Oh man... then kill me... I listen only to mp3, cause most of my favorite music is recorded so bad.... there´s really no difference...
Only... Mrs. Amos CD´s... I like their "sharp" sound... no way to get this quality with mp3´s... tried it so often, really - the wonderful TDA1541 shows you the real difference, ha! But it depends so much on the settings you use....
Only... Mrs. Amos CD´s... I like their "sharp" sound... no way to get this quality with mp3´s... tried it so often, really - the wonderful TDA1541 shows you the real difference, ha! But it depends so much on the settings you use....
Copies and jitter
-Simon
But isn't the deficiency based purely on the timing of those bits going onto the CD? ie. the problem is jitter. I've done zero A/B listening, really because I have no use for copied CDs, but the ones I have that are copies all sound unpleasant to my ears. Another poor aspect of SQ is the bass, which seems indistinct, and not well extended. These qualities (or lack thereof) seem to correlate inversely with the sort of improvements I gained in re-clocking my CD player, then (particularly) improving the clock PSU. You think I'm talking BS? I hope I'm not
-Simon
Well, the CD-R really isn't a CD anyway, it doesn't adhere to the CD audio spec, I guess, because it doesn't *really* have pits and lands, so I am sure that causes some issue. Without measuring jitter objectively I am hesitant to say it is jitter. Maybe someone with better equipment can measure and confirm. I also find this effect on very poor copies (like those made in India under license) released by the labels themselves to keep costs down.
I don't know why, but I at least find that burnt copies always gain a slight 'grainy' edge and sound a little tense. I do NOT have a high end system (NAD c521i/HK PM665 and Jamo budget towers), but I can hear it sure enough.
When I said 'modern' music I meant the bright mass-market stuff with lots of compression (anybody heard the Rush album? sounds bad even on MP3!) and limiting. The so called 'chart-busters', but that's not all they bust. But yes, some of it is niiice, my personal favourites from recent times has to be 'Come away with me'. A song that MP3 cannot handle properly without falling all over itself.
Say, that's a thought, what if someone told the recording industry to actually start making good recordings so that people would buy more CDs instead of listening to MP3s because MP3s lose so much of the music, specially if the recordings are good.
I don't know why, but I at least find that burnt copies always gain a slight 'grainy' edge and sound a little tense. I do NOT have a high end system (NAD c521i/HK PM665 and Jamo budget towers), but I can hear it sure enough.
When I said 'modern' music I meant the bright mass-market stuff with lots of compression (anybody heard the Rush album? sounds bad even on MP3!) and limiting. The so called 'chart-busters', but that's not all they bust. But yes, some of it is niiice, my personal favourites from recent times has to be 'Come away with me'. A song that MP3 cannot handle properly without falling all over itself.
Say, that's a thought, what if someone told the recording industry to actually start making good recordings so that people would buy more CDs instead of listening to MP3s because MP3s lose so much of the music, specially if the recordings are good.
Well, some of you are talking about these sharp attacks and newer
compressed recordings sounding horrible.
That´s the nature of mp3 when not encoded properly.
When a recording reaches the 0dB mark mp3 often interprets this
loud sound even louder and stores higher values for this part.
mp3 can store up to 144dB dynamics!
When you decode it for listening to a wav file these parts are cut of
to fit into the 44.1kHz wav format.
This can cause very bad clipping you immediately can hear.
You can prevent this with applying Replaygain. This alters a volume value
in the mp3 file and is regarding the mp3 lossless.
All it does is setting the volume lower so when decoded the maximum peaks
will not clip.
So for a good encoding it goes:
Good ripping software -> good encoder settings -> applying replaygain
compressed recordings sounding horrible.
That´s the nature of mp3 when not encoded properly.
When a recording reaches the 0dB mark mp3 often interprets this
loud sound even louder and stores higher values for this part.
mp3 can store up to 144dB dynamics!
When you decode it for listening to a wav file these parts are cut of
to fit into the 44.1kHz wav format.
This can cause very bad clipping you immediately can hear.
You can prevent this with applying Replaygain. This alters a volume value
in the mp3 file and is regarding the mp3 lossless.
All it does is setting the volume lower so when decoded the maximum peaks
will not clip.
So for a good encoding it goes:
Good ripping software -> good encoder settings -> applying replaygain
Hi Wombat,
I must say I've encounterd just the opposite:
CD digitaly copied to harddrive and analyzed DOES have clipped moments (goldwave analyzing program). There is a thread about this here somewhere.
MP3, also goldwave analyzing program using the LAME.dll never showed any clipping! Everything is correctly scaled to 16 bit full scale.
Regards,
Thijs
I must say I've encounterd just the opposite:
CD digitaly copied to harddrive and analyzed DOES have clipped moments (goldwave analyzing program). There is a thread about this here somewhere.
MP3, also goldwave analyzing program using the LAME.dll never showed any clipping! Everything is correctly scaled to 16 bit full scale.
