 years ago and Jean-Marc Valin committed 14 years ago

The average caps over all values of LM and C are well below the
 target allocations of the last two modelines.
Lower them to the caps, to prevent hitting them quite so early.
This helps quality at medium-high rates, in the 180-192 kbps range.

Timothy B. Terriberry authored 14 years ago and Jean-Marc Valin committed 14 years ago

Use measured cross-entropy to estimate the real cost of coding
 qtheta given the allocated qb parameter, instead of the entropy of
 the PDF.
This is generally much lower, and reduces waste at high rates.
This patch also removes some intermediate rounding from this
 computation.

Timothy B. Terriberry authored 14 years ago and Jean-Marc Valin committed 14 years ago

This extends the previous rebalancing for fine energy in N=1 bands
 to also allocate extra fine bits for bands that go over their cap.

Timothy B. Terriberry authored 14 years ago and Jean-Marc Valin committed 14 years ago

The previous "dumb cap" of (64<<LM)*(C<<BITRES) was not actually
 achievable by many (most) bands, and did not take the cost of
 coding theta for splits into account, and so was too small for some
 bands.
This patch adds code to compute a fairly accurate estimate of the
 real maximum per-band rate (an estimate only because of rounding
 effects and the fact that the bit usage for theta is variable),
 which is then truncated and stored in an 8-bit table in the mode.

This gives improved quality at all rates over 160 kbps/channel,
 prevents bits from being wasted all the way up to 255 kbps/channel
 (the maximum rate allowed, and approximately the maximum number of
 bits that can usefully be used regardless of the allocation), and
 prevents dynalloc and trim from producing enormous waste
 (eliminating the need for encoder logic to prevent this).

We use the MDCT as low-pass filter.

Timothy B. Terriberry authored 14 years ago and Jean-Marc Valin committed 14 years ago

Previously, in a stereo split with itheta==16384, but without
 enough bits left over to actually code a pulse, the target band
 would completely collapse, because the mid gain would be zero and
 we don't fold the side.
This changes the limit to ensure that we never set qn>1 unless we
 know we'll have enough bits for at least one pulse.
This should eliminate the last possible whole-band collapse.

The old constructor is renamed celt_encoder_create_custom(). Same
for the decoder.

Gregory Maxwell authored 14 years ago and Jean-Marc Valin committed 14 years ago

Prevent VBR from shooting up to the maximum rate if set to very low target rates, and prevent the encoder VBR from producing 1 byte frames (which are no longer allowed).

The energy memory can be lowered (not increased) during a transient

That way they can be exact in 16 bits once multiplied by the gain

We now encode the highest bitrate part of the split first and transfer
any unused bits to the other part. We use a dead zone of three bits
to prevent redistributing in cases of random fluctuation (or else
we will statistically lower the allocation of higher frequencies at
low-mid bitrates).

Makes celt_exp2() use Q10 input to avoid problems on very low energy.
Also makes the pitch downsampling more conservative on gain to avoid
problems later.

Also defining a 1-byte packet as triggering the PLC/CNG

Timothy B. Terriberry authored 14 years ago and Jean-Marc Valin committed 14 years ago

This changes folding so that the LCG is never used on transients
 (either short blocks or long blocks with increased time
 resolution), except in the case that there's not enough decoded
 spectrum to fold yet.

It also now only subtracts the anti-collapse bit from the total
 allocation in quant_all_bands() when space has actually been
 reserved for it.

Finally, it cleans up some of the fill and collapse_mask tracking
 (this tracking was originally made intentionally sloppy to save
 work, but then converted to replace the existing fill flag at the
 last minute, which can have a number of logical implications).
The changes, in particular:
1) Splits of less than a block now correctly mark the second half
    as filled only if the whole block was filled (previously it
    would also mark it filled if the next block was filled).
2) Splits of less than a block now correctly mark a block as
    un-collapsed if either half was un-collapsed, instead of marking
    the next block as un-collapsed when the high half was.
3) The N=2 stereo special case now keeps its fill mask even when
    itheta==16384; previously this would have gotten cleared,
    despite the fact that we fold into the side in this case.
4) The test against fill for folding now only considers the bits
    corresponding to the current set of blocks.
   Previously it would still fold if any later block was filled.
5) The collapse mask used for the LCG fold data is now correctly
    initialized when B=16 on platforms with a 16-bit int.
6) The high bits on a collapse mask are now cleared after the TF
    resolution changes and interleaving at level 0, instead of
    waiting until the very end.
   This prevents extraneous high flags set on mid from being mixed
    into the side flags for mid-side stereo.