Soon to be Newbie Stereo owner & MP3s/OGG

[voodoothunder]

I've been broke for at 10 years paying for the school thing. So now I'm soon going to purchasing my first high end system. Now I know that speakers alone can cost up 10s of 1000s of dollars, but when I say high end, its high end for me RIght now I have $200 JBL speakers hooked up to my computer

My intial idea was to get the Dynaudio Audience 52se and some sort of Denon Receiver... But my question is this...

I have a hard drive full of MP3 and OGG files (All of which btw I either own the original CD or have forked out $1000+ to see these bands live - The Stones' last tour set me back a pretty penny for 5 tickets in Toronto). If you have a higher end sound system, are you going to notice 'the compression' of the files? Do you really need the full CD in order to get a good sound?

 

[MDS]

Many people will say that they can easily distinguish a sound quality difference between compressed audio formats and a CD. However, if you use a suitably high bit-rate (192 kbps or greater), I think it is very difficult to impossible to tell the difference for *most* music. There are some kinds of music that more readily exhibit artifacts from the encoding process but for rock/pop/dance/disco/r&b/country etc it sounds just fine.

If you rip your own music from CDs you own, as I do, then you might want to save the wav file and compare it to the mp3 version made from that wav and see for yourself if you can tell any difference. I have all of the wavs as well as 192 kbps mp3s from my cd collection and for the most part, the mp3s sound just as good as playing the wav. I keep the wavs around as an archive of my music collection, but I play the mp3s because they are smaller and easier to stream.

 

[voodoothunder]

Yeah, thats good. Streaming from my home network is actually what I would like to do.

I took a Graphics and Audio course and we were told that the stuff that is removed from a track is stuff the human ear can't hear. But it one of those things that you never really 'trust'.

Is there anything I need between my computer and receiver? or can I go right from my computer's sound card to the receiver?

Thanks!

 

[annunaki]

When audio is compressed to the Mp3 format it is just that. COMPRESSED. A higher bit rate will not matter because the signal was still compressed to conform to the Mp3 format. I personally can hear a fairly significant difference between the original cd and Mp3.
If the recording is of the current (louder is better) philosophy, it becomes difficult to distinguish between the two. As so much compression is already done in the recording process. Many Mp3's and current recordings I have heard, do not sound any better than a well recorded cassette. In fact, some of the cassettes (depending upon the artist) probably have a wider dynamic range.

To spend $1000 on high end speakers, which by nature are designed to extract the maximums in dynamic range and fidelity, would be a waste with just mp3's. Basically the speaker is now limited by the source. The speaker is now capable of more than what the source will allow it to produce.

Sorry for the rant, but this is a sore subject for me. Because of the popularity of an inferior source (mp3), many artists have stopped putting as much effort (as in the past) into the fidelity of their recordings. They figure, "why spend the extra money when you will not be able to hear it (difference in quality, dynamic range, ect.) anyway."

 

[jaxvon]

Indeed, to reiterate what others have said...

Any MP3 below 192k encoding will sound like garbage. 192 isn't too bad, but anything higher is already better. If you can, I'd suggest that you rip your CDs in wav format so you don't get any compression. For the record, so-called "lossless" audio codecs are NOT the same as a wav file, they sound obviously inferior.

About the sound card on your computer, I'd suggest that you get a high-quality sound card with digital coax output (like the M-Audio Revelation 5.1 or 7.1) and let your reciever do the D-A conversion. Otherwise, you'll probably end up with computer noise in your system, which sounds bad.

 

[MDS]

Annunaki is confusing dynamic compression with perceptual coding - they are vastly different.

Perceptual codecs (mp3, wma, ogg) do exactly what voodoothunder learned in class. They use a 'perceptual model' of how the brain and ears hear and interpret sound, analyze the signal, and discard those parts that the model says we would not hear anyway. For example, a soft sound immediately preceding or following a loud sound will be 'masked' by the loud sound and not heard - so no reason to encode that bit of music. Is it perfect? No, but for the vast majority of music it works really well.

