1. 程式人生 > >【晶片手冊開發】Sil9136音訊開發詳細分析+原始碼實戰

【晶片手冊開發】Sil9136音訊開發詳細分析+原始碼實戰

[toc] --- ## 前言 * 預設在開發了視訊方面後 * 這方面的工作本來可以找技術支援拿個例程參考下,很快就可以的寫出來的,因為自己對HDMI協議不太瞭解,但是技術支援說沒有,所以沒辦法,只能自己搞了 * 看手冊不難,難的是找資料 * 記錄一下,也分享一下 ## 參考 * sil9136暫存器手冊:**《Sil-PR-1060-C》** * HDMI協議手冊:**《HDMI_1.4》** * CEA標準手冊:**《CEA-861-D[安全]》** * **使用參考例程**(*無音訊功能*) ## 手冊使用+實戰 * 以 I2S 介面為例開發 * 直接看手冊配置相關暫存器 * **記得輸入與輸出配對** * 如編碼型別 * 取樣長度 * 取樣頻率 * 等等 ### 配置 * 《Sil-PR-1060-C》手冊,28頁起 * ![](https://img2020.cnblogs.com/blog/2085252/202011/2085252-20201120213723759-111667844.png) * 圖中說明 sil9136 支援 S/PDIF, I2S or DSD模式,主機可以通過配置TPI選擇不同的模式 * ![](https://img2020.cnblogs.com/blog/2085252/202011/2085252-20201120213731654-61667321.png) * 這個暫存器表比較重要,說明了sil9136的暫存器配置 * 0x26 暫存器 * [7:6] * 選擇模式,支援 * none * S/PDIF * I2S * DSD * [5] * 通道數,支援 * 雙通道 * 8 通道 * [4] * 靜音配置 * [3:0] * 編碼型別,有 * Refer to Stream Header * PCM (*本次使用 PCM*) * AC-3 * MPEG1 * MP3 * MPEG2 * AAC * DTS * ATRAC * 0x27 暫存器 * [7:6] * 音訊取樣位元位長度 **SS** * Refer to Stream Header * 16 bit * 20 bit * 24 bit * [5:3] * 音訊取樣頻率 **SF** * Refer to Stream Header * 32 kHz * 44.1 kHz * 48 kHz * 88.2 kHz * 96 kHz * 176.4 kHz * 192 kHz * [2] * 是否支援高位元率 * **注意**:圖中說明的 0x24 和 0x25 暫存器只有在 S/PDIF 模式下有效,即是 0x26[7:6]=01 時。 ### Configuring Audio Using I2S * 直接跳到配置 I2S 流程,實現配置邏輯 * ![](https://img2020.cnblogs.com/blog/2085252/202011/2085252-20201120213743494-1600981779.png) * 上圖已經很明顯地顯示出了配置出 I2S 的流程了 * 步驟: 1. 確保有有效的 I2S 訊號進入 sil9136 2. **設定 0x26[4] 為靜音模式** * 直接呼叫例程介面:**`SetAudioMute(AUDIO_MUTE_MUTED);`** 3. **通過 0x20 來配置進來的 SD 格式** * ![](https://img2020.cnblogs.com/blog/2085252/202011/2085252-20201120213752463-1751236004.png) * *配置要和輸入的音訊配置搭配* * 以下為個人選擇 * SCK Sample Edge :Rising * MCLK Multiplier:256 * WS Polarity – Left when:WS is Low * SD Justify Data is justified:Left * SD Direction Byte shifted first:MSB * WS to SD First Bit Shift:Yes * 程式為:**`WriteByteTPI(TPI_I2S_IN_CFG, (0x80x10));`** 4. **通過多次設定 0x1F 來配置每一個 SD 輸入對映** * ![](https://img2020.cnblogs.com/blog/2085252/202011/2085252-20201120213801164-848130155.png) * SDx與FIFOn的對映 * 支援一對多 * **注意:必須順序對映,如如果要對映FIFO2,就必須先完成FIFO0和FIFO1的對映** * 我的程式碼段:**SD0-FIFO0; SD1-FIFO1; SD2-FIFO2; SD3-FIFO3;** ```c do{ WriteByteTPI(TPI_I2S_EN, 0x80); Tmp = ReadByteTPI(TPI_I2S_EN); }while(Tmp != 0x80); do{ WriteByteTPI(TPI_I2S_EN, 0x91); Tmp = ReadByteTPI(TPI_I2S_EN); }while(Tmp != 0x91); do{ WriteByteTPI(TPI_I2S_EN, 0xA2); Tmp = ReadByteTPI(TPI_I2S_EN); }while(Tmp != 0xA2); do{ WriteByteTPI(TPI_I2S_EN, 0xB3); Tmp = ReadByteTPI(TPI_I2S_EN); }while(Tmp != 0xB3); ``` 5. **通過設定 0x27[5:3] 來配置音訊取樣頻率** * 配置為48kHz:**`ReadModifyWriteTPI(TPI_AUDIO_SAMPLE_CTRL, 0x38, 0x18);`** 6. **設定 0x21-0x25 來配置傳送到HDMI的頭資訊** * ![](https://img2020.cnblogs.com/blog/2085252/202011/2085252-20201120213815694-1393027249.png) * 上圖 0x21-0x25 的描述在 I2S 模式有效,即是 0x26[7:6] = 0x10 * 主要配置兩個引數 1. 取樣頻率:48 kHz 2. 