
Audio (audio) compression comparison [mp3, wma, ogg, atrac] Part 2

[Sound source used and points of interest]
・ 1kHz sine wave
Check for noise or correction. Investigate if abnormal sounds are mixed by emphasis or noise different from the originally generated range.

· White noise
Check the frequency characteristics. Use sounds that are emitted at the same level for all sounds from 0 to 20 kHz and see if they are reproduced correctly.
·music
Use real music and investigate the differences with the original.
[Bitrate Settings]
Fixed bit rate: 96kHz, 128kHz, 256kHz, 320kHz Variable bit rate: 96-160kHz, 192-320kHz.
However, depending on the software, 320kHz cannot be set fixed and 350 can be set, or the upper and lower limit bits cannot be specified in the variable, and the sound quality standard can be specified in several steps ( medium sound quality, high sound quality). quality).be. Also, there are some that are configured with average bitrate instead of variable bitrate, so understand that it’s not a completely fair comparison.
[Software used [encoder]]
・ MP3 system
Afternoon Koda Ver.3.11a [gogo.exe ver.3.11]
Lame Ivy Frontend Encoder Ver.2.91 [Lame.exe Ver.3.93]
B’s GOLD Ver.7.12 [Unknown]
RipAudiCO Ver.3.70 [leme_enc.dll Ver.3.93]
・OGG system
oggdropXPd Ver.1.7.11 [Unknown]
B’s GOLD Ver.7.12 [Unknown]
・ WMA
B’s GOLD Ver.7.12 [Unknown]
(For WMA, I tried 3 types of software in my environment, but the result was exactly the same (maybe the encoder itself uses the same thing?) And it corresponds by software Since the bitrate range was narrow, only used a type).
・ ATRAC
nothing special. For ATRAC, we recorded analog from a CD player to an MD deck, optically connected an MD deck to a PC, and measured what was captured by WAV.
· To measure
Wave Space Ver.1.31
【others】
Although it is different from the main theme, I converted it to WAV for the visual measurement of each standard (because WaveSpace only supports wav), but the position where the sound of the WAVized data ends and the total playback. We discovered that there was a difference in time. , so we also investigated it.
3.Hardware 3.
Originally, the equipment used should be described in detail here, but the hardware environment is different for each individual, and this survey is only a guide in the first place, and it will be different if other people do the same. is a possibility of results, I will omit the detailed description of the hardware. (The thing is that I don’t have enough equipment to publish)
【Results of the test】
See the following for a summary of the results of each survey.
・White noise measurement result
・1kHz sine wave measurement result
・Music measurement results
・Simple file size and comment list
[Discussion]
ah There seems to be no big difference in file size (between the same bitrate)
stomach. Sound quality appears to be MP3 < WMA < OGG at low bit rates
(MP3: 128 = WMA: 96 < OGG: 96).
Hare. There is little difference at high bit rates
(there is a slight difference in the treble range, but it seems you won’t notice the difference unless you’re in a very good environment).
Worker. The difference in the encoder software was more than I expected
(especially in MP3)
“My conclusion”
[Less than 128]
If you’re worried about popularization (compatibility), [WMA] is good, and if you basically use it alone, [OGG] is good.
(I am worried about the amount of noise or the correction, but I sacrificed a bit on the sound quality anyway, so I chose the one that covers up to the high range as it is. Also, due to the relationship between ① and ②, mp3 is another with the same sound quality.The file size will be larger than
[With 256 and more]
The variable bit rate (192-320) of [Afternoon Koda] is good.
The fixed 320 is good for sound quality, but there is little difference between the fixed 256 and variable high-quality sound, and it seems that you can barely understand it even if you listen to it. If the sound quality is about the same, the smaller the file size, the better.
[Other impressions]
About OGG
I had high expectations for OGG, but I was concerned about the measurement result at 1 kHz, whether it was noise or correction. However, I find the relationship between sound quality (wide playback band) at low bit rates and file size to be excellent. At high bit rates the sound quality and file size are about the same as MP3s so I think MP3s are advantageous considering the penetration rate but I think they are doing pretty well considering the fact that they have just been developed. expected in the future