The 'compression' aspect when talking about perceptual codecs is data compression; ie reducing the size of the data (by discarding what we wouldn't hear anyway). Converting a wav file to mp3 will not change the average power level by any appreciable amount. The whole purpose of dynamic compression on the other hand is to do just that - increase the average power level by reducing the peaks according to whatever threshold and ratio you set.

And for the record, if a 'lossless' codec changes the data so that it sounds vastly inferior, then it is not lossless at all. Lossless codecs do not discard data the way perceptual codecs do - they simply reduce the size of the data by encoding it more efficiently, in the same way that PKZip compresses a file to a smaller size (although the algorithms used are not similar to simple zip algorithms which are usually huffman encoded). After decoding a lossless audio format it will be bit for bit identical to the original.

 

[mustang_steve]

On my PC (uses an Audigy2 ZS sound card [yeah not the best audiophile card, but i game on this computer too, plus having dvd-a capability is nice], Creek OBH-11 headphone amp, and Grado SR-125 headphones), I can easily hear the difference in MP3 and CD...actually, I can hear the difference between DVD-A and CD. It's all a matter of how much detail the end system can put out.

I know on my loudspeakers I could only tell the difference on some MP3's, usually cymbal-heavy rock music....cymbals tend to be one of those things that when they do show compression artifacts, they show in the worst way....listening to Creed, the cymbals were so distorted at 192KBPS, that even on a pair of freebie sony phones and $5 computer speakers, I could hear the artifacting on the cymbals...

Basically it's hit and miss, but I found that the tracks that sound bad in MP3, often sound good in OGG or WMA, so it might be good to use a media player that does multiple formats

 

[voodoothunder]

Thanks everybody.

Like I said I've never put a collection of mp3s through a higher end system so I really can't comment. Also being out of school for a few years has made me forget alot of the techno-babble.

I'm not a huge fan of MP3s only because of small things really. I prefer open software and MP3 is now closed. I do get better file sizes with OGG, unfortunately I'm having trouble finding MP3 players that support OGG Vorbis.

If I can remember correctly, whether it is with certain codecs or not, you can gate the input during encoding to optimize file sizes (even with the extra CD information that the ear can't pick up). So I can remember having a lot of issues with clipping and static problems.

Jaxvon, is D-A digital-to-analog conversion? I think I can remember reading somewhere that there is a seperate component that converts the signal. So I've always wondered why these weren't built into receivers. Like I said I'm a painful newbie.

Thanks a million.

 

[annunaki]

Annunaki is confusing dynamic compression with perceptual coding - they are vastly different.

Perceptual codecs (mp3, wma, ogg) do exactly what voodoothunder learned in class. They use a 'perceptual model' of how the brain and ears hear and interpret sound, analyze the signal, and discard those parts that the model says we would not hear anyway. For example, a soft sound immediately preceding or following a loud sound will be 'masked' by the loud sound and not heard - so no reason to encode that bit of music. Is it perfect? No, but for the vast majority of music it works really well.

The 'compression' aspect when talking about perceptual codecs is data compression; ie reducing the size of the data (by discarding what we wouldn't hear anyway). Converting a wav file to mp3 will not change the average power level by any appreciable amount. The whole purpose of dynamic compression on the other hand is to do just that - increase the average power level by reducing the peaks according to whatever threshold and ratio you set.

And for the record, if a 'lossless' codec changes the data so that it sounds vastly inferior, then it is not lossless at all. Lossless codecs do not discard data the way perceptual codecs do - they simply reduce the size of the data by encoding it more efficiently, in the same way that PKZip compresses a file to a smaller size (although the algorithms used are not similar to simple zip algorithms which are usually huffman encoded). After decoding a lossless audio format it will be bit for bit identical to the original.

Anonymous,

Thanks for that clarification. I was under the understanding that a perceptual codec like Mp3 used perceptual coding as well as dynamic compression at the same time? I guess I was mistaken.

