FreeCalypso > hg > freecalypso-tools

comparison doc/Audio-mode-config @ 838:02d92d49c9f8

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
doc/Audio-mode-config: update for old vs new AEC
author Mychaela Falconia <falcon@freecalypso.org>
date Fri, 30 Jul 2021 05:21:19 +0000
parents 0321cd08b19f
children 6a0fcbca8ac7
comparison
equal deleted inserted replaced
837:724a6bc2a4b8 838:02d92d49c9f8
50 /aud/modename.vol is the corresponding volume setting file 50 /aud/modename.vol is the corresponding volume setting file
51 51
52 This paradigm is a good fit for "dumbphone" handsets in which there usually 52 This paradigm is a good fit for "dumbphone" handsets in which there usually
53 will be several different voice audio configurations for classic handheld 53 will be several different voice audio configurations for classic handheld
54 operation, for the hands-free loudspeaker mode, for operation with a wired 54 operation, for the hands-free loudspeaker mode, for operation with a wired
55 headset, and if the phone uses a loudspeaker (as opposed to a piezo buzzer) to 55 headset, and if the phone uses its hands-free loudspeaker plus the Calypso DSP
56 play ringtones and uses the Calypso DSP to generate those ringtone melodies, 56 to play ringtones (as opposed to using a buzzer on BU/PWT or a ringtone player
57 there will also need to be an output-only audio configuration for ringing. 57 chip that drives the speaker bypassing the voice path), there will also need to
58 be an output-only audio configuration for ringing.
58 59
59 How do the audio mode config files under /aud come into being? It appears that 60 How do the audio mode config files under /aud come into being? It appears that
60 TI's original intent was that a configuration would be manually constructed on 61 TI's original intent was that a configuration would be manually constructed on
61 a test device via TMSH auw commands, saved in the FFS of that test device with 62 a test device via TMSH auw commands, saved in the FFS of that test device with
62 the aus command, then read out of that test device FFS in binary form and 63 the aus command, then read out of that test device FFS in binary form and
91 up or down in the UI, the *.vol file _corresponding to the current mode_ will 92 up or down in the UI, the *.vol file _corresponding to the current mode_ will
92 be updated by the running fw. Thus the fw would maintain a separate notion of 93 be updated by the running fw. Thus the fw would maintain a separate notion of
93 the current volume for ringing, for the earpiece speaker, for the hands-free 94 the current volume for ringing, for the earpiece speaker, for the hands-free
94 loudspeaker and for the wired headset, something which Pirelli's fw very 95 loudspeaker and for the wired headset, something which Pirelli's fw very
95 notoriously fails to do. 96 notoriously fails to do.
97
98 Old vs. new AEC
99 ===============
100
101 One of the settings in the audio mode config structure underwent an evolutionary
102 change within the span of history that is relevant to FreeCalypso - this setting
103 is the configuration for AEC, the Acoustic Echo Cancellation functional block
104 of the Calypso DSP. As TI's GSM DSPs evolved (before, during and after the
105 Calypso era), their AEC implementation evolved along with the rest, and
106 different evolutionary versions of AEC require different configuration and
107 tuning parameters. When the audio mode facility was first implemented, the AEC
108 block in TI's GSM DSPs of that time was controlled with a single 16-bit control
109 word; the people in the SSA group who implemented RiViera Audio Service then
110 decided to split different bits from this one DSP control word into 5 different
111 parameter words, and the result was the "old" 5-word AEC config.
112
113 But the version of AEC implemented in the DSP ROM in the Calypso silicon version
114 we work with is slightly newer; this version corresponds to what TI's L1 code
115 calls L1_NEW_AEC. However, the waters then got muddied: for reasons which we
116 (FreeCalypso team) cannot understand (perhaps miscommunication between different
117 groups at TI), TI's TCS211 reference firmware shipped with L1_NEW_AEC disabled
118 (C preprocessor symbol set to 0 instead of 1), even though the underlying DSP
119 AEC block (combination of ROM and official patches) is the "new" kind and not
120 the "old" one. There are two fallouts from this software misconfiguration on
121 TI's part:
122
123 1) If one takes stock TCS211 from TI or any derivative version in which this
124 aspect is unchanged (all mokoN firmwares, and all FC firmwares up to
125 Magnetite) and tries to enable AEC, the result will be a poor AEC
126 configuration: the old echo level and long vs short settings do nothing on
127 the new DSP, whereas the new tunable parameters will remain at their defaults
128 with no way to tweak them. I (Mother Mychaela) can only guess that this
129 situation is what Openmoko must have run into when they tried to get AEC
130 working.
131
132 2) When someone downstream of TI figures out that L1_NEW_AEC needs to be changed
133 from 0 to 1 and actually makes that change, like we did in our Tourmaline fw,
134 the format and size of the audio mode binary structure change, and all old
135 audio mode config files become invalid.
