DJing with MP3s!

[Blaze]

why not?

 

I mean, its not testing the differences between the files, but the ability of the listener to distinguish those differences..

 

I dont see why a blind test would not work.. I mean, sure, if by ´quick´ you mean using some 10 sec samples to compare and not a few whole tracks, then you could be right... but making blind tests with a few whole tracks I guess is effective enough, no?

I hope it would be

Can't find that article, but AFAIR it's main idea was that if you are used to MP3s and start listening to CDs, you're gonna hear a difference and vice versa, but listening to short samples or even a complete album may be not enough.

[Acido Domingo]

afaik, the default of mp3 is to cut off all frequencies above 20khz and below 20hz, which supposedly are the limits to the human hearing too.... So whether they cut more bass than high frequencies depends if the original file had much information lower to 20hz or above 20khz..

It's a lot more complicated; what gets left out depends on the entire spectrum of sound at the moment and even before. Like a loud hihat won't let you hear much of the high-frequency spectrum even a few dozen milliseconds after. This "psychoacousic model" is the one thing besides speed that makes encoders different, and most had years of research put into them.

if someone is willing to set up a blind test, that would be interesting... pavel maybe? (though I dont want to test myself with some full on track with billions of sounds at the same time.. rather some good production detailed fat progressive, like antix) ....

Cool, let's do it like this: someone could send me a couple of WAV snippets from various CDs, say a minute each. I'll encode them with a few different codecs and bitrates (I can do LAME and OGG, maybe someone else do Fraunhofer's, XING or whatever is teh shit in the closed-source world) and convert them back to WAV, the same way they'd be decoded when played live. Then people can download them and listen to them as often as they like and on the best equipment everyone can get their hands on, and finally everyone has to rank them in order of sound quality and identify the original (might be someone recognizes a compressed version but actually finds the sound more pleasant).

[sherlockalien]

It's a lot more complicated; what gets left out depends on the entire spectrum of sound at the moment and even before. Like a loud hihat won't let you hear much of the high-frequency spectrum even a few dozen milliseconds after. This "psychoacousic model" is the one thing besides speed that makes encoders different, and most had years of research put into them.

yeah I guessed the whole process of compressing would be more complex, but appart from all that, mp3s in the end do only keep the 20-20k hz range of frequencies, right?

 

btw.. ´psychoacoustic model´ seems interesting.. learning music theory is in my ´to do´ list for sure

 

 

Cool, let's do it like this: someone could send me a couple of WAV snippets from various CDs, say a minute each. I'll encode them with a few different codecs and bitrates (I can do LAME and OGG, maybe someone else do Fraunhofer's, XING or whatever is teh shit in the closed-source world) and convert them back to WAV, the same way they'd be decoded when played live. Then people can download them and listen to them as often as they like and on the best equipment everyone can get their hands on, and finally everyone has to rank them in order of sound quality and identify the original (might be someone recognizes a compressed version but actually finds the sound more pleasant).

get your ass on msn..

 

I will choose some cd here.

 

maybe someone else who argued they can notice the difference between mp3 should volunteer too, and prove themselves

[Blaze]

It's a lot more complicated; what gets left out depends on the entire spectrum of sound at the moment and even before. Like a loud hihat won't let you hear much of the high-frequency spectrum even a few dozen milliseconds after. This "psychoacousic model" is the one thing besides speed that makes encoders different, and most had years of research put into them.

Cool, let's do it like this: someone could send me a couple of WAV snippets from various CDs, say a minute each. I'll encode them with a few different codecs and bitrates (I can do LAME and OGG, maybe someone else do Fraunhofer's, XING or whatever is teh shit in the closed-source world) and convert them back to WAV, the same way they'd be decoded when played live. Then people can download them and listen to them as often as they like and on the best equipment everyone can get their hands on, and finally everyone has to rank them in order of sound quality and identify the original (might be someone recognizes a compressed version but actually finds the sound more pleasant).

It's a good idea, but I think people listening at home with different hardware can lead to very diverse results. Wouldn't it be better if several people got together and listened?

[Acido Domingo]

yeah I guessed the whole process of compressing would be more complex, but appart from all that, mp3s in the end do only keep the 20-20k hz range of frequencies, right?

Yes, but that doesn't make the difference to CDs because that's what a CD player does as well. Stuff below 20 Hz is usually lost in the rumble filter and since 22.05 kHz is the absolute upper limit for a CD sampler anyway you have to cut it off a little below that both to smoothen the sound a little at high frequencies (a 22 kHz sine wave digitized on a CD becomes a square wave!) and to keep aliasing artifacts out. If you sample with frequency f (44.1 kHz for CDs), all frequencies in the sound appear "mirrored" at f/2, so if an instrument produces a 25 kHz sound that we can't hear, a "mirror image" of this would appear at |25-44.1|=19.1 kHz which we can hear. Because you can't physically build a filter that filters nothing below 22.05 kHz and everything above it, it has to start a bit earlier, like at 20.

btw.. ´psychoacoustic model´ seems interesting.. learning music theory is in my ´to do´ list for sure

It's interesting but probably more technical than you'd think, doesn't have much to do with psychology, more mechanics of the ear and nerve impulses.

get your ass on msn..

