Vitor Sessak a écrit :
> On 07/13/2010 09:49 PM, Sebastian Vater wrote:
>> Vitor Sessak a écrit :
>>> On 07/11/2010 10:05 PM, Sebastian Vater wrote:
>>>
>>>> /*
>>>> * AVSequencer order list and data management
>>>> * Copyright (c) 2010 Sebastian
>>>> Vater<cdgs.basty-gM/ye1e23mwn+bqq9rb...@public.gmane.org>
>>>> *
>>>> * This file is part of FFmpeg.
>>>> *
>>>> * FFmpeg is free software; you can redistribute it and/or
>>>> * modify it under the terms of the GNU Lesser General Public
>>>> * License as published by the Free Software Foundation; either
>>>> * version 2.1 of the License, or (at your option) any later version.
>>>> *
>>>> * FFmpeg is distributed in the hope that it will be useful,
>>>> * but WITHOUT ANY WARRANTY; without even the implied warranty of
>>>> * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
>>>> * Lesser General Public License for more details.
>>>> *
>>>> * You should have received a copy of the GNU Lesser General Public
>>>> * License along with FFmpeg; if not, write to the Free Software
>>>> * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
>>>> 02110-1301 USA
>>>> */
>>>>
>>>> #ifndef AVSEQUENCER_ORDER_H
>>>> #define AVSEQUENCER_ORDER_H
>>>>
>>>> #include "libavsequencer/track.h"
>>>>
>>>> /**
>>>> * Song order list structure, This structure is actually for one
>>>> channel
>>>> * and therefore actually pointed as an array with size of number of
>>>> * host channels.
>>>
>>> Wouldn't it be better named AVSequencerChannelData?
>>
>> It is only channel based if you import from a tracker which separates
>> channels for each track, but most trackers don't do this, they have just
>> one entry for all the channels, since they only know synchronized
>> channels.
>>
>>>
>>>> * New fields can be added to the end with minor version bumps.
>>>> * Removal, reordering and changes to existing fields require a major
>>>> * version bump.
>>>> */
>>>> typedef struct AVSequencerOrderList {
>>>> /** Array of pointers containing all order list data used by this
>>>> channel. */
>>>> AVSequencerOrderData **order_data;
>>>
>>> Please replace in the comment "order list data" for a brief
>>> description of what it means. And BTW, is this an "order _list_ data"?
>>
>> Actually it is the data of the order list, order_data contains all order
>> list entries like AVSequencerTrackData contains all the rows of
>> AVSequencerTrack.
>
> Is it the list of the rows in chronological order? Is it the list of
> notes in alphabetical order? Also, if I understand well, it is an
> array of linked lists. Why?
>
>>>> /** Number of order list data used for this channel. */
>>>> uint16_t orders;
>>>>
>>>> /** Number of order list data entries to use for this
>>>> channel. */
>>>> uint16_t length;
>>>>
>>>> /** Repeat start order list data number for this channel. */
>>>> uint16_t rep_start;
>>>>
>>>> /** Volume level for this channel (defaults to 255). */
>>>> uint8_t volume;
>>>> #define AVSEQ_ORDER_LIST_VOLUME 255
>>>>
>>>> /** Sub-volume level for this channel. This is basically channel
>>>> volume divided by 256, but the sub-volume doesn't account
>>>> into actual mixer output (defaults 0). */
>>>> uint8_t sub_volume;
>>>> #define AVSEQ_ORDER_LIST_SUB_VOLUME 0
>>>>
>>>> /** Stereo track panning level for this channel (defaults to
>>>> -128 = central stereo track panning). */
>>>> int8_t track_panning;
>>>> #define AVSEQ_ORDER_LIST_TRACK_PAN -128
>>>>
>>>> /** Stereo track sub-panning level for this channel. This is
>>>> basically track panning divided by 256, but the sub-panning
>>>> doesn't account into actual mixer output (defaults 0). */
>>>> uint8_t track_sub_panning;
>>>> #define AVSEQ_ORDER_LIST_TRACK_SUB_PAN -128
>>>>
>>>> /** Stereo panning level for this channel (defaults to
>>>> -128 = central stereo panning). */
>>>> int8_t channel_panning;
>>>> #define AVSEQ_ORDER_LIST_PANNING -128
>>>>
>>>> /** Stereo sub-panning level for this channel. This is
>>>> basically channel panning divided by 256, but the sub-panning
>>>> doesn't account into actual mixer output (defaults 0). */
>>>> uint8_t channel_sub_panning;
>>>> #define AVSEQ_ORDER_LIST_SUB_PANNING 0
>>>
>>>> /** Compatibility flags for playback. There are rare cases
>>>> where order handling can not be mapped into internal
>>>> playback engine and have to be handled specially. For
>>>> each order list which needs this, this will define new
>>>> flags which tag the player to handle it to that special
>>>> way. */
>>>> uint8_t compat_flags;
>>>
>>> I suppose this is not unused ATM?
