Dragon angel]
Speaking of mp3, I'm afraid no one will say that they haven't heard of it. Even if you are not an mp3 user, you will be deeply impressed by the ubiquitous advertisements, publicity activities, discussions among friends and rich resources on the Internet. For fashionable young people, especially friends who like music and digital devices, mp3 is probably a word to be memorized every day. But what is mp3? How to determine the sound quality of MP3? How to listen to high-quality MP3? I think the following article can help you solve many problems.
Looking at the current mp3 users, the recognized universal production standard is eac ripping +lame compression. I also use this to match. In this production process, experienced friends will explore some tips. Different music uses different parameter settings and compression ratios, ranging from the standard 128kbps to the maximum of 320kbps. But what are the differences and differences between these bit rates? Is the most suitable compression ratio. Which is better, cbr or vbr, and so on. These problems are often discussed by everyone. In order to make these details clear, I specially made a targeted experiment. Let me share some feelings with you.
I like listening to classical music very much, so Bach's track 1: Munich Bach Orchestra. Trajectory capture software eac. Compression software cd`ex. Play the software fooba2000 v0.8 version. Audio-visual headphones are Shure's er6 and E3. Because there are many details in classical music, large bands and high quality requirements in all aspects, it can obviously reflect the differences in details between different processing methods.
(A spectrum comparison chart will be provided later.)
I grabbed the tracks with rac first, and then processed the wav files with lame MP3 encoder (Vision 1.92 Engine 3.92) engine in cd`ex software. I tried the lame parameters one by one to choose a good result:
The first parameter, thread priority, is the highest and lowest respectively. When other parameters are the same, comparison is compressed. It is found that thread priority has no effect on sound. The generated files are the same size. Compared with the sound, this parameter has no effect on the sound quality.
The second parameter is the version. There are mpegI.mpegII and mpegII. You can choose v. Other parameters are the same. They were compressed three times with these three options. By comparison, although the file sizes of the three methods are the same, the actual listening sense of mpegI is better. The compression ratio of middle and low frequency is slightly smaller, but the high frequency distortion is slightly more. It is more suitable for listening to human voices and pop music. Listen. The classical mpegI type is also good, and the sound base is better. However, if it is high-frequency solo music such as violin, mpegII.v can be recommended for better effect.
The third parameter is the most important. This is the code rate. Choosing it will directly affect the size and hearing of your mp3 file. High compression ratio will cause great distortion, while low compression ratio will cause little distortion. But how can we find a balance that is acceptable to both sides? This needs to be explored carefully in the experiment. Considering that the sound quality of files with low bit rate is not suitable for playing music, the minimum is 128kbps. Four fixed bit rate files (128.192.256.320 * *) are used for comparison and testing.
The compression ratio of 128kbps is still rough. After compression, the distortion of high frequency part is very obvious. It sounds hollow, shriveled and harsh. There are often flashing sounds. These instruments are of poor quality. Some musical instruments have changed their taste. You may even have misheard. The compressed volume of a piece of music at 3: 39 is 34 14kb. Although the volume is not big, the sound is not satisfactory. There are great defects.
Compared with 192kbps, the compression effect of 128 is much better. First of all, the sound is three-dimensional, at least there is no sense of emptiness, and the high-frequency distortion is much smaller. The sound is compact, the noise is small and clean, and the ideal listening effect is achieved. It's just that the compression is still quite strong, and the details are still not very good. The texture of musical instruments, especially wind instruments, is still hard, unreal and lacks musical sense. 5 123kb。 I think this compression ratio is better for an mp3 player with a capacity of 128-256m. It can not only meet the basic hearing, but also be the right size. 128m can store about 95 minutes of music, and 256m can double to 190 minutes of music.
Compared with 192, the compression rate of 256kbps is naturally higher in sound quality. Take Quxian 10 second as an example. The low-frequency graininess of the cello part is much smaller. The sound is smoother and more natural. Texture and texture are clearer. There are many details. The rendering of the atmosphere is more prominent. The voice rotation in the following tracks is also more incisive. The clarity of large and small signals is also improved. Some. The voice is more detailed and pleasant to hear. But at the same time, the file size suddenly increased to 68,365,438+0 KB, which is still affordable for a 256m mp3 player. Through calculation, it is not difficult to know that under the condition of code rate of 256, about 135 minutes of music can be stored. On the whole, it is enough. 128m is a little less.
