JP2001318696A - Device and method for fast feeding/listening and rewinding of compressed audio contents - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method to execute a fast-feeding/listening or a rewinding of audio contents expressed in a compressed format. SOLUTION: Repeated skips are made for a beforehand set amount to forward or to backward a compressed audio data stream, data blocks are retrieved and these data blocks are mutally rejoined. During the decoding, these data block joints are detected, resynchronization is rapidly conducted and fast feeding/listening and rewinding are realized.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

FIELD OF THE INVENTION The present invention relates to a method and apparatus for using digital communication signals, and more particularly to a method and apparatus for audibly fast-forwarding or rewinding compressed audio content.

[0002]

BACKGROUND OF THE INVENTION Providing audible fast-forward / rewind of compressed audio formats is generally a challenge due to the resynchronization problem associated with such formats. Figure X shows a typical compressed audio
Indicates the format of the data stream. The data stream is divided into units called frames. Each frame represents a segment of audio data that can be decoded. At the beginning of each frame is a header, which contains general information about the data stream, such as sampling rate, bit rate, profile, and the like. The first word of the header is the sync word, which is 1
A bit string that identifies the "start" of one frame.

The current generation of personal digital audio
The fast forward and rewind functions are silent in the player.
That is, the user does not hear anything during the fast forward (or rewind) operation. A simple fast-forward technique involves jumping and sending in the data stream by an amount related to the desired fast-forward rate, followed by resynchronization based on the frame header. To perform this resynchronization, a search is performed which searches the data stream for a bit string that matches the sync word. Once this sync word is found, the decoder can begin parsing the frame. In addition, a cyclic redundancy code (CRC) can be used for detecting errors in the data stream.

[0004] However, even this simple technique has many practical problems associated with it. Many compressed audio file formats, eg M
PEG-1 Layer 3 (MP3) or MPEG-2AC
C uses a large number of variable length codes. Certain acceptable sequences of these variable length codes can actually simulate a sync word, ie, the sync word is not unique. When searching within such a data stream, it is common to find a match with a "false" sync word.

[0005] In addition, CRC checking is typically not required and therefore is not always transmitted. When this is sent, a lot of parsing and / or computation would be needed to determine the validity of the frame. Some audio transfers use a one-bit field to indicate whether CRC is used. Parsing an incomplete header with this bit set to 0 (due to incomplete sync word detection) should actually override the CRC check.

[0006] Because of the use of variable length codes, parsing errors can occur undetected (and thus decoded / reproduced), but even random noise can be a valid sequence of codewords. That's why. As a result, the output of the decoder will be significantly distorted.

[0007] Furthermore, in some audio data streams, such as those associated with MP3, the data that requires decoding of the frame actually occurs before the sync word and header. One field in the header tells the decoder where to look for the data. This pointer should always point backward, but in some cases may point to the location of the previous frame header. If breaks or discontinuities occur in the data stream, synchronization becomes difficult because the data is not complete despite the header being found.

[0008] The synchronization of hierarchical or multiplexed data streams poses another problem. In such a data stream, the encrypted external bits
The stream carries fragments of the internal data stream as payload. This multiplexed data stream is decoded at the outer decoder, which extracts the payload data and provides it to the inner decoder.
Following the seam of the data stream, the outer decoder / decryptor must first acquire synchronization, followed by the inner decoder. This complicates the difficulty of problems such as resynchronization time and error resilience.

[0009]

SUMMARY The present invention provides a method and apparatus for audibly fast-forwarding or rewinding compressed audio content.

The invention is not limited to, for example, MPEG-1 Layer 3 (MP3) or MP3.
A method is provided for performing audible fast forward / rewind of audio content represented in a compressed format such as EG-2 Advanced Audio Code (ACC). A fast-forward controller is employed, which repeatedly skips forward or backward within the stored compressed audio data stream, retrieves large chunks of data, and subsequently retrieves these large chunks of data. , Stored in an end-to-end fashion such that they are tied back together as a single data stream to perform fast forward or rewind. A decoder is then used to decode each of these large chunks to detect that a large chunk switch has occurred (a splice in the data stream) and to quickly resynchronize upon each switch. The hierarchical or multiplexed data stream is decrypted,
Each is decoded using a serial decoder employing this technique. The decoder performs resynchronization and error recovery using a novel and robust sync word search.

For a more complete understanding of the present invention and its features, a more detailed description is given below with reference to the accompanying drawings.

[0012]

DETAILED DESCRIPTION The present invention provides a method and apparatus for audible fast forward or rewind of compressed audio content.