>>
>> Fixed.
>>
>>>
>>>> /** Order list playback flags. Some sequencers feature
>>>> surround panning or allow initial muting. which has to
>>>> be taken care specially in the internal playback engine.
>>>> Also sequencers differ in how they handle slides. */
>>>> uint8_t flags;
>>>> #define AVSEQ_ORDER_LIST_FLAG_CHANNEL_SURROUND 0x01 ///< Initial
>>>> channel surround instead of stereo panning
>>>> #define AVSEQ_ORDER_LIST_FLAG_TRACK_SURROUND 0x02 ///< Initial
>>>> track surround instead of stereo panning
>>>> #define AVSEQ_ORDER_LIST_FLAG_MUTED 0x04 ///< Initial
>>>> muted channel
>>>>
>>>> } AVSequencerOrderList;
>>>>
>>>> /**
>>>> * Song order list data structure, this contains actual order list
>>>> data.
>>>> * New fields can be added to the end with minor version bumps.
>>>> * Removal, reordering and changes to existing fields require a major
>>>> * version bump.
>>>> */
>>>> typedef struct AVSequencerOrderData {
>>>> /** AVSequencerTrack pointer to track which should be played. */
>>>> AVSequencerTrack *track;
>>>
>>>> /** Next order list data pointer if seeking forward one
>>>> frame. */
>>>> AVSequencerOrderData *next_pos;
>>>>
>>>> /** Previous order list data pointer if seeking backward one
>>>> frame. */
>>>> AVSequencerOrderData *prev_pos;
>>>
>>> These do not look to belong to the BSS.
>>
>> They are used to allow the composer to define alternative order
>> sequences in case of forward and backward seeking. These alternatives
>> can pre-initialize note data in that case. This is a very rarely used
>> feature, but avoids the problem that instruments are missing if you skip
>> one channel.
>>
>> Example:
>> We have a track playing a new instrument at row 60. If the user skips
>> this track before arriving row 60, it will not be played at all, thus
>> causing silence. Now consider the normal next row, would change data
>> here or even has a complete empty row (but expecting to play a long
>> looped instrument), in that case it would confuse the actual output.
>> These pointers allow the composer to initialize that in case.
>
> Is this about an human user seeking the file he is playing or a
> tracker formats that specify seeking commands?
>
>>>> /** Tempo change or zero to skip tempo change. */
>>>> uint16_t tempo;
>>>
>>> Cryptic comment. This is the timestamp where the tempo change? An
>>> index to a list a tempo changing structures? The instrument number
>>> that has a different tempo?
>>
>> Fixed.
>>
>>>
>>>> /** Played nesting level (GoSub command maximum nesting
>>>> depth). */
>>>> uint16_t played;
>>>
>>> Again, BSS?
>>
>> Would require duplicating each track information in the player and
>> therefore add a lot of unnecessary complexity.
>
> So that means that the player _writes_ to this struct? This seems not
> optimal design in a modularity point of view. IMHO, everything that is
> feeded to the player should be read-only and the player should only
> write in his own context structure.
>
>>>> /** Track volume (this overrides settings in AVSequencerTrack).
>>>> To enable this, the flag AVSEQ_ORDER_DATA_FLAG_SET_VOLUME
>>>> must be set in flags. */
>>>> uint8_t volume;
>>>
>>> This is ugly, why not making it a signed int and summing it to
>>> AVSequencerTrack.volume no matter what?
>>
>> This again would seriously break compatibility with almost any tracker.
>> Trackers either set volume by track data XOR by order data, never both.
>> It is not a volume transpose but a set volume command.
>
> What is the difference between a track volume and a order data volume?
> If they are the same, since only one or the other is specified per
> file, why not using a single struct field for it, without caring where
> in the file it is defined (and the module loader should know where to
> look)?
>
Hi I have excellent news!
libavsequencer now flawlessly integrates into FFmpeg, just check out my
latest git. Please do a git pull --rebase, Stefano had problems without
using it.
Here are the order.[ch] part of the BSS to review.
This version compiles perfectly.
--
Best regards,
:-) Basty/CDGS (-:
diff --git a/libavsequencer/order.c b/libavsequencer/order.c
new file mode 100644
index 0000000..f54f8e4
--- /dev/null
+++ b/libavsequencer/order.c
@@ -0,0 +1,72 @@
+/*
+ * Implement AVSequencer order list and data stuff
+ * Copyright (c) 2010 Sebastian Vater <cdgs.ba...@googlemail.com>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file
+ * Implement AVSequencer order list and data stuff.