The problem I have is that most Mp3 users are not using lossless compression or codec tecniques. They are downloading lossy files and in turn they sound significantly inferior. By doing this they are limiting the capability of a high end audio system. Essentially defeating its purpose though it is doing exactly what it was intended for. Reproducing the material as true as possible.

I would download music IF I could be sure that it would sound exactly the same as it does on the pre-recorded cd. I could care less about file sizes or how many songs I can get on one cd.

 

[MDS]

About dynamic compression and lossy codecs

Mustang_steve mentioned Creed and I just bought their Greatest Hits CD, so let's use that as an example.

The My Own Prison wave file that I ripped from the CD has the following statistics (according to SoundForge):

RMS power: Left = -11.24 dB, Right = -11.70 dB.
Min sample value: -32,768 (0dB, -100%)
Max sample value: 32,762 (-0dB, 99.98%)

This is pretty typical of modern CDs. Very loud (-11dB) and lots of peaks right at 0dB. It never lights the clip indicators, but the Peak meter hits zero every second. The 'crest factor' is the difference between average level and peak level. For this song, the level meters occasionally drop as low as -12dB, but for the majority of the song they never go below -6dB. So the entire dynamic range of the song has been compressed to roughly the top 6dB.

Converting the wav to mp3 (192 kbps):
L = -11.34 dB, R= -11.81 dB.
The min and max sample values are now both at exactly 0dB (-32,768 and +32,767 respectively). So, the rms power was decreased ever so slightly, but there are now far more peaks at 0dB and the clip indicators light repeatedly throughout the whole song.

The clipping may be audible on some systems, but usually not because the interval is very short. I think for these cases, some people may very well notice a difference that they don't like. Now, is that a condemnation of the mp3 format itself or should we blame the recording engineers that pushed the limit and compressed the song to be as hot as possible?

I can hear the occasional sibilance and 'not quite right' sound on the mp3 too, but it really isn't much worse than the wav. I accept the subtle degradation in sound quality, but place the blame on the original mastering of the CD.

Compare that to an older CD that was mastered conservatively (say -18dB) and no peaks anywhere near zero. I won't bore you all with an example of the statistics for a such a file, but suffice it to say that it exhibits none of the subtle 'artifacts' that does the Creed example and it sounds excellent.

So it really is dependent on the quality of the source, but for my ears mp3 works well enough that I like the convenience of smaller file sizes and almost-as-good sound quality. Naturally, I keep the wavs around too for my 'archive' of my cd collection, but I generally play the mp3s.

 

[mustang_steve]

I wasnt pointing fingers at either direction, just saying some songs sound better encoded in WMA or OGG.

Sadly my Nomad Zen can't do OGG, so I'm stuck with MP3 and WMA


edit: by the way, even some classical pieces show some distortions under MP3, Grieg's "in the hall of the mountain king" is a perfect one....I lost the information on what orchestra played that rendition, but the violins were extremely muddy at the peak of the piece.

 

[karltl]

Excellent example!

Mustang_steve mentioned Creed and I just bought their Greatest Hits CD, so let's use that as an example.

The My Own Prison wave file that I ripped from the CD has the following statistics (according to SoundForge):

RMS power: Left = -11.24 dB, Right = -11.70 dB.
Min sample value: -32,768 (0dB, -100%)
Max sample value: 32,762 (-0dB, 99.98%)

This is pretty typical of modern CDs. Very loud (-11dB) and lots of peaks right at 0dB. It never lights the clip indicators, but the Peak meter hits zero every second. The 'crest factor' is the difference between average level and peak level. For this song, the level meters occasionally drop as low as -12dB, but for the majority of the song they never go below -6dB. So the entire dynamic range of the song has been compressed to roughly the top 6dB.

Converting the wav to mp3 (192 kbps):
L = -11.34 dB, R= -11.81 dB.
The min and max sample values are now both at exactly 0dB (-32,768 and +32,767 respectively). So, the rms power was decreased ever so slightly, but there are now far more peaks at 0dB and the clip indicators light repeatedly throughout the whole song.