320kbps is the maximum bit rate that lame can provide. The final generated file is 8592kb, which is about 8.4M larger than the 37M of wav file, and the compression ratio is basically 4.5: 1. However, the generated mp3 file sounds very distorted. Compared with other code rates, the natural advantages of 320 are obvious. The timbre, details and so on are exquisite. Basically, the sound quality of the cd has been copied as it is. Don't sound like it on a cd player with an mp3 player. There is basically no difference. But I use high-resolution high-end earplugs. Coupled with my experience and ability in music and equipment, I still hear a lot of differences compared with wav files. First of all, compressed mp3 sounds a little shriveled. It sounds lively and dynamic without wav files. The ending details, overtones and sense of space are not as high as wav files. However, the timbre is quite close, but the expressive force is poor. Digital tastes strong. So if you are using a mini hdd player such as an ipod, I still recommend you to use 320kb. Compression ratio of ps. This can get the best hearing. Of course, it is best to listen to wav directly-without compression. No loss. Unfortunately, there is no walkman to support ape lossless compression. Otherwise, there are many choices.
What I said above is the compression ratio of fixed bit rate. In lame, this is called CBR (constant bit rate). In fact, the biggest feature of lame is to provide users with a variable bit rate VBR compression method. This method will automatically reduce the bit rate and reduce the file size during some pauses. This is a very good coding method. But how to choose the lowest and highest bit rate range of VBR to get the most suitable file and sound quality? This is another problem that needs to be solved by experiments. Similarly, considering 128kbps as the basic value, we choose 96kbps to 160kbps for compression. The compressed file size is 380 1kb, which is only 387kb more than CBR's 128kbps. But the sound has definitely raised a big level. At least half as small. Although there is a lot of noise in the details, the first hearing is still much better than 128kbps. The average bit rate after compression is 147kbps, which is also very space-saving. Later, experiments were carried out from 96kbps to 192kbps.96kbps to 224kbps.96kbps to 320kbps. It is found that they are similar to the maximum cbr compressed sound quality, that is, the sound with vbr from 96kbps to 192kbps is similar to CBR, but the former is 448 1kb and the latter is 5 123kb. Therefore, vbr is really useful as a compromise method to pursue high sound quality and save space. Of course, on the other hand,
Fourth, the mode parameters are three-dimensional. J stereo. Force stereo and mono. The contrast test shows that the standard stereo has the best effect. Although the compressed file size is the largest, considering the small file size difference, sound quality difference and hearing sense, stereo is still ideal.
Fifth, compression method. There are two kinds of software, such as vbr-old.vbr-new, but only the first two are easy to use. Comparing the old and new vbr coding methods, it is found that the old one is more delicate in sound quality, but the old one has a slow compression speed, almost 5.6 times that of the new one. It takes about 3.4 minutes to make a song, which is very difficult to use, and the file size is 6540.
Mp3 bar is not a word invented out of thin air, but an abbreviation of a technical term, which is the abbreviation spelling of MPEG- 1 AudioLayer-3. What does this mean? Let me explain to you: MPEG- 1 AudioLayer-3 is the third layer protocol of the audio part in MPEG- 1 international standard technical protocol. It describes an audio format. Does it look complicated? It doesn't matter if you say it bit by bit. First of all, the word mpeg-is the abbreviation of Moving Picture Exp-erts Group, which means moving image compression processing group. This group is quite powerful. It specializes in developing dynamic still video (including audio). Almost all international technical standards. We use their research results from TV to movies, from vcd to dvd. Mp3 is a part of their mpeg 1 protocol. The audio part is on the third floor, so it's called mp3. As far as I know, the birth of mp3 was discovered unconsciously by the staff of mpeg Group. It is produced as a method of capturing cd audio tracks. At most, it was an accessory product in the experiment at that time. Let's go and have a look.
First of all, the size is small: different file sizes can be obtained according to different compression ratios. However, compared with the original wav format, the size is much smaller. The file size is small, the storage space is small, and the size and cost of hardware devices used for playing will be significantly reduced, so there is a good market. In addition, mp3 files are downloaded everywhere. Compared with cd player users, it is naturally a saving to keep buying software. Again: use. Whatever you want. A little player can easily carry it with him. I can use it at will. Moreover, because the power consumption of the circuit part of the player is relatively low, the working time of a single battery is also long, which saves a lot of trouble in replacing the battery. In addition, users can listen to their favorite songs together at will, without being subject to the molding software of audio-visual publishers. There is a lot of freedom. This reminds me of a slogan that Jay Chou advertised for the M-Zone [My Site]. I'm in charge! "That's cool! !
Having said so many related words, everyone must be impatient. Let's go deep into the main body and thoroughly analyze all aspects of mp3 for everyone.
