lavc: remove unused traces of fmtconvert usage

Those decoders have been switched to float output and so do not use
fmtconvert anymore.

libavcodec/aac.h

@@ -36,7 +36,6 @@
 #include "fft.h"
 #include "mpeg4audio.h"
 #include "sbr.h"
-#include "fmtconvert.h"
 
 #include <stdint.h>
 
@@ -293,7 +292,6 @@ typedef struct AACContext {
     FFTContext mdct_ld;
     FFTContext mdct_ltp;
     IMDCT15Context *mdct480;
-    FmtConvertContext fmt_conv;
     AVFloatDSPContext fdsp;
     int random_state;
     /** @} */

libavcodec/aacdec.c

@@ -86,7 +86,6 @@
 #include "get_bits.h"
 #include "fft.h"
 #include "imdct15.h"
-#include "fmtconvert.h"
 #include "lpc.h"
 #include "kbdwin.h"
 #include "sinewin.h"
@@ -1043,7 +1042,6 @@ static av_cold int aac_decode_init(AVCodecContext *avctx)
 
     ff_aac_sbr_init();
 
-    ff_fmt_convert_init(&ac->fmt_conv, avctx);
     avpriv_float_dsp_init(&ac->fdsp, avctx->flags & CODEC_FLAG_BITEXACT);
 
     ac->random_state = 0x1f2e3d4c;

libavcodec/atrac3.c

@@ -41,7 +41,6 @@
 #include "avcodec.h"
 #include "bytestream.h"
 #include "fft.h"
-#include "fmtconvert.h"
 #include "get_bits.h"
 #include "internal.h"
 
@@ -107,7 +106,6 @@ typedef struct ATRAC3Context {
 
     AtracGCContext  gainc_ctx;
     FFTContext mdct_ctx;
-    FmtConvertContext fmt_conv;
     AVFloatDSPContext fdsp;
 } ATRAC3Context;
 
@@ -910,7 +908,6 @@ static av_cold int atrac3_decode_init(AVCodecContext *avctx)
 
     ff_atrac_init_gain_compensation(&q->gainc_ctx, 4, 3);
     avpriv_float_dsp_init(&q->fdsp, avctx->flags & CODEC_FLAG_BITEXACT);
-    ff_fmt_convert_init(&q->fmt_conv, avctx);
 
     q->units = av_mallocz(sizeof(*q->units) * avctx->channels);
     if (!q->units) {

libavcodec/binkaudio.c

@@ -34,7 +34,6 @@
 #include "get_bits.h"
 #include "dct.h"
 #include "rdft.h"
-#include "fmtconvert.h"
 #include "internal.h"
 #include "wma_freqs.h"
 #include "libavutil/intfloat.h"

libavcodec/nellymoserdec.c

@@ -37,7 +37,6 @@
 #include "libavutil/random_seed.h"
 #include "avcodec.h"
 #include "fft.h"
-#include "fmtconvert.h"
 #include "internal.h"
 #include "nellymoser.h"
 #include "sinewin.h"

libavcodec/vorbisdec.c

@@ -30,7 +30,6 @@
 #include "avcodec.h"
 #include "get_bits.h"
 #include "fft.h"
-#include "fmtconvert.h"
 #include "internal.h"
 
 #include "vorbis.h"
@@ -123,7 +122,6 @@ typedef struct vorbis_context_s {
     GetBitContext gb;
     VorbisDSPContext dsp;
     AVFloatDSPContext fdsp;
-    FmtConvertContext fmt_conv;
 
     FFTContext mdct[2];
     uint8_t       first_frame;
@@ -1018,7 +1016,6 @@ static av_cold int vorbis_decode_init(AVCodecContext *avctx)
     vc->avctx = avctx;
     ff_vorbisdsp_init(&vc->dsp);
     avpriv_float_dsp_init(&vc->fdsp, avctx->flags & CODEC_FLAG_BITEXACT);
-    ff_fmt_convert_init(&vc->fmt_conv, avctx);
 
     avctx->sample_fmt = AV_SAMPLE_FMT_FLTP;
 

libavcodec/wma.c

@@ -94,7 +94,6 @@ av_cold int ff_wma_init(AVCodecContext *avctx, int flags2)
         avctx->bit_rate    <= 0)
         return -1;
 
-    ff_fmt_convert_init(&s->fmt_conv, avctx);
     avpriv_float_dsp_init(&s->fdsp, avctx->flags & CODEC_FLAG_BITEXACT);
 
     if (avctx->codec->id == AV_CODEC_ID_WMAV1)

libavcodec/wma.h

@@ -25,7 +25,6 @@
 #include "libavutil/float_dsp.h"
 
 #include "fft.h"
-#include "fmtconvert.h"
 #include "get_bits.h"
 #include "put_bits.h"
 
@@ -131,7 +130,6 @@ typedef struct WMACodecContext {
     float lsp_pow_e_table[256];
     float lsp_pow_m_table1[(1 << LSP_POW_BITS)];
     float lsp_pow_m_table2[(1 << LSP_POW_BITS)];
-    FmtConvertContext fmt_conv;
     AVFloatDSPContext fdsp;
 
 #ifdef TRACE