Zipping/Compressing files - letter to editor in Issue #75

[digitaldrummer]

I was surprised to see this "letter to the editor" claiming that WAV files should never be zipped because zip methods use "lossy" file compression.

I was equally surprised to see that it was not completely refuted!!! So let's clear it up, OK?

Zipping files is perfectly safe and LOSSLESS - it will not lose bits or change bits. If it does, your computer is broken! if you don't believe me test it yourself.

compress a file with zip or stuffit. Now unzip it to another folder. Compare the original file with the one you zipped and unzipped. To do the compare you need to do a binary file compare or you can do an md5 checksum.

For Windows, go to a command prompt and do:

fc /b <orig_file> <new_file>

if it says "no differences found" then they are exactly the same.

I'm not a Mac guy so sorry if this is incorrect, but i believe you can do this from a command prompt:

md5sum <orig_file>
md5sum <new_file>

compare the long string that is calculated - they will be the same.

If you run Linux, then I don't need to tell you what to do...

The bottom line is that zip and file compression is SAFE. Tons of people use it every day (sometimes without even knowing it) to compress not only audio files, but important data files as well.

WMA, MP3, AAC, etc. are all lossy compression and should not be used if you want to revert to the original format without "loss". these are great formats for emailing samples to someone or storing on you iPod for casual listening.

Hope this helps.

Mike

[Waltz Mastering]

The bottom line is that zip and file compression is SAFE.

Absolutely true.

[0-it-hz]

Second that, Zip away kids.

[digitaldrummer]

and Larry, John, if you are reading this and ever need to verify something "computery" (did I just make up a word?) don't be afraid to ask your very own Tape-Op community -- I guarantee there is no shortage of computer geekness around here.

[dcsimon]

If you run Linux, then I don't need to tell you what to do...

[b3groover]

I got into a stupid argument on another board about this several years ago. All it takes is a smidgen of logic to debunk the idea that zipping (or rar'ing) an audio file will change it. Would anyone use a lossy compression scheme on a data file, such as an executable? No, because the file would not run once you uncompressed it.

[TapeOpLarry]

I thought zipping wouldn't change anything, but I guess Adam was trying to be wary of stating anything absolute! Amazing how people get the two types of compression confused...

[Andy Peters]

I thought zipping wouldn't change anything, but I guess Adam was trying to be wary of stating anything absolute!

I think if he wasn't sure, perhaps he could have asked an expert?

LZW compression has been around for four decades, and every piece of software you can think of is delivered in a compressed format, for obvious reasons. And since you're able to actually use your computer, clearly lossless data compression works.

Amazing how people get the two types of compression confused...

That's true, and I think that instead of making things murkier, perhaps an article that discusses the three types of compression used by audio guys would go a long way towards helping.

I say three types, because we have audio dynamic-range compression (as performed by your dbx 160), lossy data compression (as in MP3's lossy versions, JPEG, and others) and finally lossless data compression (LZW as used in zip, gz, etc).

-a

[Submersible]

Ha! I logged on for the first time in a while just to ask about this. I think it was about 15 years ago when I first discovered that compressing digital audio to a ZIP file doesn't result in any loss. It doesn't really reduce the file size, either.

[Andy Peters]

It doesn't really reduce the file size, either.

That's true; an audio file is basically random data, so the simple run-length coding has nothing to reduce since it cannot find redundancies. (In extreme cases the compressed file may be bigger than the original!)

Zipping is still a good idea, because as I mention, the dearchive program will tell you if the archive is corrupt.

-a

[birk]

...lemme just add that although you're probably only get a 10-20% reduction in size (with audio files) you can count on:
1) simplicity of transfer... one file, not a folder of, say 100 files
2) faster transfer... i get the "available bandwidth" bit but are you really going to rely on that all the time (at both ends) AND what if you're in the business of doing a heckuva lot of transfers every day?

just got my issue today... sorry for the rehashing.

[Brian Brock]

a final two penny nail in the coffin of this argument -

I would imagine that the infamous letter-writer has installed software on a computer. Almost every piece of software uses an installer with ZIP or similar compression. The need for bit-identical decompression is even greater with software than with data. The letter-writer therefore depends on ZIP archives all the time whether he wants to or not.

[digitaldrummer]

Zipping is still a good idea, because as I mention, the dearchive program will tell you if the archive is corrupt.

-a

Absolutely! I just sent some files to a client yesterday through an online studio service. Three out of the four files were corrupt when he tried to unzip them so I had to re-upload them to my own site for delivery (I've had trouble with that service before). I always zip the files, not only for convenience of delivering multiple files, but also because, as mentioned, you will know immediately if your files are corrupt - if they unzip with no errors then they are good.

A couple years ago I also uploaded WAV files for a CD duplication - after I got 100 CDs back, we noticed that the last song had several minutes of complete silence. I called the company and we checked the files - they had the same file size, but sure enough, they only got half the audio.

Why? Because ftp and http are not reliable transports - in this case, the ftp utility allocated the entire file size before it had the file - at some point there was a timeout or network error but it looked like the file was complete because the size was correct. Luckily the company was willing to work with me and sent me 100 new CDs. I think they learned from the experience too - now they accept zipped files on upload. Some other companies use an upload manager that does some kind of checking (crc, md5, etc.).


Not only is zipping good, it's good for you...

[;ivlunsdystf]

http://holyfuckingshit40000.blogspot.co ... s-mp3.html

Funny thread about an insane false premise taken to its extreme.

My favorite quote: "Just run a scan disk and you will recover all the lost and neglected 1's and 0's."

[jason brian merrill]

posted a dupe thread about this, sorry.

glad to see alot of other people caught that too!