[jagorev]

Quote:

Originally Posted by richard For those who say they hear a major difference between different bit-rates or encoding technologies, are you running double blind tests (ABX or whatever) or do you know which bit-rate you're hearing?

Yup, I use the Foobar2k ABX feature. Difference is quite audible between FLAC and LAME-alt-preset-standard. Not so much of a difference between alt-preset-standard and alt-preset-extreme, though, so I stick to standard (~192 kbps VBR) for my portable needs, keeping FLAC at home.

 

[loopsider]

Quote:

Originally Posted by richard For those who say they hear a major difference between different bit-rates or encoding technologies, are you running double blind tests (ABX or whatever) or do you know which bit-rate you're hearing?

I can definetely tell you if something is below 192 kbps, and if orchestral I can tell you if it is 192 kbps. If its less than or equal to 64 kbps you will know because I will vomit.

 

[AdamWill]

About 80% of my collection is in q4 (usually works out between 100k and 128k) ogg and can't be re-ripped as the CDs are all stuck in the U.K. (shipping 600 CDs across the Atlantic is excitingly expensive). I listen to them through my HF-1s and ER4s and they're fine. Easily better than bad headphones. If I listen for the compression artifacts I can kinda hear them but, put it this way, I could choose not to listen to all this and just listen to the stuff I bought in Canada and have in FLAC, but...I don't. Draw your own conclusions.

 

[Krause]

I am in the process of reripping my cds. I used WMP the first time and this time I am using EAC with lame. I will tell you the difference is staggering.

http://wiki.hydrogenaudio.org/index.php?title=LAME

Is where I learned everything on how to rip nearly perfect copies in about an hour of messing around. Those forums are priceless for a newb like myself looking to back up his cd collection or get ready for a new portable

 

[EvilDwarf]

With MP3s at 128kbit I can very clearly hear compression artifacts and they sound somewhat "thin". If I listen to web radio for some time, it doesn't keep me from enjoying music. With Lame encoded mp3s at 192 kbps VBR I don't hear any significant differences to CD. Maybe you need golden ears or extremely high quality equipment to find differences there (or turn up to insane levels to pick up the details). Maybe with classic music, it would matter also. Otherwise lossless encoding is not compulsory - just like golden cable contacts (mostly imagination makes them sound better).

 

[Tachikoma]

While it isn't hard at all to tell the difference between 128k and 320k MP3s, 128k MP3s aren't THAT bad if you don't have anything better. Don't let that stop you from getting better equipment, you can always get new music with higher bitrates in the future

 

[halcyon]

[offtopic]
I do apologize for this off-topic content in advance, but this opportunity is way too good to let pass.

All you who think it's easy to hear artifacts with lossy encoders at 128, 192 or even 320 kbps, now is you chance to prove it and prove others wrong.

Go to this website: http://www.maresweb.de/listening-tests/mf_128_1.php

1. Download: http://www.maresweb.de/listening-tests/ABC-HR_bin.zip

2. Read the readme.txt inside the zip file

3. Download the test samples and unpack them as instructed. Then run the .bat files (or manually decode them) as instructed in the readme.txt

4. Start abchr.jar file. Load a file set config file (again, as instructed)

5. Start testing (help: http://ff123.net/64test/practice.html )

6. Save results from the program and upload them to the test organiser

If I miswrote something in haste, the proper instructions are in the readme.txt in the aforementioned zip file.

In my humble opinion, I doubt the majority of you will be able to detect current lame 128kbps (above test) from the original reliably for most of the samples. Not to mention pulling it off for Ogg or Apple's AAC implementation. Reliable detection here could be defined as a result in which you get at least <0.05 probability of guessing in the ABX module of the ABCHR.jar program.

Now here's your change to prove to yourself that you are right and I'm mistaken (which I'm more than willing to admit, if somebody shows proof).

Please do not take this as me claiming you absolutely cannot hear the differences. There are many people who can distinguish some tracks at 128kbps from original reliably. However, they are usually very experienced in blind subjective evaluation of lossy encoders.

As such, I find it somewhat unlikely based on my own testing on myself and others, that most of you will be able to do it reliably under blind conditions, unless of course you happen to be very experienced in subjective blind evaluation of perceptual encoders output.

And don't take it as an offence. Take it as a sign of how far the lossy encoders have developed since the crappy encoders of the bygone days.

Your listening test results (if done as instructed) would also be more than welcome to the test organisers.
[/quote]

PingPongPay,

the only way for you to know is to find it out for yourself. Get the headphones, if you can hear the problems, then think how to solve it (better files, different headphones, etc).