取樣長度:24 bits * 程式碼段 ```c WriteByteTPI(TPI_I2S_CHST_0, 0x00); WriteByteTPI(TPI_I2S_CHST_1, 0x00); WriteByteTPI(TPI_I2S_CHST_2, 0x00); WriteByteTPI(TPI_I2S_CHST_3, 0x02); WriteByteTPI(TPI_I2S_CHST_4, 0x0B); ``` 7. **設定 0xBF-0xCD 來配置 audio infoframe** * 這步驟先給出最終程式碼再分析:**`SetAudioInfoFrames(TWO_CHANNELS, 0x00, 0x00, 0x18, 0x00);`** * ![](https://img2020.cnblogs.com/blog/2085252/202011/2085252-20201120213829449-1023336344.png) * 說明了配置 audio infoframe 的必要性和需要參考的檔案 **HDMI Specification**,根據本檔案說明,瞭解到sil9136 支援 HDMI1.4 協議,所以準備好檔案《HDMI_1.4》,並找到關於 **audio infoframe** 的說明。 * 圖A * ![](https://img2020.cnblogs.com/blog/2085252/202011/2085252-20201120213858023-1300178249.png) * 圖B * ![](https://img2020.cnblogs.com/blog/2085252/202011/2085252-20201120213903780-1784533537.png) * 圖C * ![](https://img2020.cnblogs.com/blog/2085252/202011/2085252-20201120213909903-1325677670.png) * 圖Cc1,*《CEA-861-D安全》* * ![](https://img2020.cnblogs.com/blog/2085252/202011/2085252-20201120213917708-1032554316.png) * 圖Cc2,*《CEA-861-D安全》* * ![](https://img2020.cnblogs.com/blog/2085252/202011/2085252-20201120213925424-1404842599.png) * 圖D * ![](https://img2020.cnblogs.com/blog/2085252/202011/2085252-20201120213930718-497220773.png) * 結合例程原始碼 `bool SetAudioInfoFrames` 函式,得出只需要瞭解幾個引數配置即可。 * **byte ChannelCount** * 參照*圖C和圖Cc1*,C0...C2,選擇雙通道,得出值為 **0x01** * **byte CodingType** * CT0...CT3: The CT bits shall always be set to a value of 0 (“Refer to Stream Header”).即是置為0即可 * **SS** * The SS bits shall always be set to a value of 0 (“Refer to Stream Header”). 即是置為0即可 * **Fs** * ***參考原始碼、0x27暫存器、圖C和圖Cc2*** * 推測 `Fs` 就是取樣頻率 SF0...SF2 * `B_Data[6] = (Fs >> 1) | (SS >> 6);` * Fs在圖C中的PB2[4:2],而上述程式碼中右移一位,所以 Fs 的值佔用[5:3],參考0x27。 * 取樣頻率為 48kHz,得出 Fs=x018 * **SpeakerConfig** * 參考原始碼 `B_Data[8] = SpeakerConfig;` 得出 SpeakerConfig 為 圖C中的PB4 * 這裡為 LPCM ,所以 SpeakerConfig = 0; ```c ////////////////////////////////////////////////////////////////////////////// // // FUNCTION : SetAudioInfoFrames() // // PURPOSE : Load Audio InfoFrame data into registers and send to sink // // INPUT PARAMS : (1) Channel count (2) speaker configuration per CEA-861D // Tables 19, 20 (3) Coding type: 0x09 for DSD Audio. 0 (refer // to stream header) for all the rest (4) Sample Frequency. Non // zero for HBR only (5) Audio Sample Length. Non zero for HBR // only. // // OUTPUT PARAMS : None // // GLOBALS USED : None // // RETURNS : TRUE // ////////////////////////////////////////////////////////////////////////////// bool SetAudioInfoFrames(byte ChannelCount, byte CodingType, byte SS, byte Fs, byte SpeakerConfig) { byte B_Data[SIZE_AUDIO_INFOFRAME]; // 14 byte i; TPI_TRACE_PRINT((">>SetAudioInfoFrames()\n")); for (i = 0; i < SIZE_AUDIO_INFOFRAME +1; i++) B_Data[i] = 0; B_Data[0] = EN_AUDIO_INFOFRAMES; // 0xC2 B_Data[1] = TYPE_AUDIO_INFOFRAMES; // 0x84 B_Data[2] = AUDIO_INFOFRAMES_VERSION; // 0x01 B_Data[3] = AUDIO_INFOFRAMES_LENGTH; // 0x0A B_Data[5] = ChannelCount; // 0 for "Refer to Stream Header" or for 2 Channels. 