Send the link below via email or IMCopy
Present to your audienceStart remote presentation
- Invited audience members will follow you as you navigate and present
- People invited to a presentation do not need a Prezi account
- This link expires 10 minutes after you close the presentation
- A maximum of 30 users can follow your presentation
- Learn more about this feature in our knowledge base article
Do you really want to delete this prezi?
Neither you, nor the coeditors you shared it with will be able to recover it again.
Make your likes visible on Facebook?
Connect your Facebook account to Prezi and let your likes appear on your timeline.
You can change this under Settings & Account at any time.
MP3 Vs. AAC (Media Formats)
Transcript of MP3 Vs. AAC (Media Formats)
Ipod ect. So what's the difference? References What compression does. How a basic compression encoder works a basic encoder will have 4 key areas in the audio chain. Compression is the act of reducing the data of a file. Less data = Less storage space required This means that we can store more data than before More stored data = a lot more music tracks on your player What is Quantization? the bit rate determines the amount of quantization levels that can be performed. the higher the number the more levels will be taken more often resulting in a much clearer digital 'picture' of the original wave. Making it smaller whilst trying to keep it the same. Filter bank Perceptual model Quantization and coding Encoding of bitstream How a basic compression encoder works What is the analysis filter bank? used to decompose the input signal into subsampled spectral components. Along with the filter bank in the decoder it forms an analysis/ synthesis system. How a basic compression encoder works What are Quantization and coding? How a basic compression encoder works what is the encoding of a bitstream? How a basic compression encoder works What is the Perceptual model? by using either the time domain input or the output from the filterbank, an estimate is made of the actual masking threshold which is worked out but using rules known from phychoacoustics. spectral components are coded and quantized with the aim of keep noise down below the masking threshold, the noise is introduced in the quantizing phase. this can be done a number of ways: from simple block companding to analysis-by-synthesis systems using additional noiseless compression to assemble the bitstream a formatter must be used, this typically consists on coded and quantized spectral coefficients and side information such as bit allocation information. What is Quantization? Common artifacts Distortion Quantization noise Pre-echoes Double-speak Pre-echo Pre-echoes are a very common artifact for perceptual coding systems running at high frequency. Quantization noise This is introduced at the quantizing part of the coding.
Quantizing is only really audible if the bit rate is not high enough. the absolute minimum is double the highest frequency that will occur. Double-speak Occurs most at low bit-rates and lower sampling frequency when there is a mistake between the time resolution coder and the time structure of some signals. Analysis filter bank. Perceptual model. Quantizing and Coding. Bitstream encoding. The artifact is noise spread out over some time even before the music event causing the noise. Most noticeable for speech signals when listening with headphones. A single voice sounds like the same recording was done twice and overlayed. This means that it follows the concepts of the human ear. It has 2 basic rules to follow for it to be effective. 1: Remove frequencies that go above or below the human hearing range ( 15/20Hz - 20kHz) 2: If two sounds occur at the same time the louder one will only be heard (this is called simultaneous masking) so it can remove the quite sound. MP3 was created in 1991 and published in 1993. MPEG stands for Motion Picture Experts Group. Fig 2: Anon (online) viewed on 6th February 2013 http://www.digikey.com/Web%20Export/techzone/microcontroller/article-2011december-audio-coding-and-compression-fig1.jpg Fig 1: Anon (online) viewed on 6th February 2013
http://www.tufts.edu/programs/mma/mrap/analogdigitalhistory/D2quantization.gif Fig 2 Fig 1 Compression amount can be set by user Sharing files is easy -It deletes the lesser audible content. Data is susceptible to virus attack. Due to easy of copying the piracy market uses it a lot. Portable music players. Copying songs (while this still reduces the quality of the song, as it is being compressed each time it is copied) Became a standard in 1997 Quantization is how a audio wave is transformed into a digital form. The wave will be sampled throughout the time. The higher the sample rate the more 'steps' will be created. Lossy form of compression AAC has an optional backward prediction, this helps to achieve better coding efficiency for tone-like signals. Types of compression Lossy Compression Lossless Compression Lossy Compression To make the file smaller than the original during the compression stage unnecessary data is deleted. Lossless Compression Reduces the size of the data with out deleting any of it. To do this it re-writes the data as simple code and then de-codes it when needed. Types of Masking Temporal Masking Simultaneous Masking Temporal Masking Simultaneous Masking Simply put simultaneous masking occurs when two sounds happen at the same time in close frequency proximity to each other, however you will only ever be able to hear the louder sound because the basilar membrane will be dominated by the louder frequency and thus you will only perceive that frequency. Huffman Coding Huffman coding is a simple way of compression. The basilar membrane in the resonant cavity in our ear cannot start or stop vibrating quickly. because of this a quieter sound just before or after a louder sound will not be heard as it will have been masked, this occurs the pre and post masking phase. Types of compression Types of masking Hoffman coding This technique is a lot more desired as none of the data is lost however because of that fact file sizes will never be able to get as small as lossy. What it does is get the two lowest values and add them together creating a new value, this continues until all the values that can be added together have been. Once it has done this it goes up a layer and repeats the process added the lowest values together. This repeats until you get 100%. Once this is done you then go down the 'tree' labeling it either 0 or 1.