First, the basic knowledge:
Digital compressed audio and mp3 technology;
To say mp3, I have to say digital compressed audio first. Well, digital compressed audio, as its name implies, is compressed digital audio. But what is digital audio? What is compression? Here are some professional explanations. Interested friends should look carefully!
In the digital world of computers, sounds are all stored by digital coding, which is different from analog audio in traditional life. Because computers can only record numbers of 0 and 1, analog audio must be sampled quantitatively. According to Nyquist sampling theorem. Sampling at twice the frequency of sine wave can completely and truly restore the waveform. Therefore, the sampling frequency of digital recording wave is directly related to its highest recovery frequency index. For example, sampling at the sampling frequency of 44. 1KHZ can restore the highest frequency of 22.05KHZ, which is slightly higher than the hearing limit of human ears. Therefore, a/ D conversion usually uses a sampling frequency above 44. 1KHZ. However, due to the sampling frequency (usually in Hz) and sampling ratio (usually in bits), the file size of acoustic data is different. The higher the sampling frequency, the more storage space is needed. The higher the sampling rate, the larger the required storage space. This brings difficulties to practical operation. Therefore, the emergence of compression technology becomes inevitable.
Digital compression refers to "slimming" the sampled original digital audio file. By using some effective algorithms and methods, redundant information in files can be removed. This can reduce the volume and make it easy to use. The specific operation is too complicated and professional (even I am a student of computational mathematics). So I won't explain them to you one by one. Here is a brief explanation of MP3 compression coding technology involved in compression: MP3 compression coding uses five important technologies: minimum auditory threshold, masking effect and bit storage slot. JointStereo and huffman coding are combined. Through the computer calculation of these technologies, we can greatly reduce the file size of digital audio. The new audio format generated after compression is the familiar mp3.
2. Compression principle:
No matter what technology, we must have our own reasons. This is especially true for audio technology, because it is directly related to our hearing. Why does the compressed mp3 sound very close to the sound of a cd? This has a lot to do with people's physiological structure. Experiments show that the frequency range (audio) of sound that humans can hear. It's 20 Hz-20 Hz. But the human ear does not respond directly to the sound in the whole audio frequency band. 2-5kHz is the most sensitive frequency band for human ears. According to its characteristics, the whole audio frequency band is divided into several critical frequency bands. Because human auditory system distinguishes sound energy according to frequency, small sound of any frequency will be covered by loud sound in critical frequency band because of masking effect. MP3 doesn't quantify it, thus removing what the human auditory system doesn't have. It can be seen that mp3 is a lossy audio compression coding. So no matter how high the bit rate, MP3 is lossy compression. The sound quality is lower than that of cd, but it is more or less acceptable. It varies from person to person.
3. Sound quality after compression:
This is the topic that everyone is most concerned about. What standard can the compressed sound reach? Let me give you a chart first. Let's look at the reference values in the mpeg standard.
It can be seen that due to the different compression ratios, the quality of the obtained sound file is gradually improving. 128kbps basically reaches the sound quality of cd. Is that so? My answer is: of course not. This is just an official reference data table. In practical use, the bit rate of. 128 can't express the sound quality of a cd at all. Although the software algorithm used in mp3 compression will be different, the compression ratio is the most fundamental factor that limits its performance.
Lame is not the only compression software on the market, and mp3 is not the only audio compression format. Other companies have developed many useful compression formats, but they are not as popular as mp3 for various reasons. However, as a member of the compressed audio family, I want to introduce them to you here to let you know more.
1. Ape:
APE format: audio of monkeys (
www.monkeysaudio.com
). Simply put, ape is a lossless compression format of audio. It is only half the size of wav file, but the sound quality has not decreased at all. Can restore the sound quality of the CD to the greatest extent. Of course, it also depends on the quality of rail grabbing. If the quality of rail grabbing is good, the sound quality of ape will be similar to that of CD, which is impossible for lossy compression such as mp3. Of course, the price of lossless is the increase of volume. 1CD ranges from 200mb to 400mb. Therefore, if there is enough bandwidth, ape brings more convenience for users to exchange CDs and meets the needs of many classical music lovers to transmit high-quality CDs. An 80-minute music CD is about 700 meters. It is obviously inconvenient to transmit such a large size. At this time, use Monkey ""audio- to compress WAV files into APE files of half size (even smaller). Ideally, when you put it.
Advantages: high compression ratio and good sound quality. After loading the plug-in, you can play it directly with Winamp. The sound quality is better than Mp3 or Mp3 Pro with any parameters.