Referring to FIG. 1, there is shown a hardware platform or system 100 for playing digital audio files for an individual end user, such as a portable audio player. Digital audio files can be in various compressed and / or encoded formats. The platform 100 is included in a suitable holder so that it can be easily transported, manipulated and replaced stored audio files (audio content).

As can be seen from FIG. 1, the platform includes a digital signal processor (DSP) 110 and a microcontroller 120, which are interconnected. Typically, the DSP will store the stored program 11
Perform 2, 114, or 116 to decode and decompress the stored audio data file.
Microcontroller 120 provides suitable displays 122 and controls 124 to allow a user to operate the platform. Microcontroller and DS
P includes on-chip memories 132 and 130 and stores programs for operating the microcontroller and the DSP to provide desired functions to the portable audio player. Platform 100 also includes flash memory 140, which is preferably removable and stores digital audio data.

[0015] The platform includes a battery 150, which is replaceable and / or rechargeable, which provides power to other devices on the platform. There is a power supply block 160 connected in connection with the battery 150. The power supply block 160 is D
A C / DC converter 162 is included to convert the battery voltage to the voltage required to operate the device on the platform. A voltage regulator 164 is provided to regulate the selected voltage to a desired regulation level. Further, a voltage monitoring device 166 is provided as part of the power supply,
It monitors and controls the operation of the / DC converter and voltage regulator. For simplicity, FIG. 1 does not show all of the interconnections between the power supply and the devices in platform 100.

The DSP converts the audio bit stream to a stereo digital-to-analog (DAC) converter 1.
70, which converts the digital signal to an equivalent analog. A power amplifier 180 is interconnected with the DAC to amplify the signal and provides the signal to an external device such as a speaker, an earphone set, or other device for converting an electrical signal to an audible signal. I do.

Preferably, the DSP, in addition to decoding, provides equalization, volume, tone and balance control functions 185 to control signals from the microcontroller as a result of a user interface with control keys in the keypad 124. Execute in response. Alternatively, the power amplifier can respond to control signals from a microcontroller (or DSP) for volume, tone and balance control.

The platform preferably includes a crystal 186 for controlling the clock frequency of the DSP, and may include another crystal for controlling the clock frequency of the microcontroller. Instead of a DSP
Supply clock signals to the microcontroller or vice versa.

The flash memory 140 contains audio files, for example, audio content that can be heard by the end user. The audio files are typically stored as separate files for each song, which may be in different audio formats such as, but not limited to, MP3 and AAC. The flash memory also contains program files for decoding each type of audio format, as well as controlling other operations associated with the method of the present invention. These files are typically stored with a file extension indicating their format. Thus, the system controller can then recognize the format of the song that the user wishes to play and recognize that the decoder program currently loaded in the local memory associated with the DSP is of a different format; The system controller then discards the "old" decoder program and reloads the DSP's local memory with the decoder program for the format of the file desired to be played. For each such data file stored in flash memory, the audio data is typically stored as a continuous sequence of data in an adjacent memory arrangement. Therefore, sequential data in memory
The stream is accessed and retrieved for decryption and / or decryption, which addresses the corresponding sequential memory address location and the output data width of the particular memory element employed in the flash memory. Is performed by retrieving a data block having a size corresponding to.

External personal computer (PC) 9
0 may be suitably connected to the platform 100. In this way, the audio file is stored in memory 1
It can be loaded into PC 40 directly from a PC or into memory 140 via microcontroller 120. Instead, store audio files in memory 1.
It is also possible to employ a PC to load into the platform 40, where the memory 140 is
0 and the audio file is stored in memory 14
It is implemented by being inserted into special hardware interconnected to a PC for downloading into a PC.

Referring now to FIG. 2, there is shown a high-level block diagram of a software hierarchy 200 suitable for implementing the present invention on the hardware platform 100 of FIG. More specifically, FIG. 2 shows data stream 2 stored in flash memory 140.
05 is shown. Fast-forward controller 210 is a programmable address generator that moves a sequence of data from memory to temporary data buffer 220 when operating in "normal" playback mode.
The fast-forward control device 210 operates in the fast-forward mode in response to the user performing an appropriate fast-forward (or rewind) operation on the audio player.

When operating in the fast-forward mode, the fast-forward controller divides the sequential data stream into large chunks of data (blocks of data), which are separated in time. The time interval between large chunks is determined by the desired rapid traverse speed. This speed is preferably set or selected via software or other control signals.
The fast-forward control 210 is preferably a microcontroller 120
It is implemented as a program executing on the above; however, as will be apparent, this program also executes on the DSP 110. In each case, the fast-forward controller moves data from memory to temporary data buffer 220.