The clipping may be audible on some systems, but usually not because the interval is very short. I think for these cases, some people may very well notice a difference that they don't like. Now, is that a condemnation of the mp3 format itself or should we blame the recording engineers that pushed the limit and compressed the song to be as hot as possible?

I can hear the occasional sibilance and 'not quite right' sound on the mp3 too, but it really isn't much worse than the wav. I accept the subtle degradation in sound quality, but place the blame on the original mastering of the CD.

Compare that to an older CD that was mastered conservatively (say -18dB) and no peaks anywhere near zero. I won't bore you all with an example of the statistics for a such a file, but suffice it to say that it exhibits none of the subtle 'artifacts' that does the Creed example and it sounds excellent.

So it really is dependent on the quality of the source, but for my ears mp3 works well enough that I like the convenience of smaller file sizes and almost-as-good sound quality. Naturally, I keep the wavs around too for my 'archive' of my cd collection, but I generally play the mp3s.

This is an excellent writeup and demonstrates a curious side effect of the conversion from analog recording methods to digital. I used to be a partner in a recording studio years ago. At the time, digital recording systems were just coming out. We used 2", 24 track MCI tape machines for master recording. One of the things about using analog tape is it's ability to manage (within reason) record levels above 0db often without introducing any perceptable distortion.

As I recall, in those days it was the habit to try and drive the tape as hot as possible (again within reason) as this helped to address/reduce the normal noise floor of analog tape and the associated electronics.

Where digital recording is concerned, it's been my understanding that 0db is effectively a "hard stop" and that running input levels above this point cause hard clipping of the signal when it is converted back to analog form to send to the speakers. Imagine a sine wave with tops that are clipped flat. It effectively becomes a square wave (anyone who's fiddled with synthesizers might recognize that using a square wave source gives you a particular/distorted type of sound. I'd always related these two things as being similar.

 

[krabapple]

When audio is compressed to the Mp3 format it is just that. COMPRESSED. A higher bit rate will not matter because the signal was still compressed to conform to the Mp3 format. I personally can hear a fairly significant difference between the original cd and Mp3.

? *data* compression can be totally inaudible -- hence lossless compression formats such as FLAC and MLP (the latter used for DVD-Audio).

Bit rate absolutely *does* matter for the sound of *lossy* data compressed formats, as does the codec, and software used, and the source material, and the playback rig, and the listener's training for ability to hear mp3 artifacts. Using variable bitrate mp3 encoding with LAME with ~192 average, several people for whom I've made
A/B/X test discs have been unable to tell mp3 from source; this accords with results
published on the hydrogrenaudio.org website, where codec developers and users
congregate, and where all comparisons are done using blind protocols.

see their recommended mp3 settings here:

http://www.hydrogenaudio.org/forums/index.php?showtopic=28124

If the recording is of the current (louder is better) philosophy, it becomes difficult to distinguish between the two. As so much compression is already done in the recording process. Many Mp3's and current recordings I have heard, do not sound any better than a well recorded cassette. In fact, some of the cassettes (depending upon the artist) probably have a wider dynamic range.

Mp3s are NOT inherently *dynamically* compressed. You are confusing two meanings of compression.

 

[jaxvon]

To voodoothunder, D/A is indeed digital to analog conversion. Most recievers have this function and do it via their digital inputs (usually coaxial and optical S/PDIF). I would suggest using the digital out on your soundcard so you can just run one cable from your computer to reciever. This also allows you to utilize the high-quality DACs (digital-analog converters) built into your reciever.

 

[annunaki]

? *data* compression can be totally inaudible -- hence lossless compression formats such as FLAC and MLP (the latter used for DVD-Audio).