136
137 Our FreeCalypso work is affected by point 2 above: we started working with audio
138 mode config files in 2017, using the old AEC configuration, and only made the
139 switch to L1_NEW_AEC in 2021. We now have two kinds of audio mode config binary
140 files: the old kind that are 164 bytes long, and the new kind that are 176 bytes
141 long. Our Tourmaline firmware has L1_NEW_AEC enabled, while Magnetite (our
142 legacy backward compatiblity fw) has it disabled.
143
144 To prevent loading of garbage into AEC config when an audio mode file of the
145 wrong kind is loaded, we have implemented the following workaround in both
146 Tourmaline and Magnetite: if the loaded mode config file has the wrong length,
147 the AEC config is set to the default disabled state instead of whatever is in
148 the mode file - loading an AEC config of the wrong format is not possible.
96 149
97 Default audio configuration 150 Default audio configuration
98 =========================== 151 ===========================
99 152
100 The default audio config set in the Iota ABB registers and in the DSP when no 153 The default audio config set in the Iota ABB registers and in the DSP when no
121 fir 24 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 174 fir 24 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000 0x0000
122 } 175 }
123 sidetone -5 176 sidetone -5
124 aec 0 0 0 0 0 177 aec 0 0 0 0 0
125 178
179 The above version is the one produced by Magnetite and earlier firmwares without
180 L1_NEW_AEC; in the new Tourmaline version the last line changes to:
181
182 aec-new 0 0 0x1 0x7FFF 0x1FFF 0x4000 0x32 0x1000 0x1000 0 0 0
183
126 The meaning is as follows: 184 The meaning is as follows:
127 185
128 * voice-path is the DSP digital voice path setting, 0 means the standard 186 * voice-path is the DSP digital voice path setting, 0 means the standard
129 configuration with the voice channel going between GSM and the local analog 187 configuration with the voice channel going between GSM and the local analog
130 voice hardware attached to the ABB. 188 voice hardware attached to the ABB.
142 is NOT set, i.e., the normal configuration. 200 is NOT set, i.e., the normal configuration.
143 201
144 * DSP FIR filters do nothing, as coefficient 0 is set to unity and all other 202 * DSP FIR filters do nothing, as coefficient 0 is set to unity and all other
145 coefficients are set to zero. 203 coefficients are set to zero.
146 204
147 * The AEC mechanism in the DSP is disabled. 205 * The AEC mechanism in the DSP is disabled, although the format of the bits that
206 say so is different between old and new AEC versions. In the new version
207 there are a number of tunable settings that only kick in when AEC is enabled;
208 when AEC is disabled by default, these tunable knobs still have sensible
209 defaults that aren't all zeros.
148 210
149 Creating your own audio mode configurations 211 Creating your own audio mode configurations
150 =========================================== 212 ===========================================
151 213
152 The input to our tiaud-compile utility can contain every setting shown in the 214 The input to our tiaud-compile utility can contain every setting shown in the
181 input to tiaud-compile you can put each coefficient on its own fir line, put 243 input to tiaud-compile you can put each coefficient on its own fir line, put
182 all 31 coefficients on the same line, or group them in any other way you like. 244 all 31 coefficients on the same line, or group them in any other way you like.
183 The grouping used in the tiaud-decomp output has been chosen for line length 245 The grouping used in the tiaud-decomp output has been chosen for line length
184 reasons. 246 reasons.
185 247
186 * The numbers given on fir and aec lines are 16-bit values that go directly into 248 aec vs aec-new in tiaud-compile input
187 the DSP; the former are FIR coefficients and the latter are bit masks. They 249 =====================================
188 can be given as either decimal or hexadecimal with 0x prefix in the ASCII 250
189 input to tiaud-compile. 251 tiaud-compile accepts both aec (old) and aec-new settings; aec must be followed
252 by 5 numbers, aec-new must be followed by 12 numbers. Each number is a 16-bit
253 value, and they go into the binary structure without further interpretation by
254 tiaud-compile - instead the firmware is the entity that gives them meaning.
255 Numbers without 0x prefix are interpreted as decimal.
256
257 tiaud-compile will generate one type or the other of the binary output file,
258 following these rules:
259
260 * If an aec setting is given, a 164 byte file will be produced, with the 5 AEC
261 words being the given ones.
262
263 * If an aec-new setting is given, a 176 byte file will be produced, with the 12
264 AEC words being the given ones.
265
266 * If neither setting is given, a 164 byte file will be produced, with the 5 AEC
267 words of the old type being all zeros. Thanks to the modified audio mode
268 loading code in our firmwares, these 164 byte mode files can still be used
269 with current Tourmaline fw, with AEC set to its default disabled state.