Do remember that any 128kbps encoding is not the same as any other 128kbps encoding. Encoder maker, encoder flags user and possible transcoding all affect the quality of the end result.

A good 128kbps encoding can be transparent to most of people most of time. However a crappy 128kbps encoding, once transcoded from AAC to mp3 can sound like utter crap to almost anyone.

It's all down to how it was done and who does the listening.

 

[Skylab]

No disrespect to Halcyon, but there are plenty of people who think that this sort of testing is not a valid way to listen for or "prove" audio differences. I recommend people who are interested to look at recent Stereophile articles on the topic of ABX testing and music.

I would encourage people to LISTEN, to music they are familiar with, and make up their own minds in this way, not based on a test designed by someone else.

Many people have an axe to grind about this topic, on both sides. If someone believes they can detect differences, saying they cannot is not productive. Music is about ENJOYMENT, and anything that increases someone's enjoyment of it is OK in my book. Deciding on things like bit rate should be a personal decision based on one's own situation of what they can hear, priorities related to how much hard drive space they have/can afford, how they listen, etc.

 

[CookieFactory]

Stereophile dismissing ABX? Who would've thought!

 

[Skylab]

Quote:

Originally Posted by CookieFactory Stereophile dismissing ABX? Who would've thought!

Yeah, I know, but the article was written by someone who does not work at Stereophile, and it's pretty scientific.

 

[EvilDwarf]

Quote:

Originally Posted by halcyon [offtopic] I do apologize for this off-topic content in advance, but this opportunity is way too good to let pass. All you who think it's easy to hear artifacts with lossy encoders at 128, 192 or even 320 kbps, now is you chance to prove it and prove others wrong.

That was exactly my point. Many beliefs about the "necessity" of losless enconding are really related to the imagination of the listener. This might apply to other aspects of the audiophile hobby as well.

 

[chrisfromalbany]

Quote:

Originally Posted by halcyon [offtopic] I do apologize for this off-topic content in advance, but this opportunity is way too good to let pass. All you who think it's easy to hear artifacts with lossy encoders at 128, 192 or even 320 kbps, now is you chance to prove it and prove others wrong. Go to this website: http://www.maresweb.de/listening-tests/mf_128_1.php 1. Download: http://www.maresweb.de/listening-tests/ABC-HR_bin.zip 2. Read the readme.txt inside the zip file 3. Download the test samples and unpack them as instructed. Then run the .bat files (or manually decode them) as instructed in the readme.txt 4. Start abchr.jar file. Load a file set config file (again, as instructed) 5. Start testing (help: http://ff123.net/64test/practice.html ) 6. Save results from the program and upload them to the test organiser If I miswrote something in haste, the proper instructions are in the readme.txt in the aforementioned zip file. In my humble opinion, I doubt the majority of you will be able to detect current lame 128kbps (above test) from the original reliably for most of the samples. Not to mention pulling it off for Ogg or Apple's AAC implementation. Reliable detection here could be defined as a result in which you get at least <0.05 probability of guessing in the ABX module of the ABCHR.jar program. Now here's your change to prove to yourself that you are right and I'm mistaken (which I'm more than willing to admit, if somebody shows proof). Please do not take this as me claiming you absolutely cannot hear the differences. There are many people who can distinguish some tracks at 128kbps from original reliably. However, they are usually very experienced in blind subjective evaluation of lossy encoders. As such, I find it somewhat unlikely based on my own testing on myself and others, that most of you will be able to do it reliably under blind conditions, unless of course you happen to be very experienced in subjective blind evaluation of perceptual encoders output. And don't take it as an offence. Take it as a sign of how far the lossy encoders have developed since the crappy encoders of the bygone days. Your listening test results (if done as instructed) would also be more than welcome to the test organisers.

this is a pretty good test and I agree with halcyon.

 

[Ayreonaut]

Some music sounds OK at 128. But not all 128s are equal.

The better 128s (that you rip yourself with the programs suggested here) suffer mostly from the lack of high frequency extension. But downloaded tracks often were ripped with poor software and some pretty bad artifacts are present. There's no real remedy for that. The only thing I've tried that helps a little is to ruthlessly hack off the upper octaves with an equalizer. It sounds dull, but the artifacts are gone.

 

[taymat]

Quote:

Originally Posted by Skylab Yeah, I know, but the article was written by someone who does not work at Stereophile, and it's pretty scientific.

That doesn't mean he wasn't paid by Sterophile, or that it's fact because it includes science.