Bit rate absolutely *does* matter for the sound of *lossy* data compressed formats, as does the codec, and software used, and the source material, and the playback rig, and the listener's training for ability to hear mp3 artifacts. Using variable bitrate mp3 encoding with LAME with ~192 average, several people for whom I've made
A/B/X test discs have been unable to tell mp3 from source; this accords with results
published on the hydrogrenaudio.org website, where codec developers and users
congregate, and where all comparisons are done using blind protocols.

see their recommended mp3 settings here:

http://www.hydrogenaudio.org/forums/index.php?showtopic=28124





Mp3s are NOT inherently *dynamically* compressed. You are confusing two meanings of compression.

Read my posts after the original, this has been addressed. I am sure a properly done Mp3 can sound very much like the original cd. However, the VAST majority of Mp3s are made using a lossy codec and do not sound like the original.

 

[miklorsmith]

mp3's are inherently lossy

There is no "lossless" mp3, it's only a matter of how much loss there is. Uncompressed CD's, or .wav files in compu-speak, run at a bitrate of 1,411 kb/s. The .mp3 format is capable of no better than 320 kb/s. This means that AT BEST, 78% of the signal has been removed. At 192 kb/s, the removal percentage increases to 86%.

It is frequently said that the part that is removed is the part you don't hear. Logically, it is hard to imagine that the remaining 14% constitutes 100% of what you hear. In practice, reasonably resolving equipment quickly shows the reduced content to be sonically inferior. Body and dynamics become highly compromised as bit-rate decreases.

All compression software is not created equal. Variable bitrate (VBR) encoding will produce a superior result at a given figure. Also, the Exact Audio Copy/Lame encoding combination so promoted (free) at Hydrogen Audio is obviously superior to three other flavors I have tried.

I love .mp3 music. It sounds very good and is highly portable. It is not "CD quality" at any bitrate and is only played in my home at parties. Fine as background music, but for critical listening you're only cheating yourself.

Your best bet, as stated elsewhere, is to keep your .wav files intact. Then, as encoding software improves, you are prepared without re-ripping all your discs over again. Hard drive storage is so cheap now that it's cheap to keep a second copy somewhere else.

Second on the thought about downloaded files. When I looked about 6 months ago, all the services offered 128 kb/s, constant bitrate (CBR). This is nearly intolerable for anyone not listening to pop music.

 

[MDS]

There is no "lossless" mp3, it's only a matter of how much loss there is. Uncompressed CD's, or .wav files in compu-speak, run at a bitrate of 1,411 kb/s. The .mp3 format is capable of no better than 320 kb/s. This means that AT BEST, 78% of the signal has been removed. At 192 kb/s, the removal percentage increases to 86%.

That is not even remotely close to accurate. So you took 320 kbps and divided by 1411 kbps to come up with 78% loss??? The bit rate has been reduced by 78%; it hasn't discarded 78% of the signal.

The typical 5.1 Dolby Digital (DD is 'lossy' too) track runs at 384 kbps. The bit rate naturally can be varied if you encode fewer than 6 channels, but let's say we encoded 2 channels at the full bit rate of 384 kbps. So by the above logic, DD has discarded 73% of the signal? Not in any way shape or form!

Again, if anyone truly believes that mp3 or other lossy codecs produce vastly inferior sound compared to an uncompressed PCM version, so be it - but let's not invent whacky explanations that stem from a complete misunderstanding of how they actually do what they do.

 

[miklorsmith]

Terms

OK, 78% or 86% of the original information. I said signal, whatever. And, if you can't hear the difference, get some real gear.

If you have a better explanation than the even-more-dubious "throws away the unhearable stuff", have at it.

 

[MDS]

I'd be happy to give a layman's explanation of how lossy codecs work. By neccessity it would have to be slightly more complicated than 'discards what we cannot hear', but I'm sure you will quickly revert back to your convenient 'audiophile' misconceptions you've posted in countless other threads, rather than make an attempt to truly understand it.

 

[miklorsmith]

The floor is yours. . .

I'm sure others would also be interested. The more scientific would probably also appreciate a detailed explanation.