Hi, about a year ago i encoded all my cd's with lame V3.95 with the settings --preset standard -q 2 -V 2 now im getting around to encoding more cd's. This time i decided to compair the wav to the mp3. Both me and my brother agree on this, the mp3 sounds slightly more bassy and the top end symbols are duller. Although i am more than happy with the quality that lame produces i decided to see if i can make it more like the original. Ive tested this on V3.95, V3.96.1 and V3.97b, settings i have tried, preset standard, extreme and insane and also tried -V 0 -k -q 0 -m s. Non of these settings seem to make a difference. I tried ogg libvorbis 1.1.1 with oggdropXPd V1.8.6 quality setting 7.54, this setting both for me and my brother could barely tell the difference between it and the wav.

Basically am i missing something, i have to use mp3 as my car stereo only supports mp3.

Thanks

This is a strong indication that either you fell victim to the placebo effect or that your player applies different equalizer to mp3s and wavs.

Before people start flaming you for your silly combinations of lame switches and for claiming you hear differences without providng any evidence, I urge you to read the terms of service rule No.8. Then read the FAQ and sticky threads about recommended lame mp3 settings. Then read FAQ and search forums about "placebo" and "ABX test"



This is a strong indication that there is probably all these settings produce mp3s which are transparent for you, but you need check the equalizer and try ABX to be sure.

i use winamp, now that strange thing is i check the equiliser and its the same, but as you suggested it must do something to mp3 files as i ripped the song to mp3, then used besweet to convert back to wav. I cannot turn the stereo up loud enough to here it at the moment as people sleeping will kill me . But looking at the spectrum analyser in winamp the converted wav file looks like the original wav in that all the song starts on the left of the scale where-as the mp3 starts in the centre and is concentrated there.

Winamp has a tradition of applying different EQs depending on filetype.

Dont use winamp to compare files - its unreliable.

Maybe a notice about winamp should be added to the FAQ..... these sort of issues tend to come up again and again.

If you encoded a WAV from the MP3 it should sound identical. It is a non-compressed version of the MP3 file. Am I wrong about this?? I has happened before, once or twice.

i found the problem, once i knew about the equiliser + winamp + mp3 problem i knew what to search for. I found it in this forum somewhere. Basically winamp uses a faster method to decode. To stop this, goto winamp, options, preferences, input find the Nullsoft Mpeg decoder, on the decoder tab untick Fast Layer 3EQ. The mp3 now sounds like the Wav in Winamp i can't remeber who said this soloution but thanks it makes a nice difference.

If you encoded a WAV from the MP3 it should sound identical. It is a non-compressed version of the MP3 file. Am I wrong about this?? I has happened before, once or twice.

[/quote]


not really as the wav to mp3 is lossy, so going back to the wav you can't gain what you have lost even if it is un-compressed.

Ive just noticed i was warned for this post, which i think is completly unfair as what i heard is a known problem, it was definatly not a placebo and the problem was winamp and the problem was found and rectified. The difference before the Fast Eq was off was massive and did not need a ABX test as a deff person could have been able to tell the difference between loud sharp symbols and near enough no symbols. etc

Send a PM to one of the mods or admins
(and say that you promise to always make ABX tests and never, never ever use the evil "-k -m s" switches)

(Edit: I have no idea why exactly you got a warning, only moderator can tell you, but I can guess...)

Seriously:
If you really could hear the difference at a qual setting this high, then Ogg-developers would be glad to have some feedback from you. But you must prove it with ABX results, because otherwise it seems more likely that you perhaps couldn't really hear the difference, and saying that you "could barely hear it at q=7.54" is one of those useless comments which are not welcome, unless proven.
If you are going to test Ogg, try a recent "Autovu"-encoder (or how it is spelled now)


you mean "cymbals" ?