At present, there is an extensive discussion about the MP3 and other similar formats online. Sound compression utilities are also subjects of heated debates like which the better coder is: Fraunhofer or LAME, as well as whether or not there are obvious differences between 256 and 320 Kbits/s. [no_toc]
While there are proponents of the MP3 format, other alternative music storage methods also have their own fans. LQT or OGG Vorbis, for instance, are also popular. However, at 9 years old, the MP3 still has no viable heir, and everyone is waiting.
When considering the improvements, 2 points of view must be considered: the increase in sound quality after file compression at a specific bitrate and the compressed files’ smaller size while retaining the quality. Now, let us examine the goal of the developers here.
The Situation Before
You probably know that data transfer size has a lot of bearing, and the size of audio files is usually staggering. What do you do then? You have the option to use the Microsoft-promoted WMA format that comes with bitrates that are comparable to MP3 in terms of quality, (although limited to a mere 192 Kbits/s) reduces the size of the file and is also lacking in ID tags; and are copy-protected. On the other hand, there is the successor to the all-time favorite MP3, and that is in the form of the MP3Pro.
The MP3Pro comes from Coding Technologies, Inc., a developer established back in 1977. The company is involved in marketing and development of SBR technology-based codecs. SBR stands for Spectral Band Replication.
Working in collaboration with Thomson Multimedia and Fraunhofer Institute, Coding Technologies is backed up by notable investors like Heinz Gerhauser, the Fraunhofer Institute head. Thus, CTI is privy to all Fraunhofer projects. In fact, MP3Pro was a name provided by Thomson Multimedia that was, together with RCA, involved in the promotions of the product.
The Fine Details
MP3Pro is an SBR technology-based MP3 codec. The technology was borne out of the necessity to transfer digital music via the Net in real time, and it was developed for mobile computers and various types of portable digital players.
A small-sized memory or a limited data rate allow the use of only low bitrates when compressing sounds into AAC or MP3 formats. Faster modes of connection like xDSL and ISDN don’t offer a constant data flow since the Internet is often overloaded.
Decent quality can more or less be achieved through 128 Kbit/s or higher compression rate. Issues sometimes arise with low bitrates. For instance, it may become necessary to lower a frequency range for audio data transfer, or artifacts may sometimes arise following encoding. This is a proof that merely being a psychoacoustic model isn’t sufficient when working with low bitrates such as 128 Kbits/s and below. SBR technology is designed to narrow the frequency range a bit during the encoding stage by cutting the highs off. The frequency range can then be brought back to its previous size during decoding. This is based on the provided information about low frequencies.
For testing purposes, we used various synthetic fragments and styles. First, the original WAVE file was coded using each codec. It was later decoded back to WAVE, then cleared using the wave editor. After that, a comparison between the original’s AFCs and the final WAVE files.
Microsoft’s WMA format was used since it is considered as the MP3Pro’s direct competitor. More than that, however, the WMA 8.0 is quite interesting to study and take a closer look at.
The tests involve comparing the MP3 128 Kbits/s with the WMA and MP3Pro, both at 64 Kbits/s. At 96 Kbits/s, the MP3 192 can be pitted against the 96 Kbits/s of WMA and MP3Pro.
Music types used for testing include:
- Modern dance music
- Jazz that feature live performances
- Pop music with vocals
- Satisfactory sound quality at lower bitrates
- High degree of compression
- Low system requirements
- High frequency syntheses from low frequencies
- Insufficient support for high bitrates