Seam detector and synchronization block 230
Detects transitions between each of these large chunks in fast-forward (or rewind) mode, detects when a large chunk switch occurs (e.g., a seam in the data stream), and immediately proceeds with each seam. Resynchronize later (eg, find sync word and its associated header). The seam detector and synchronization block 230 maintains synchronization and performs resynchronization whenever a break or seam occurs in the data stream. The details of such operation execution will be described below. Once the data stream is synchronized, the data is sent to detector 240, which detects audio data and outputs audio data stream 25.
Provide 0.

Referring now to FIG. 3, there is shown a simplified flow diagram 300 for the fast forward (or rewind) method of the present invention. More specifically, the fast forward controller 210 jumps forward (or backward) within the data stream 310 stored in memory. A large chunk of data is sent to temporary buffer 220.
The fast-forward controller jumps forward again and subsequently sends another large chunk of data to the buffer. This continues in a progressive manner. Next, the seam detector and the synchronization block 23
A zero searches for the sync word 320 in at least a portion of the first large chunk of data in the buffer. If the sync word has not been placed, the sync word of another data of the first chunk 230 from the buffer is examined.
If a sync word is located, the header is examined to determine 340 whether a CRC check has been performed. If no CRC check is performed, the data is sent to decoder 240 and decoded 350. Otherwise, a CRC check is performed and the generated checksum is evaluated. If the CRC check passes, the data is made available at decoder 240 and decoded 350 based on the audio format of the data file. If the CRC check fails, a new search for a sync word is started 320 to find another header.

Referring now to FIG. 4, a simplified flowchart 400 for the synchronization method of the present invention is shown. More specifically, as can be seen from FIG. 4, data from the temporary data buffer is first searched or analyzed for its sync word. If no sync word is found, more data is retrieved from the buffer and the search for the sync word continues. If the sync word is not located, the associated header is parsed to determine some of the things. One criterion is where in the data stream the next sync word should be located. This is then used to determine the frame length. Other decisions are sampling rate, bit rate, profile,
Performed on layers and identification fields.

Subsequently, a jump is made to the next header via the corresponding synchronization word. The next header is then parsed and the same information as described above is determined. Next, the consistency of the information from the two headers is compared. A return jump to the first header is performed in parallel with the consistency comparison. If the information from the two headers is consistent, a check is made as to whether a CRC check needs to be performed. If the information is inconsistent, the data
The data from the stream is advanced and the search for the first sync word and header is started again. In a similar manner, CR
If a C-check is performed and the checksum is incorrect, the search for the header associated with the first synchronization word is started again.

If the CRC check passes, the data is supplied to the decoder. This type of double header search is particularly useful for formats such as those associated with the ACC format. The ACC frame format is shown in FIG. As can be seen from FIG. 6, the frame contains an initial synchronization word, immediately followed by a header. The header is followed by a portion of fixed length data. This fixed-length data is followed by a packet of variable-length data. As indicated above, information about the length of the frame is included in the header.

FIG. 7 illustrates a frame format of the MP3 audio format. Again, a header follows immediately after the sync word. However, MP3
With respect to the format, the data is partially or completely present "before" the header. In FIG. 7, the header N
Is in the "middle" of its associated data packet. Therefore, header N has a pointer to the location where the data starts (main_data_begin). However, until the header N + 1 is decoded,
It is uncertain whether the data of this data ends and the packet N + 1 starts. Therefore, it is often necessary to evaluate two headers to determine the packet length. The remainder of FIG. 7 illustrates a situation where three headers must be decoded to determine the packet length of packet N + 2. The frame is extended from one sync word to the next sync word,
Note that the data for a particular packet is spread over one or more frames.

Referring now to FIG. 5, there is shown a simplified flowchart 500 of the synchronization method of the present invention modified for the MP3 format. "TotalamountMainData" is a variable, which adjusts (removes) the header and sync word
Serves to capture the length of the later data packet. It is initialized to zero at the beginning of the first sync word search. Similarly, the variable “mainDataThisFrame” stores the data length in the frame being analyzed or determined. FIG. 5 adds the step of determining these variables for the header and frame.

For encrypted audio data, an outer layer has been added to the inner layer of the data that needs to be decrypted, as indicated above. These layered or multiplexed data streams are
Each is decrypted and decrypted using a serial decoder that employs this technique.

With respect to seam detection during resynchronization, the present invention provides a "candidate" that correctly identifies the location of the subsequent frame header.
Requires a header frame. For the general case where the input data stream data is sent to the decoder in turn, the data spanning between these two headers is stored in memory. In the case of a hierarchical data stream with large outer frames, this results in an unacceptable increase in memory costs beyond what is required to decode the inner data stream. In addition, as the data stream is sent to the decoder at a finite data rate, the extra memory increases the delay before the external decoder makes a decision, which is noticed in the decoded audio. Yet another feature of the present invention is a method for minimizing memory and delay increases in the case of a hierarchical data stream. Next, the details of this method will be described.