 

I will choose some cd here.

 

maybe someone else who argued they can notice the difference between mp3 should volunteer too, and prove themselves

Uhm...what's your nick there again?

[NEMO.BOFH]

Live acts?? Maybe you are talking about ''live'' acts cause as far as i remember in the the before laptop's ''century'' things were different

you dont get teh point, or you dont want to get the point? as much as I love you, STOP BEING A DICK!

[reger]

It's a lot more complicated; what gets left out depends on the entire spectrum of sound at the moment and even before. Like a loud hihat won't let you hear much of the high-frequency spectrum even a few dozen milliseconds after. This "psychoacousic model" is the one thing besides speed that makes encoders different, and most had years of research put into them.

Cool, let's do it like this: someone could send me a couple of WAV snippets from various CDs, say a minute each. I'll encode them with a few different codecs and bitrates (I can do LAME and OGG, maybe someone else do Fraunhofer's, XING or whatever is teh shit in the closed-source world) and convert them back to WAV, the same way they'd be decoded when played live. Then people can download them and listen to them as often as they like and on the best equipment everyone can get their hands on, and finally everyone has to rank them in order of sound quality and identify the original (might be someone recognizes a compressed version but actually finds the sound more pleasant).

some might use audio editors/spectrographs [or how was it called?] to cheat

[Acido Domingo]

some might use audio editors/spectrographs [or how was it called?] to cheat

Maybe Pavel could tell the difference...

There was this contest on a phoneticists' site where the admin posted some spectrogram of a word, and people had to guess what was said. That was hard enough for trained pros who had seen and analyzed thousands such spectrograms.

[sherlockalien]

It's a good idea, but I think people listening at home with different hardware can lead to very diverse results. Wouldn't it be better if several people got together and listened?

yeah in general it will be diverse results, yes... but the primary question, if that specific person can notice the difference of mp3 and wav in his own system, is still answerable by this test...

 

Yes, but that doesn't make the difference to CDs because that's what a CD player does as well. Stuff below 20 Hz is usually lost in the rumble filter and since 22.05 kHz is the absolute upper limit for a CD sampler anyway you have to cut it off a little below that both to smoothen the sound a little at high frequencies (a 22 kHz sine wave digitized on a CD becomes a square wave!) and to keep aliasing artifacts out. If you sample with frequency f (44.1 kHz for CDs), all frequencies in the sound appear "mirrored" at f/2, so if an instrument produces a 25 kHz sound that we can't hear, a "mirror image" of this would appear at |25-44.1|=19.1 kHz which we can hear. Because you can't physically build a filter that filters nothing below 22.05 kHz and everything above it, it has to start a bit earlier, like at 20.

 

I guess I will have to trust on you for this one

 

It's interesting but probably more technical than you'd think, doesn't have much to do with psychology, more mechanics of the ear and nerve impulses.Uhm...what's your nick there again?

yeah I bet it doesnt have to do with psychology.. but its the technical part I would be after anyways.. I was always bad at maths but loved physics, and music/vibrations being an integral part of my ideas, I would like to learn the patterns in their workings and make sense out of some things I need to know

 

my display nickname there is Rafael

[slyman2.0]

A different guy slyman604

 

naa..thats my old account, this is new improved 2.0 version of me...or i'm dumbass, cant remember my password and dont have that email address anymore.....

[slyman2.0]

in that article i posted up top, it said something interesting in that some people who actually have hearing problems in one ear probly have a better chance of hearing the difference with mp3 because of the way that breaks down the psychoacousic model.

 

i think a test would be fun, get as many people to do it as want to. i would be up for it and dont expect to hear the difference.

 

a minute would be some big files. i think it would be far better to have 20 - 10 second clips, 10 guesses on what one is what. i would rip them but there is no way i could resist making them all actually wav or all mp3 and fuck with everyone

[Colin OOOD]

Maybe Pavel could tell the difference...

There was this contest on a phoneticists' site where the admin posted some spectrogram of a word, and people had to guess what was said. That was hard enough for trained pros who had seen and analyzed thousands such spectrograms.

Sorry but no. The difference between ANY compressed and uncompressed format is blatantly obvious when viewed with a spectrum analyser. MP3s all have everything above a certain point cut off, the exact point and nature of the audio above the cutoff point being determined by the encoder and bitrate.

 

The only way a blind test can tell us anything is if it takes place under controlled conditions.

[sammyhkhan]

It's a lot more complicated; what gets left out depends on the entire spectrum of sound at the moment and even before. Like a loud hihat won't let you hear much of the high-frequency spectrum even a few dozen milliseconds after. This "psychoacousic model" is the one thing besides speed that makes encoders different, and most had years of research put into them.