Regards,
Thijs
I don´t know about this Goldwave bug. A CD can´t store higher values than
the 44.1kHz spec defines !?
You mean clipping at the mastering point? Sadly more and more recordings
introduce this due to bad mastering to make the music feel louder
A good explanation:
http://www.prorec.com/prorec/articles.nsf/files/8A133F52D0FD71AB86256C2E005DAF1C
Here is what i mean with en/decoding:
the 44.1kHz spec defines !?
You mean clipping at the mastering point? Sadly more and more recordings
introduce this due to bad mastering to make the music feel louder
A good explanation:
http://www.prorec.com/prorec/articles.nsf/files/8A133F52D0FD71AB86256C2E005DAF1C
Here is what i mean with en/decoding:
Attachments
You can find a tool doing replaygain here:
http://www.geocities.com/mp3gain/
When i encode an entire album i analyze all files and apply a constant gain to all songs to the point none of it clips anymore.
http://www.geocities.com/mp3gain/
When i encode an entire album i analyze all files and apply a constant gain to all songs to the point none of it clips anymore.
Hi,
Very interesting findings Wombat! I must say I'm surprised how far you and some others go to make MP3 sound more acceptable.
sangram,
I too love Come Away With Me, by Norah Jones. On the digital clipping thread someone told me this mostly pleasant sounding album is taken into clipping on multiple occasions! I was not too surpised though, the sound can get slightly edgy now and then, which I'd previously put down to my speakers midrange performance...
Makes me wonder how much nasty stuff I hear is really just on the recording, but smudged into obscurity by low-res systems. e.g. a friend has B&W DM603 speakers driven by a cheap amp and CDP, with out-of-the box interconnect! Everything sounds smooth and (bass)rich, with way too much sloppy, wobbly mid-bass. Never ever harsh though! I wish mine could have that quality, with the other good qualities it already has
-Simon
Very interesting findings Wombat! I must say I'm surprised how far you and some others go to make MP3 sound more acceptable.
sangram,
I too love Come Away With Me, by Norah Jones. On the digital clipping thread someone told me this mostly pleasant sounding album is taken into clipping on multiple occasions! I was not too surpised though, the sound can get slightly edgy now and then, which I'd previously put down to my speakers midrange performance...
Makes me wonder how much nasty stuff I hear is really just on the recording, but smudged into obscurity by low-res systems. e.g. a friend has B&W DM603 speakers driven by a cheap amp and CDP, with out-of-the box interconnect! Everything sounds smooth and (bass)rich, with way too much sloppy, wobbly mid-bass. Never ever harsh though! I wish mine could have that quality, with the other good qualities it already has
-Simon
Hi Wombat,
Nope.. goldwave isn't buggy... it real in the data.. instantly wave-cut-off because Full-Scale is reached.. it is recorded like that!!!
Every, (yes every !) MP3 I open, I open not with a media-player of any kind, but with a sound analyzing program (Goldwave (standard) or Audacity or Matlab) , and I have never, (yes never !) seen a clipped MP3.
What software are you using for visualisation/encoding/decoding etc, Wombat ?
Kind regards,
Thijs
Nope.. goldwave isn't buggy... it real in the data.. instantly wave-cut-off because Full-Scale is reached.. it is recorded like that!!!
Every, (yes every !) MP3 I open, I open not with a media-player of any kind, but with a sound analyzing program (Goldwave (standard) or Audacity or Matlab) , and I have never, (yes never !) seen a clipped MP3.
What software are you using for visualisation/encoding/decoding etc, Wombat ?
Kind regards,
Thijs
@tschrama
No clipping in my experience is very uncommon. At least with newer recordings.
Most likely it is encoded with an encoder taking away lots
of highs. So the overall energy goes down!?
You would see this with a spectral analysis like Cooledit
provides.
@Craddle22
Now there we go further. The decoder
There is a page comparing many different decoder for accuracy. Many do have faults.
I don´t have the link atm
The Fraunhofer decoders work nearly perfect and it is integrated in Nero for example.
The decoding engine build into lame also is accurate.
The mad decoder is different and offers best performance in theory. It has 24bit precission and can dither if remember right.
There are for sure informative threads at hydrogenaudio.org
No clipping in my experience is very uncommon. At least with newer recordings.
Most likely it is encoded with an encoder taking away lots
of highs. So the overall energy goes down!?
You would see this with a spectral analysis like Cooledit
provides.
@Craddle22
Now there we go further. The decoder
There is a page comparing many different decoder for accuracy. Many do have faults.
I don´t have the link atm
The Fraunhofer decoders work nearly perfect and it is integrated in Nero for example.
The decoding engine build into lame also is accurate.
The mad decoder is different and offers best performance in theory. It has 24bit precission and can dither if remember right.
There are for sure informative threads at hydrogenaudio.org
Hi Wombat,
I s does not depend on the recording offcoarse, but on de encoder/decoder. I have many mp3 recordings downloaded from the internet (properbly with encoded diffrently), but no clippping! So it must be an error in a your decoder? What decoder are you using... ?
can you give us an example of
WAV->encoded->decoded->clipping?