0x07 for 8 Channels B_Data[5] |= (CodingType << 4); // 0xC7[7:4] == 0b1001 for DSD Audio B_Data[4] = 0x84 + 0x01 + 0x0A; // Calculate checksum // B_Data[6] = (Fs << 2) | SS; B_Data[6] = (Fs >> 1) | (SS >> 6); //write Fs to 0x27[5:3] and SS to 0x27[7:6] to update the IForm with the current value. // ReadModifyWriteTPI(TPI_AUDIO_SAMPLE_CTRL, BITS_7_6 | BITS_5_4_3, (B_Data[6] & BITS_1_0) << 6 | (B_Data[6] & 0x1C) << 1); B_Data[8] = SpeakerConfig; for (i = 5; i < SIZE_AUDIO_INFOFRAME; i++) B_Data[4] += B_Data[i]; B_Data[4] = 0x100 - B_Data[4]; g_audio_Checksum = B_Data[4]; // Audio checksum for global use WriteBlockTPI(TPI_AUDIO_BYTE_0, SIZE_AUDIO_INFOFRAME, B_Data); #ifdef DEV_EMBEDDED EnableEmbeddedSync(); #endif return TRUE; } ``` 8. **I2S 模式, 設定音訊通道數,並關閉靜音** 1. 程式碼:**`WriteByteTPI(TPI_AUDIO_INTERFACE_REG, AUD_IF_I2S | TWO_CHANNEL_LAYOUT | 0x01);`** 2. ***注意:audio inframe中的通道數配置必須和 0x26 配置的一樣*** ### 總結實現 * 得出一下程式碼,並把一下函式放到熱插拔的插入後執行即可 ```c /** * @brief setPrivateAudio(void) * @param * @retval * @author lzm */ void setPrivateAudio(void) { byte Tmp = 0; /* Select I2S input mode using TPI 0x26[7:6], with Mute enabled (bit [4] = 1). */ SetAudioMute(AUDIO_MUTE_MUTED); /* Write register TPI 0x20 to select the general incoming SD format. */ WriteByteTPI(TPI_I2S_IN_CFG, (SCK_SAMPLE_EDGE | 0x10)); /* Write register TPI 0x1F up to four times, to program each of the SD inputs. */ do{ WriteByteTPI(TPI_I2S_EN, 0x80); Tmp = ReadByteTPI(TPI_I2S_EN); }while(Tmp != 0x80); do{ WriteByteTPI(TPI_I2S_EN, 0x91); Tmp = ReadByteTPI(TPI_I2S_EN); }while(Tmp != 0x91); do{ WriteByteTPI(TPI_I2S_EN, 0xA2); Tmp = ReadByteTPI(TPI_I2S_EN); }while(Tmp != 0xA2); do{ WriteByteTPI(TPI_I2S_EN, 0xB3); Tmp = ReadByteTPI(TPI_I2S_EN); }while(Tmp != 0xB3); // /* Program register TPI 0x27 with the correct audio about. */ // WriteByteTPI(TPI_AUDIO_SAMPLE_CTRL, AUDIO_SAMPLE_SIZE_24BIT | AUDIO_SAMPLE_FREQ_48KHZ | AUDIO_SAMPLE_HBR_DISABLE); /* Program register TPI 0x27[5:3] with the correct audio rate */ ReadModifyWriteTPI(TPI_AUDIO_SAMPLE_CTRL, 0x38, AUDIO_SAMPLE_FREQ_48KHZ); /* Program registers TPI 0x21-x25 with the correct header information for the stream that will be sent over HDMI. */ WriteByteTPI(TPI_I2S_CHST_0, 0x00); WriteByteTPI(TPI_I2S_CHST_1, 0x00); WriteByteTPI(TPI_I2S_CHST_2, 0x00); WriteByteTPI(TPI_I2S_CHST_3, 0x02); WriteByteTPI(TPI_I2S_CHST_4, 0x0B); /* Write registers TPI 0xBF-xCD with the appropriate Audio InfoFrame information. */ SetAudioInfoFrames(TWO_CHANNELS, 0x00, 0x00, 0x18, 0x00); /* Set the audio packet header layout indicator to 2-channel or multi-channel mode as needed using the sequence described below. Note that Audio InfoFrame byte 1 must also have this same setting. */ /* Again write register TPI 0x26 with I2S selected, this time with Mute disabled (bit [4] = 0). */ WriteByteTPI(TPI_AUDIO_INTERFACE_REG, AUD_IF_I2S | TWO_CHANNEL_LAYOUT | 0x01); SetAudioMute(AUDIO_MUTE_NORMAL