Talk:WAV

From Wikipedia, the free encyclopedia

This article is within the scope of the Professional sound production WikiProject, a collaborative effort to improve Wikipedia's coverage of the technology, equipment, companies and professions related to professional sound production. If you would like to participate, you can visit the project page, where you can join the project and see a list of open tasks.
Start This article has been rated as Start-Class on the assessment scale.


Contents

[edit] Type 350

Does anyone know what this codec type is, or where I could find out? Mr. Jones 10:07, 1 Jul 2004 (UTC)

[edit] Lossy/lossless

I'm confused, this article describes PCM as lossless whereas all other articles seem to describe it as lossy (Lossy data compression, List of codecs). --gb 08:31, Feb 16, 2005 (UTC)

Don't mix up PCM and ADPCM. The first is "lossless", but the latter can be lossy.
According to the article, PCM is also uncompressed (which is why it it lossless -_-). This would seem to indicate that ADPCM is compressed.
You could also store data lossy without using compression, e. g. if you have a format with floating point values instead of integers :o) — 91.4.31.52 20:13, 5 November 2006 (UTC)
Why would one want to "store data lossy without using compression" ? The loss that the article is about is loss whose consequence in the data (image, sound or whatever) is never desireable but may be considered tolerable as the price for the benefit of compacting the data. Floating point values just have smaller (finite and non-constant) quantising steps than integer values. (cuddlyable3) 84.210.139.189 19:53, 18 November 2006 (UTC)
32bit Floating point can easily cover all 16bit integer numbers without any quantization effects. Quite the contrary: 32 bit floating is able to cover integers up to 25 bit (signed). The definition of lossless can be stated as follows: If a certain number format is capable to represent the input format without quantization or similar the encoding process is lossless (i.e. if the input signal can be completely reconstructed from the encoded signal). This does not depend only on the number format but also on the input signal. This implies that for each input signal which has to be encoded an appropriate encoding must be found which can fully (non-destructively) cover the input number range. E.g. if the input signal has a dynamic of 8 bits, the signal can be encoded lossless using 8 bits PCM. Compression has nothing to do with it. Compression is per definition a lossless process. If something can be lossy than it is encoding. Take mp3 for instance: The encoding is based on psycho-acoustical models what can result in a data reduction which can be (to trade space for exactness) lossy. Once encoded, the output stream gets compressed by a runlength compression. This runlength compression is completely reversible and that's the reason why it's a compression. Please respect the simple statutes for editors of encyclopaedias: First get the knowledge, then write. Wikipedia does not profit when it spreads misinformation only because of some writing exercises of some self-exposers.Cls.nebadje 11:07, 8 February 2007 (UTC)
Actually, PCM is not really lossless. PCM is a way to store sound, which by definition is an analogue signal, in a digital way. Analogue signals have an infinite amount of possible sample values, whereas PCM represents samples with a finite number of values. Any "real" sample value that is in between PCM's possible values is rounded to the nearest representable sample value. So, PCM is lossless only in the sense that no data is lost apart from the loss caused by digitalization.Cassandra B 17:49, 22 April 2007 (UTC)

[edit] Amount of channels

How many audio channels can a wav file have? --Zilog Jones 00:01, 10 February 2007 (UTC)

Technically 65535 channels, because WAVEFORMATEX::nChannels is 16 bits. Motonari 17:52, 29 September 2007 (UTC)

[edit] The Codec Comparison Table

The values in "1 min" column are wrong. For example "11,025 Hz 16 bit PCM" is said to be "1291k" per minute. Thats just the Prefix it doesn't have any unit of measurement. I guessing it should be Byte? And shouldn't the value be ((Bitrate * 60)/8)? —Preceding unsigned comment added by 88.112.71.53 (talk) 02:19, 25 September 2007 (UTC)

[edit] Limit

I'm don't think the limit calculation is accurate. If the maximum file size is 2^32 bytes, then by my calculation the file can be up to 13.5 hours.[1]. The issue may be that some programs treat the file size as an signed integer, which would limit it to 2^31 bytes. However, I think the format itself uses unsigned. Superm401 - Talk 10:46, 9 November 2007 (UTC)

The limit calculation is correct, because it assumes stereo.Motonari 02:55, 11 November 2007 (UTC)
Implementations often take -1 (0xFFFFFFFF) in the size field as "length unknown" and therefore will play until the file ends. If you only have one chunk, that does the right thing. Otherwise, one uses superchunks to overcome this limitation — movies packaged into a RIFF/AVI are commonly superchunked at 1 GB boundaries. Not sure how exactly the binary layout is, but I suppose it is just a concatenation of multiple RIFFs. ("RIFF" <length> <payload> "RIFF" <length> <payload>...) -j.engelh (talk) 23:32, 5 March 2008 (UTC)

[edit] Why is there no example file?

I want an example wav file on this site. Yes sir. —Preceding unsigned comment added by 83.254.4.90 (talk) 21:48, 12 March 2008 (UTC)