+ */
+
+#include "libavutil/log.h"
+#include "libavsequencer/avsequencer.h"
+
+static const char *order_list_name(void *p)
+{
+ AVSequencerOrderList *order_list = p;
+ AVMetadataTag *tag = av_metadata_get(order_list->metadata, "title", NULL, AV_METADATA_IGNORE_SUFFIX);
+
+ return tag->value;
+}
+
+static const AVClass avseq_order_list_class = {
+ "AVSequencer Order List",
+ order_list_name,
+ NULL,
+ LIBAVUTIL_VERSION_INT,
+};
+
+int avseq_order_open(AVSequencerSong *song) {
+ AVSequencerOrderList *order_list;
+ uint16_t channels;
+
+ if (!song)
+ return AVERROR_INVALIDDATA;
+
+ channels = song->channels;
+
+ if ((channels == 0) || (channels > 256)) {
+ return AVERROR_INVALIDDATA;
+ } else if (!(order_list = av_mallocz(channels * sizeof(AVSequencerOrderData)))) {
+ av_log(song, AV_LOG_ERROR, "cannot allocate order list storage container.\n");
+ return AVERROR(ENOMEM);
+ }
+
+ song->order_list = order_list;
+
+ do {
+ order_list->av_class = &avseq_order_list_class;
+ order_list->volume = 255;
+ order_list->track_panning = -128;
+ order_list->channel_panning = -128;
+ order_list++;
+ } while (--channels);
+
+ return 0;
+}
diff --git a/libavsequencer/order.h b/libavsequencer/order.h
new file mode 100644
index 0000000..70bcf0b
--- /dev/null
+++ b/libavsequencer/order.h
@@ -0,0 +1,193 @@
+/*
+ * AVSequencer order list and data management
+ * Copyright (c) 2010 Sebastian Vater <cdgs.ba...@googlemail.com>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifndef AVSEQUENCER_ORDER_H
+#define AVSEQUENCER_ORDER_H
+
+#include "libavsequencer/track.h"
+
+/** AVSequencerOrderData->flags bitfield. */
+enum AVSequencerOrderDataFlags {
+ AVSEQ_ORDER_DATA_FLAG_END_ORDER = 0x01, ///< Order data indicates end of order
+ AVSEQ_ORDER_DATA_FLAG_END_SONG = 0x02, ///< Order data indicates end of whole song
+ AVSEQ_ORDER_DATA_FLAG_NOT_IN_ONCE = 0x04, ///< Order data will be skipped if you're playing in one-time mode
+ AVSEQ_ORDER_DATA_FLAG_NOT_IN_REPEAT = 0x08, ///< Order data will be skipped if you're playing in repeat mode
+ AVSEQ_ORDER_DATA_FLAG_TRACK_SYNC = 0x10, ///< Order data is a track synchronization point.
+ AVSEQ_ORDER_DATA_FLAG_SET_VOLUME = 0x20, ///< Order data takes advantage of the order list volume set
+};
+
+/**
+ * Song order list data structure, this contains actual order list data.
+ * New fields can be added to the end with minor version bumps.
+ * Removal, reordering and changes to existing fields require a major
+ * version bump.
+ */
+typedef struct AVSequencerOrderData {
+ /**
+ * information on struct for av_log
+ * - set by avseq_alloc_context
+ */
+ const AVClass *av_class;
+
+ /** Metadata information: Original order entry file name, order
+ * entry name, artist and comment. */
+ AVMetadata *metadata;
+
+ /** AVSequencerTrack pointer to track which should be played. */
+ AVSequencerTrack *track;
+
+ /** Next order list data pointer if seeking forward one frame. */
+ struct AVSequencerOrderData *next_pos;
+
+ /** Previous order list data pointer if seeking backward one
+ frame. */
+ struct AVSequencerOrderData *prev_pos;
+
+ /** Number of row to jump to if forward seeking one frame. */
+ uint16_t next_row;
+
+ /** Number of row to jump to if backward seeking one frame. */
+ uint16_t prev_row;
+
+ /** Beginning row for this track. If this is a track synchronization
+ point, the high byte is interpreted as the first track number
+ to be synchronized with and the low byte as the second track
+ number or for all channels when all 4 tracks are 0. */
+ uint16_t first_row;
+
+ /** Last row for this track. If this is a track synchronization
+ point, the high byte is interpreted as the third track number
+ to be synchronized with and the low byte as the fourth track
+ number or for all channels when all 4 tracks are 0.
