Geez, just to demonstrate how far over my head this has gone, for the longest time, I thought you all just were spelling "lousy" wrong.


I've never heard the word "lossy" before, somehow I doubt it's in any dictionary.


Interesting thread though.

Mp3 is really all about convenience. Most people aren't that concerned about sound quality. If they were, nobody would install radios in cars anymore.

 

No, there isn't an audible difference. BUT YES, they are a lossy form of media, BUT you are missing my point. THE LOSS IS INAUDIBLE FREQUENCIES. That is the whole point. There is also a compression of mp3s, just like how zip files work, but that compression is completely reversible. That is beside the point, but does contribute anywhere from 15-40% of the compression. The rest is achieve by the filtering out of inaudible sounds. Plain and simple. I just burned another cd to illustrate my point. I ripped the cd to mp3 (I even made the bitrate 192kbit/sec), then burned it back to another cd. I've gone through this with a few people now, from my roomates back at school to a couple friends here. I played the 2 cd's, not telling which was which, and they could not tell me which was the original, and which was the burned. And this was not on crappy audio equipment.

*You* may be unhappy that your huge cd collection has nothing over a smaller collection of mp3s, and all that money you spent on your beloved cd collection seems a waste. (See the resemblance to the record lovers' case?) The *fact* is, I can't tell the difference, and haven't many ANY who have. Sure they can tell the difference between 128kbit/sec or less especially, but above 192kbit/sec, nobody I know can tell. And btw, most mp3s out there are 128kbit/sec, maybe you should take a look at bitrate you are listening to, rather than assuming they are so *inferior*.

 

http://dictionary.reference.com/search?q=lossy

lol this further proves my point, from the definition:
" The lost information is usually removed because
it is subjectively less important to the quality of the data
(usually an image or sound) or because it can be recovered
reasonably by interpolation from the remaining data."

What's less important to the data than sounds that we can't hear? Hmmmm. Again, with higher quality encoders, it will take only this data that can't be heard.

 

Actually, I'm not unhappy at all, because it would take my huge cd collection to generate a smaller collection of MP3's should I chose to go an inferior route. I don't know where you get your CD's from (unless you're file sharing which is a whole other issue), that you're burning, but mine are bought and paid for. So regardless if my CD collecion is huge or not, the source has to come from somewhere. Secondly, if you're comfortable with inferior quality music, then that's certainly your right....I however am not. I guess my ears are a little more sensative than yours. I can definitely tell the difference between a CD track and an MP3 that has been ripped at 192kbps. It gets a little more difficult @ 256kbps though.

I could compare your argument to aftermarket rims. You've got originals, and then you have knockoffs. On the surface, you can't really tell the difference, but when it comes to quality and craftsmanship, they're a world apart.

We could go on all day, so why don't we just agree to disagree?

 

Please explain this a little more in detail, because I don't think I'm interpreting this correctly. Are you saying that when you rip an MP3, it gets compressed, but when you convert back to a .wav file the lost information somehow magically reappears? I don't think that's what you're saying, but I just want to make sure.

When you rip an MP3 from a source file, it is what it is. The loss has already taken place, and there's no way to retrieve it once you've done this. A Shorten file (.shn) however, does act like a .zip file and absolutely nothing is lost when it is extracted back to a .wav file.

 

So many people have tried to convince me that the frequencies rolled off are INAUDIBLE and have called upon experiences where they have compared the MP3 to the original source and heard no difference. Go ahead...blind test me. I CAN hear the difference, and those who dont hear the difference are not used to listening to their friends "$3000 system" and finding detail in music. Yes, it is minute, but the compression is most noticeable in the frequencies that I pay most attention.

The only thing I can say as a counterpoint, is that supposedly different programs rip and encode better than others. I use musicmatch...if thats a bad program then I certainly can reevaluate after getting a better one!

BTW...so how about those green lights guys?? anyone? (cricket sounds on the background) oh, carry on!

 

Get the 7893 so I can see what it looks like in the Z before I get it!

 

lol, ok lets agree to disagree. To each his own, but I'll stick with my mp3s. For the record, the mp3s that I use for comparison are bought, I do have a decent cd collection. I just find mp3s much more convenient, w/ no quality loss.

 

but wait lithoss..before yoy leave, please tell me what mp3 program you use, maybe I can try it out...

 

While I'm not opposed to the good use of MP3s, LithossZ, man, you need to do some studying.

If all that sound comes back when you uncompress it, do this. Rip a .wav and encode an MP3 from a CD at 192. From that MP3, create a .wav. Reencode that .wav at 192 again. Repeat the encoding/decoding about 10-20 times, then create one last .wav file.

Play that wav file through good quality headphones or stereo (preferably headphones). Do an A/B comparison to your original .wav that was ripped from the CD.

Get back to us.