For a hierarchical or multiplexed data stream, the sync word search for the outermost encryption layer is relaxed to a single sync word and associated header search. This is possible in situations where encryption is weak, i.e., the data immediately after the encrypted header is encrypted, but the remaining data in the frame is unencrypted.

The present invention can be realized by software, hardware, or a combination of hardware and software. Having described the invention and its features in detail, it will be appreciated that various changes, substitutions and substitutions can be made without departing from the spirit and scope of the invention as defined in the appended claims.

With respect to the above description, the following items are further disclosed. (1) A method for providing audible fast-forward or rewind for an audio format data file: jumping forward or backward in the data file by a preselected amount, and its length Switching of the stored chunks of data to assemble the chunks of data that are at least a function of the sampling rate and the bit rate, store the chunks of data, and look for data packets with data. Analyzing the data packet, selectively decoding the data packet, and playing back the data packet.

(2) A method for determining a decodable data packet: scanning and evaluating a synchronization word of a data stream, evaluating at least one header indicating a data position, and indicating an index on the data. Applying the method.

(3) Apparatus for providing audible fast-forward or rewind of an audio format data file: repeatedly skipping forward or backward within a compressed audio data stream, and blocks of data And then retrieve these data
A fast-forward controller for splicing blocks back together and a decoder for decoding each data block, detecting when a block switch has occurred (a seam in the data stream), and quickly resynchronizing at each switch. Said apparatus comprising: a decoder.

(4) This disclosure is not limited to compressed formats, such as, but not limited to, MPEG
Describes a method for performing audible fast-forward or rewind of audio content expressed in a-1 layer 3 (MP3) or MPEG-2 advanced audio code (AAC). In this technique, a fast forward control (210) is employed, which skips forward or backward repeatedly in the compressed audio data stream forward or backward, and blocks blocks of data (22).
0) and perform the splicing of these data blocks back to each other. Subsequently, the decoder (24
0) is used to decode each of these blocks, to detect the occurrence of a block switch (a seam in the data stream) and to quickly resynchronize at each switch. The layered or multiplexed data stream is decoded using a serial decoder, each employing this technique. The decoder implements resynchronization and error recovery using a novel robust synchronization search in the present invention.

[Brief description of the drawings]

FIG. 1 is a block diagram of a hardware platform capable of playing digital audio files and implementing the method of the present invention for an individual end user, eg, a portable audio player. Show.

FIG. 2 shows a high-level block diagram of a software hierarchy suitable for use on the hardware platform of the present invention shown in FIG.

FIG. 3 shows a simplified flow diagram of the fast forward (or rewind) method of the present invention.

FIG. 4 shows a simplified flow diagram of one synchronization method of the present invention.

FIG. 5 shows a simplified flowchart of the synchronization method of the present invention applied to the MP3 audio format.

FIG. 6 shows a block diagram of a frame structure for the AAC audio format.

FIG. 7 shows a block diagram of a frame structure for the MP3 audio format.

[Explanation of symbols]

 90 Personal Computer (PC) 100 Platform 110 Digital Signal Processor (DSP) 112, 114, 116 Program 120 Microcontroller 130, 132 On-Chip Memory 140 Flash Memory 150 Battery 160 Power Block 200 Software Hierarchy 205 Data Stream 210 Fast forward controller 220 Temporary data buffer 230 Seam detector and synchronization block 240 Detector 250 Output audio data stream

Continued on the front page (72) Inventor Alexey, Robinson United States of America Texas, Dallas, McComas 5830 (72) Inventor Jonathan Elle, Rowlands United States of America Massachusetts, Somerville, Webster Avenue 25 Number 303 F-term (reference) 5D044 AB05 DE03 FG09 FG24 GK08 5D045 DA20 DB01

Claims (2)

[Claims] 1. A method for providing an audible fast forward or rewind of an audio format data file, comprising: jumping forward or backward in a data file by a preselected amount; To assemble a large chunk of data that is at least a function of the sampling rate and bit rate, store the large chunk of data, and find a data packet with the data:
Analyzing the stored data chunk transitions, selectively decoding the data packets, and playing back the data packets. 2. Apparatus for providing audible fast forward or rewind of an audio format data file, comprising: repeatedly skipping forward or backward in a compressed audio data stream to block blocks of data. A fast-forward controller for retrieving and subsequently splicing these data blocks together, decoding each data block, detecting when a block switch occurs (a seam in the data stream), and switching each And a decoder for quick resynchronization at the same time.