Ok, I downloaded a couple of songs from The Bends so I went and bought the CD. I then went to rip the album in EAC using the 192 variable LAME sound quality option. After it ripped it I played it and compared it to the downloaded version. The downloaded version was so much more louder. The peak quality I was getting was 75.8%. The downloaded version sounded much more sharper and was obviously more louder. I would like to know why the downloaded version sounded better. I thought EAC was the best. It is (downloaded song) encoded at 192 that is all I know. It's just frustrating that I'm using an option where it should] sound better but it's not.
Help!
Thanks

Dowloaded version may have been normalised louder before encoding.

Ok, thanks. How can I go about achieving this? I tried encoding it at the same settings in dBpoweramp and it sounded like the EAC also.

It's possible that the person who made the file processed the file prior to encoding using a dynamics compressor.

It's also possible that it came from a source other than The Bends CD (is there a DVD?).

Nope, there aint no DVD.

Louder is not better.

Rip and encode your CD without modifying the audio data. If you want it to sound louder, then turn up your speakers.

LAME at 192k should be transparent to almost anyone on this planet.

ARGH!!!! And we all wonder why the loudness race continues?!?!?

More louder is not more better!

Louder isn't always worse: I've got the Telarc recordings of 1812 Overture, Bolereo, and Carmina Burana on my iFP-895 along with a bunch of Dire Straits albums, a selection of Celtic music, some Moody Blues, and a little Art of Noise tossed in.

When I encode from Wavpack to Vorbis (aoTuV beta4, -q 2.25) I use Foobar and the vLevel DSP, otherwise I'd be spending all my time at work just adjusting the volume up and down.

bleedlikeme, people here at Hydrogen Audio generally hold that the best-quality mp3 is the one that sounds closest to the original cd. The word "transparent" is used to describe an mp3 that doesn't sound any different than the original cd from which it was encoded. What you should do is use EAC to rip a song to uncompressed wav file, and then compare that to the one that you downloaded. If the downloaded song is much louder, then someone probably increased its volume before encoding, or possibly increased the volume of the mp3 file after encoding.

Songs that are louder usually seem to sound "better" to our ears, because we can hear details more clearly. But louder doesn't mean that the mp3 is actually better-quality - as AtaqueEG says, if you turn up the volume on the mp3 that you encoded from the cd using EAC and LAME, it will sound just as good (and quite possibly better) than the mp3 that you downloaded.

Interestingly, a common trick of some encoders (past and possibly present versions of Windows Media) is to boost the volume of the encoded track slightly, thereby making it "seem better" to a quick listen, without actually increasing the quality.

A danger of mp3's being too loud is clipping - if the volume is boosted a lot, then the loud sounds will start to hit the maximum "ceiling" of allowable digital sound, which can hurt the crispness of sound quality.

One tool that will be useful to you here is mp3gain. It can be used to raise and lower the volume of mp3 files, but its main purpose is to standardize the volumes of mp3 songs, so taht you don't have different albums and songs that are much louder or softer than each other. Most albums taht are mastered at loud volumes will "clip" when encoded to mp3, and so mp3gain can be used to lower the volume.
Mp3gain also tells you what the maximum no-clip gain. I've seen some songs that I've downloaded (including the songs that I got off Radiohead's "The Bends" before I bought the album) had a maximum noclip gain of -7.5 dB - which means that the volume had been raised at least 6.0 dB after encoding, by someone who didn't know what they were doing or by someone who thought that louder is always better. perhaps the files that you downloaded are similar to these? (I know you said 75% max volume, but you might have meant something else by this.)

If you used 192 CBR (Contant Bitrate) I would encourage you to try LAME's VBR setting --alt-preset standard as detailed in the List of recommended LAME settings. If you're wanting to archive your audio CD's using the --alt-preset switches is the way to go.

It's called Exact Audio Copy, not "Magically Louder and Better Audio Copy"...

EAC is undeniably one of the top tools for error free audio extraction. It's job is to produce an exact copy of the original, and that is precisely what it does when used correctly. It's responsibility ends at audio extraction, and the sound quality of the converted music is dependant upon your choice of a 3rd party encoder and it's settings.

So your problem here is not really with EAC, but in general with mp3 encoding. Lame is definitely the best mp3 encoder, and all you need to do is find a method of encoding with Lame that will suit your preference. Though I would really rather recommend that you learn to keep your music encodings sounding exactly like the original source material, and instead alter your playback equipment and methods to produce a more pleasing output.