I really really have no examples in my music library...
Regards,
Thijs
I s does not depend on the recording offcoarse, but on de encoder/decoder. I have many mp3 recordings downloaded from the internet (properbly with encoded diffrently), but no clippping! So it must be an error in a your decoder? What decoder are you using... ?
can you give us an example of
WAV->encoded->decoded->clipping?
I really really have no examples in my music library...
Regards,
Thijs
Where do you know from it doesn´t clip?
When you opened it in an editor it gets decoded in the way i described above. The higher peaks are not there anymore when you look at your waveform.
Did you try Mp3gain i linked to and tested the original mp3s for clipping?
A nice one was Rammsteins Mutter, it was released at the time we encountered these problems.
One Thread here with some:
http://www.hydrogenaudio.org/index.php?showtopic=7140&hl=
Wombat
When you opened it in an editor it gets decoded in the way i described above. The higher peaks are not there anymore when you look at your waveform.
Did you try Mp3gain i linked to and tested the original mp3s for clipping?
A nice one was Rammsteins Mutter, it was released at the time we encountered these problems.
One Thread here with some:
http://www.hydrogenaudio.org/index.php?showtopic=7140&hl=
Wombat
Hi,
You can see it when you look at the transientwave form. I do this on default.
The original recording (on CD) you mean. I trust it clips here and there. But if you encode it (what encoder did you use), and then decode it (what decoder did you use), did it clip? Did it clip more? Did it not clip at all? Did it sometimes hit Full Scale but without cutting off the waveform?
To make my point: I can send you several MP3 without clipping (at least on using decoder), but I really can't send you any example of clipping after encoding/decoding. If I encounter such an example I will be convinced.
Regards,
Thijs
You can see it when you look at the transientwave form. I do this on default.
I do this allways (default program to open MP3 is goldwave) and I see no waveforms being cut-off. So no clipping.
I looked at it but decided that not every musical piece is intended to be equal loud. I trust the program does achieve it's goal fine, but I don't agree with the goal.
The original recording (on CD) you mean. I trust it clips here and there. But if you encode it (what encoder did you use), and then decode it (what decoder did you use), did it clip? Did it clip more? Did it not clip at all? Did it sometimes hit Full Scale but without cutting off the waveform?
To make my point: I can send you several MP3 without clipping (at least on using decoder), but I really can't send you any example of clipping after encoding/decoding. If I encounter such an example I will be convinced.
Regards,
Thijs
btw. There were often people bringing samples to the mp3 community that obviously sounded wrong. Afterwards it was only a problem due to clipping.
This is for all lossy compressors, even musepack (mpc, best lossy compressor) isn´t immun of this. These people often have "lousy" equipment against people on this board here but were able to hear it.
When your music files don´t clip, lucky you.
All others may try the mp3gainanalysis on samples they hear distorted sound.
Me is tired of all this, i talked about such issues for years now and had some experiences with so called audiophiles that weren´t able to pick everything right when comparing compressed music to their originals.
This is for all lossy compressors, even musepack (mpc, best lossy compressor) isn´t immun of this. These people often have "lousy" equipment against people on this board here but were able to hear it.
When your music files don´t clip, lucky you.
All others may try the mp3gainanalysis on samples they hear distorted sound.
Me is tired of all this, i talked about such issues for years now and had some experiences with so called audiophiles that weren´t able to pick everything right when comparing compressed music to their originals.
Dear Wombat,
Now comes the fun part....
I read the link you provided and , allthough I find the explanation a bit quirky, understand quiet well what's going on... and got me really interested....
So I did my onw test.....
1] made a wav file containing quare wave (2Hz) ('metronome') using the Full Dynamic Range (mono, 5 sec, fs=44100Hz).
2] using LAME to encode to MP3 (setting = 'hifi' / 160kbps)
3] open the mp3-file with cold wave....
And.............
Behold ... (due to restricted bandwidth ?) the quare wave rings, and the ringing is clipped.... even if it is still a digital respresentation (NOT analog) .. it is clipped... I am confinced: it can happen.
I think this is not all ... there more to consider.. but I'll leave that for a nother time...
Cya,
Thijs
Now comes the fun part....
I read the link you provided and , allthough I find the explanation a bit quirky, understand quiet well what's going on... and got me really interested....
So I did my onw test.....
1] made a wav file containing quare wave (2Hz) ('metronome') using the Full Dynamic Range (mono, 5 sec, fs=44100Hz).
2] using LAME to encode to MP3 (setting = 'hifi' / 160kbps)
3] open the mp3-file with cold wave....
And.............
Behold ... (due to restricted bandwidth ?) the quare wave rings, and the ringing is clipped.... even if it is still a digital respresentation (NOT analog) .. it is clipped... I am confinced: it can happen.
I think this is not all ... there more to consider.. but I'll leave that for a nother time...
Cya,
Thijs
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.