+ If last row is set to 65535 in non synchronization mode,
+ the last row is always taken from AVSequencerTrack. */
+ uint16_t last_row;
+
+ /** Order list data playback flags. Some sequencers feature
+ special end markers or even different playback routes for
+ different playback modules (one-shot and repeat mode
+ playback), mark synchronization points or temporary
+ change volume), which has to be taken care specially
+ in the internal playback engine. */
+ uint8_t flags;
+
+ /** Relative note transpose for full track. Allows playing several
+ tracks some half-tones up/down. */
+ int8_t transpose;
+
+ /** Instrument transpose. All instruments will be relatively
+ mapped to this if this is non-zero. */
+ int16_t instr_transpose;
+
+ /** Tempo change or zero to skip tempo change. A tempo value of
+ zero would be zero, since that would mean literally execute
+ unlimited rows and tracks in just one tick. */
+ uint16_t tempo;
+
+ /** Played nesting level (GoSub command maximum nesting depth). */
+ uint16_t played;
+
+ /** Track volume (this overrides settings in AVSequencerTrack).
+ To enable this, the flag AVSEQ_ORDER_DATA_FLAG_SET_VOLUME
+ must be set in flags. This allows have a basic default track
+ volume by still allowing to override the track volume in case
+ the track is used multiple times, e.g. for creating echoes. */
+ uint8_t volume;
+
+ /** Track sub-volume. This is basically track volume
+ divided by 256, but the sub-volume doesn't account
+ into actual mixer output (this overrides AVSequencerTrack). */
+ uint8_t sub_volume;
+} AVSequencerOrderData;
+
+/** AVSequencerOrderList->flags bitfield. */
+enum AVSequencerOrderListFlags {
+ AVSEQ_ORDER_LIST_FLAG_CHANNEL_SURROUND = 0x01, ///< Initial channel surround instead of stereo panning
+ AVSEQ_ORDER_LIST_FLAG_TRACK_SURROUND = 0x02, ///< Initial track surround instead of stereo panning
+ AVSEQ_ORDER_LIST_FLAG_MUTED = 0x04, ///< Initial muted channel
+};
+
+/**
+ * Song order list structure, This structure is actually for one channel
+ * and therefore actually pointed as an array with size of number of
+ * host channels.
+ * New fields can be added to the end with minor version bumps.
+ * Removal, reordering and changes to existing fields require a major
+ * version bump.
+ */
+typedef struct AVSequencerOrderList {
+ /**
+ * information on struct for av_log
+ * - set by avseq_alloc_context
+ */
+ const AVClass *av_class;
+
+ /** Metadata information: Original order list file name, order
+ * list name, artist and comment. */
+ AVMetadata *metadata;
+
+ /** Array (of size orders) of pointers containing all order list
+ data used by this channel. */
+ AVSequencerOrderData **order_data;
+
+ /** Number of order list data used for this channel. */
+ uint16_t orders;
+
+ /** Number of order list data entries to use for this channel. */
+ uint16_t length;
+
+ /** Repeat start order list data number for this channel. */
+ uint16_t rep_start;
+
+ /** Volume level for this channel (defaults to 255). */
+ uint8_t volume;
+
+ /** Sub-volume level for this channel. This is basically channel
+ volume divided by 256, but the sub-volume doesn't account
+ into actual mixer output (defaults 0). */
+ uint8_t sub_volume;
+
+ /** Stereo track panning level for this channel (defaults to
+ -128 = central stereo track panning). */
+ int8_t track_panning;
+
+ /** Stereo track sub-panning level for this channel. This is
+ basically track panning divided by 256, but the sub-panning
+ doesn't account into actual mixer output (defaults 0). */
+ uint8_t track_sub_panning;
+
+ /** Stereo panning level for this channel (defaults to
+ -128 = central stereo panning). */
+ int8_t channel_panning;
+
+ /** Stereo sub-panning level for this channel. This is
+ basically channel panning divided by 256, but the sub-panning
+ doesn't account into actual mixer output (defaults 0). */
+ uint8_t channel_sub_panning;
+
+ /** Order list playback flags. Some sequencers feature
+ surround panning or allow initial muting. which has to
+ be taken care specially in the internal playback engine.
+ Also sequencers differ in how they handle slides. */
+ uint8_t flags;
+} AVSequencerOrderList;
+
+#endif /* AVSEQUENCER_ORDER_H */
_______________________________________________
FFmpeg-soc mailing list
FFmpeg-soc@mplayerhq.hu
https://lists.mplayerhq.hu/mailman/listinfo/ffmpeg-soc