First thing is that making wave louder than it originally is doesn't make it sound better in any way. Only when your sound card od aplifier generates some noise and you must have loud signal to not hear it

On the other side decreasing volume make sound worse! WAVE on mixer should be on 100% if you play only one sound source and quality matter the most.

BTW, MP3 192kpps has shitty sound quality so it's for "preview and buy" only and we cannot talk about "quality" in case of using mp3 192...

Maybe 192kbps IS transparent for the music in question, you can't say that 192kbps has "shitty" quality until you hear the samples both compressed and uncompressed. Before VBR was as accepted as it is now, 192kbps was pretty much the standard.

Not always, a lot of "standard" built in soundcards will clip at 100% sometimes.

192kbps in MP3 is transparent for most people. I'm sure that you would have a hard time identifying any difference against the original in an ABX test.
Anyway, post back some ABX results proving that you can hear a difference against the original and that way you could contribute better than just throwing personal statements that most probably don't hold any ground.

TOS #8 violation

Yep!! Idon't want to sound like a TOS #8 zealot but sometimes i see things and I really can't help it!!

Their amp is already set to 10 for most people, so they need to make their MP3's sound louder.
Of course, mine goes to 11.

You're right, I forget to mention it. I don't know if new sound cards has that problem but old ones has expecially in sb16 era



Maybe my words was to "strong"
First, hardware must be able to play sound sample with enough fidety to hear any artefacts, and second one must be sensitive to them. That is the thing that you can gain with every sample in mp3 compared to original losless sample, so if you did some comparsion between MP3s and WAVs then you can say easly if sample is compressed with Mp3 or losless.
But if someone doesn't matter about quality much and/or use bult-in motherboard audio codec, speakers with low fidety or have narrow frequency response ears then MP3 192 will be perfect solution

ps. Full ABX test whatever it is I will make tomorrow when I find time that I now haven't

Well, I did a lot of testing in the past, but I still cannot tell *easily* if a sample is compressed or not at high bitrates (>= 192 kbs). Only problem/killer samples are easy.


This is still a strong statement. I care about quality, have a professional sound card, good speakers and good ears, but 192 kbs sounds fine for me, for most music. (to make it fine for almost *all* music, and to use efficient encoding I use Lame --aps). By the way, it also depends on the encoder you use, Lame is considered to be the overall best MP3 encoder.


Good luck, I'm really curious what the results are.

Way to go. I'm looking forward to see your results.

Sorry but I haven't yet made this test, and reason is that I have headache and I'm sick and in that state I'm not good candidate of testing anything To fullfil my tragedy my jack -> big-jack converter is lost and test cannot be done untill I get one

My speakers stand now very close and I don't have any stereo, only mono sound Why do I need good stereo separation? Turn on "voice cancelation" or "vocal remover" on mp3 192kbps and later on losless or mp3 320kbps to find out what I mean a lots of artefacts and in stereo headphones it can be easly heard

MP3 192kbps is ideal compromise between audio quality and file size, but we cannot say it is "transparent" in any way! To mp3 320kbps I would agree that it is transparent cos even voice cancelation and other sound postprocessing shown no audible diffrence between losless samples.

I have a question, does other mp3 encoders have much worse quality than lame? Some of my mp3s are rather lo-quality with lots of artefacts and others sounds almost unaffected by process of encoding. So doest some stupid %#@@! recompres from lower bitrate or other mp3 codecs (there is few that I know names but not quality) are much worse in sense of quality

ps. From what I readed in this forum about ABX tests I should not hurry and prepare everything so there can be later any "If I have better...".
ps2. On my speakers with my current state I got ~~ p=0.3 for now

*I so badly want to comment... but I don't want to waste my time.*

No problem, we'll wait


Postprocessing should always be done on the lossless files, if available. The whole point of MP3 is that it should sound to our ears in a certain way in the state it is in. As soon as you alter the file that concept is ruined.


Again, postprocessing has nothing to do with transparency.


Yes, there are quality differences between different encoders, even between different versions of Lame. When you perform a test, you should always include the encoder/version/settings you have used along with your results.


Correct. Especially at high bitrates your ears can get very tired after some time. For high bitrates, I normally take a break of about 3 sec between each trial.


That is far too high.

Good luck