If your theory holds true, all you will lose is the inaudible sounds.

 

IMO, the best combo I've tried for MP3 creation is EAC to rip and LAME to encode (LAME binaries here).

If you need help pointing EAC to LAME, let me know. There are some tutorials out there, and it isn't as hard as it sounds.

I've been encoding with the --r3mix switch VBR for awhile, but I may start using some different settings to see if there are improvements in the newer version of LAME. I haven't updated yet to the latest.

 

Ok what is was saying is there is 2 levels of compression in an mp3, from what I've read. The first is the removal of inaudible sounds, whether you agree that they are inaudible or not.

The second, which was what I was referring to there, is a compression similar to that which zip or rar files undergo. What happens is when a redundant pattern or section of bits is found, all subsequent identical sections are replaced by a pointer to the original, identical section. But because in music sections like these do not occur all that often, the resulting compression isn't that great. Later, when the mp3 is decoded to .wav format to play, the decoder replaces these pointers with the original section of bits, resulting in a perfect duplicate of the original, non-compressed file, disregarding the inaudible sound filtering. This is how .shn files work as well.
When you rip a cd the .wav file is read from the cd. After this, you chose to take the path of compressing it into an mp3, or into an .shn file. The two work very similar, but mp3s use the extra audio filtering algorithm.
That's probably more than you wanted explained, but I did anyway.

 

This must be brand spanking new technology then, because once an MP3 is ripped at any bitrate and the loss has occured, there's no way to get it back.

Can you post a link or something to this, so I can educate myself about it. If this is true, then I'll definintely be getting an MP3 player for the car....lol!

 

This is true Mannish, the wav file that has been raped by the encoding over and over will not sound like the original. No matter what program you use to date, you can be sure that there will be some tiny bits of audible sounds that are lost. Still, these losses are so minute with good algorithms that it can't possibly be heard by the imperfect hearing of the human ear. Repeating the algorithm over and over only multiplies this problem over and over, making it noticeable. In theory (also in the worst case scenario), it also results in a file that would probably be equal to a mp3 encoded at below 14kbit/sec the very first time.

Now boost the bitrate up to 256kb/sec, and encode it just once, which is all that's necessary, and the resulting bit loss is extremely minute. Now I don't care how good your stereo is, you should not be able to hear these losses. Hence, inaudible.

I think I'm going to try your experiment, it'll be interesting, then I can post it on the internet somewhere for reference. We can make it our my350z.com acoustical experiment, hehe.

 

I use the same as MannishBoy, and the newest version of LAME, 3.93.

 

MP3 compression is not like zips and rars. It is probably closer to jpgs. Zips and rars basically just reduce duplicated information in a file to compress things, while jpgs and mp3s use tricks to make it seem like little data is lost.

MP3s use psychoaccoustical models to cut out stuff that people supposedly wouldn't notice. For instance, if there are two different sounds, and one is much louder than the other, the compression model drops the very soft sound. Very different than zip type compression, where after you unzip, all the old stuff is still there. On an MP3, lots of stuff gets averaged out and dropped, so there are generational differences in the files if you do like I mentioned in my prior post. Every compress/uncompress cycle of the file is different than the last to a degree.

 

I'm not saying the loss can be recoverd, lol. I'm talking about the 2nd part of the compression, the part that's very similar to the zip file compression. This part is 100% reversible. However, the inaudible filtering part of the compression (the 2nd part, which is completly different from the 1st part), the part that you claim is so detrimental to the resulting sound quality, is not reversible. Ever.

 

I can hear the difference with quality headphones on about any bitrate MP3 from the original source. I guess what I'm differening with you on is the impression I'm getting from you that there is no way anybody can hear the difference on a good quality MP3, because all the frequencies lost are irrelevant to the human ear.

I also agree that for me, the convenience outweighs that loss in many cases, just like you seem to be saying.

I think where I differ is just some of the minor details. Lossy compression will never sound as good as the original. That doesn't mean lossy compression is horrible. Just look at DVDs.

 

Exactly. Well put, I'm just saying the the psychoaccoustical models used to cut out this stuff is not as bad as assumed, and there is an additional compression on top of that that is similar to the zip/rar file compression. This is why when you rar an mp3, the resulting difference is a matter of >10000bytes on average, about .2% compression, which is most likely all attributed to the compression ID3 tag that mp3s have. rar a .wav file, and the resulting compression is about 30%. If this additional compression was not used on mp3s, when you zip an mp3 the resulting compression would also be around 30%.

I can honestly not hear a difference with headphones between the original and a 256kbs mp3, let alone a stereo. And everyone else I've talked to about it can't either. Oh well, I think we understand each other though, that's all that really matters I guess.

I agree about the convenience and we're just differing in the minor details.

 

well guys, what are your thoughts on WMA files? more compression than mp3, is it even as good as MP3?