eg, 0 on all the 'branches' that go left, 1 on the ones that go right.
This means that 'B' would have
a value of '100', and 'F' would be
'11101'. AAC is an improved design to go over MP3. While MP3's use a filter bank, AAC uses discrete cosine transform to find the spectral elements of the audio. Discrete Cosine Transform (DCT) Huffman Coding Huffman coding is used at the end of the coding sequence. As you have seen the way it codes the symbols, the ones with higher probability have a shorter code enabling the code to be reduced in length. The DCT divides the signal into blocks of differing importance. It then scans these 'Blocks' or 'Windows'
and works out the different values of all the blocks inside it.
once this is done it just sends the value of these blocks. AAC MP3 Only Predominately used by Apple Uses DCT to encode, meaning a lot more accuracy Uses DCT to code signal Has Backward Prediction Can Have Improved Huffman coding. Compression amount can be set by the user Fig. 3: Anon, Viewed on 6th February 2013 https://www.google.co.uk/search?hl=en&sugexp=les%3B&gs_rn=2&gs_ri=hp&tok=wWDVhhbH3MvvJWJSj6vhEw&cp=16&gs_id=3&xhr=t&q=temporal+masking&bav=on.2,or.r_gc.r_pw.r_qf.&bvm=bv.42080656,d.d2k&biw=1422&bih=703&um=1&ie=UTF-8&tbm=isch&source=og&sa=N&tab=wi&ei=w1EWUcy1AaWV0QXvnYCQBw Fig 4: Viewed on 7th February 2013
https://www.google.co.uk/search?hl=en&sugexp=les%3B&gs_rn=2&gs_ri=hp&tok=rahrxgjd8oXKgvIrJQleeA&cp=20&gs_id=3&xhr=t&q=simultaneous+masking&bav=on.2,or.r_gc.r_pw.r_qf.&bvm=bv.42080656,d.d2k&biw=1422&bih=703&um=1&ie=UTF-8&tbm=isch&source=og&sa=N&tab=wi&ei=dlEWUdDLOeqj0QXIsIGgDg References Anon, Advanced Audio Coding, Online - Viewed Fri 8th February 2013
http://www.digitalpreservation.gov/formats/fdd/fdd000036.shtml Karlheinz Brandenburg, MP3 and AAC explained, online Viewed 6th February 2013.
http://graphics.ethz.ch/teaching/mmcom12/slides/mp3_and_aac_brandenburg.pdf Fig. 3 Fig. 4 Fig. 5 Fig. 5: viewed on 7th February 2013
https://www.google.co.uk/search?q=huffman+coding+explained&hl=en&tbo=d&source=lnms&tbm=isch&sa=X&ei=guoTUceSI8qa1AXUvICwBA&ved=0CAoQ_AUoAA&biw=1422&bih=703#hl=en&tbo=d&tbm=isch&sa=1&q=huffman+coding+tree&oq=huffman+coding+tree&gs_l=img.3...0.0.0.155719218.104.22.168.0.0.0.0.0..0.0...0.0...1c..2.img.csrM3twKLrE&bav=on.2,or.r_gc.r_pw.r_qf.&fp=6a7669599341ce96&biw=1422&bih=703 References Anon, What is an MP3, BBC, Online viewed 7th February 2013
http://www.bbc.co.uk/webwise/guides/about-mp3s Anon, What is an AAC File, Online viewed 8th February 2013
http://ipod.about.com/od/glossary/g/what_is_aac.htm Anon, Advanced Audio Coding, Online viewed 7th February 2013