Disadvantages: the compression and decompression time is too long.
2.MP3 Professional Edition
In order to reduce the distortion, a new MP3 format, MP3 Pro, appeared. It divides the whole audio frequency band into two channels, the middle and low frequency channel and the high frequency channel. The traditional MP3 encoder is responsible for the middle and low frequency channels, and the other uses SBR technology (band duplication/). The decoder is responsible for the high frequency part. Finally, both of them are played in MP3 Pro software at the same time, which makes the sound quality of MP3 with the same bit rate obviously better than that of MP3. Not only the high-frequency details are rich, but also the trembling phenomenon is not easy to detect, which is very close to the original WAV file.
SBR technology adopted by MP3 Pro is not as simple as directly separating and coding the high frequency in music, but based on analyzing the low frequency signal transmitted by the original encoder, thus reconstructing the high frequency signal.
3. Real audio
Real audio. Extended RA: This format is really the soul of the network. Its powerful compression and minimal distortion make it stand out among many formats. Like MP3, it is also to solve the network transmission bandwidth resources. So the main goal is compression ratio and fault tolerance, followed by sound quality. Therefore, we usually use this format to audition new music online.
3.WMA
WMA is a file format encoded by Windows Media Audio. It was developed by Microsoft. The goal of WMA is not the stand-alone market. It's the internet! The competitor is the famous Real Networks in the online media market. Microsoft claims that WMA can achieve sound quality close to CD with only 64kbps bit rate. Unlike previous coding, WMA supports anti-copy function. She supports protection through Windows Media Rights Manager, which can limit the playing time and times, even the playing machine. WM。 A supports streaming media technology, that is, watching while playing, so WMA can easily realize play online. Because it is a masterpiece of Microsoft, Microsoft has added support for WMA in Windows. WMA has excellent technical characteristics. With the vigorous promotion of Microsoft, this format has been accepted by more and more people. Compared with ra, copyright is probably the most annoying place for users.
4.ACC:
AAC (Advanced Audio Coding) is a technology provided by Dolby Laboratories for the music community. AAC claims that "it can accommodate up to 48 audio tracks with a sampling rate of 96 KHz, and can provide music programs with 5. 1 channel, with the quality equivalent to ITU-R broadcasting and a data rate of 320Kbps". Compared with MP3, it has better sound quality and can save about 30% storage space and bandwidth. It is a technology developed according to the specification of MPEG-2. Panasonic's mp3 products all adopt this coding method, which is of course compatible with mp3 format. I found aac is a very useful audio format when I used it myself. 128kbps aac is enough to compete with 224kbps mp3, but the space is almost half smaller. But aac and mp3 coding styles are different in space and structure. Whether I like it or not is a matter of opinion.
5.ATRAC 3/ATRAC 3 plus:
ATRAC 3 (Adaptive Transform Acoustic Coding 3) was developed by Sony Corporation of Japan. It is an upgraded version of ATRAC adopted by MD, and its compression ratio is about twice that of ATRAC. The compression principle includes simultaneous masking, aging masking and equal loudness curve, similar to MP3. The copyright protection function of ATRAC3 is OpenMG. At present, the portable player corresponding to ATRAC3 is mainly Sony's own products. However, the company signed contracts with semiconductor manufacturers such as Fujitsu, Hitachi and NEC. Rohm, Sanyo and TI manufactured and sold A in February 2000. TRAC3 uses the patent license agreement of LSI. Compared with mp3, recording cd with mdlp compression still has certain sound advantages. However, due to copyright issues, it is very inconvenient to use. ATRAC3 Plus has been further improved because of its advanced algorithm, but it is only widely used in Sony products. This is quite depressing.
6.VQF:
The so-called VQF refers to Twinvq (weighted indirect vector quantization in transform domain), which is NTT (Japan Telegraph and Telephone Company) in Japan. Audio compression technology developed by NTT man-machine interface laboratory. This technology is supported by the famous Yamaha company. VQF is an extension of its archives. The implementation method is similar to MP3. It compresses sound by using a distortion algorithm. However, compared with MP3 compression technology, it is fundamentally different: the purpose of VQF is to compress music without. It is the compression of sound. So vqf adopts a compression technique called "vector quantization". This technology firstly vectorizes the audio data, and then unifies and smoothes the similar waveform parts in the audio waveform to strengthen the sensitive parts of the human ear. Finally, the processed vector data is quantized and then compressed. I feel that VQF is much better than mp3 with the same sound quality in my own use.