Cool, let's do it like this: someone could send me a couple of WAV snippets from various CDs, say a minute each. I'll encode them with a few different codecs and bitrates (I can do LAME and OGG, maybe someone else do Fraunhofer's, XING or whatever is teh shit in the closed-source world) and convert them back to WAV, the same way they'd be decoded when played live. Then people can download them and listen to them as often as they like and on the best equipment everyone can get their hands on, and finally everyone has to rank them in order of sound quality and identify the original (might be someone recognizes a compressed version but actually finds the sound more pleasant).

Wicked idea! Let's bloody do it!

 

We could just use a simple poll? Who is keen to provide Acido Domingo with some files?

 

I would do it, but I just do not know how to cut WAV files into 1 minute snippets. Give me some advice and I will do it...

[Acido Domingo]

Sorry but no. The difference between ANY compressed and uncompressed format is blatantly obvious when viewed with a spectrum analyser. MP3s all have everything above a certain point cut off, the exact point and nature of the audio above the cutoff point being determined by the encoder and bitrate.

 

The only way a blind test can tell us anything is if it takes place under controlled conditions.

Looks like it.... I stand corrected!

But then...that's pretty low bitrates, and probably a sample read directly from CD, not postprocessed through the filters that are there in any CD player. In a 320 kbps file, more bits will be assigned to the lower frequencies as well, so it wouldn't just double the frequency range as compared to 160 kbps, but I still wonder if it couldn't be pushed close enough to 20 kHz that the difference is hard to tell.

 

Wicked idea! Let's bloody do it!

 

We could just use a simple poll? Who is keen to provide Acido Domingo with some files?

Well, Reger and Colin still have a point, if people just used a spectrum analyzer it would make telling at least the low-bitrate compressed files trivial.

[frozen dream]

IMO the problem is that Mp3s (even in highest quality) always sound/feel inferior to CD quality/WAV (even FLAC) when played at very high volume.

I guess it depends on what kind of environment you are djing in and if you are satisfied by the sound quality yourself. (And maybe try to get WAV files, sometimes the labels offer them also directly).

I also find it more rewarding to have a physical copy- somehow it lets me appreciate the music more and connect to it better. However if it really is a pain to buy/order psy in NZ and takes too long for your needs, the speed and ease of obtaining it online may be more important to you and makeup for it.

Next is......which program do you intend to use for djing or do you intend to burn the files and play them traditionally (if you have equipment at hand)?

Overall I think there is still alot of prejudice and disrespect for "lap top" djs, simply because it narrows the barriers of entry for djing tremendously , takes away some of the basic skills (beatmatching...) and automates many processes (cueing, music management etc). Also some people assume all lap top djs play pirated music:/ However in future this will surely change and become more accepted as people realise the possibilities and benefits these programs offer.

hehe i've had the mp3-cd thing too, but let me explain

; i listen bizarre contact - love simulation in mp3 on my computer(192 kbit), a day later i lost the file so what do i do, i rip it from the cd burned one day earlier, and as i listen to it again (on the same speakers yes) and what do i notice? it really appears to be more in-depth, fatter, rounder, cleaner, crisper sound, surely something in the nero program. funny that it's just the same file but only ripped after burning.(i did not check level volumes option before burning!)

 

question: do some cd burning programs add or diminish quality?

 

for nero i'd say yes (add to quality, smooth them and gentle push) but i don't know for sure for other programs

 

Oh noes...

 

I agree that I can listen to some background music, like radio, or ass-shaking house/electro in almost any quality, but when it comes to the 'state of trance' I can't appreciate low-quality tracks with frequencies cut off.

 

No high-frequencies and these dithered basslines... omg!

i've already had 128 kbps tracks that sounded better than 192 or even 320 kbit tracks, with no ressy basslines or whatever, it all depends on how and with what it is encoded. mostly you don't know if a file has been reconverted from a lower to a higher bitrate. you need a very powerfull spectrum analyzer like the one in pro tools or cool edit pro to tell for sure.(if you know how to interprete all that shiet) (i've seen a super audio editor once, the price for just the software was over $2500!!)

 

for instance on mac 128 kbps is considered 'high' bitrate while in windows 320 is the highest, mac just uses other convertors/encoders shiet who produce better results

 

I'll tell you from my experience... spinning with mp3's just doesn't feel right... I prefer cd...

to me as well, i've always spun with mp3 burned on cd or originals, i once had to do it with a laptop because of a burn-error (my cds didn't work ) and it didn't feel right indeed, mainly because you have the hang of controlling eqs and faders, and you dont have the ability to control 2 or even 3 things at once(headbanging frustrating to a real dj), though if i never got the hang of spinning with cds and pitch/beatmatching, i guess i wouldve picked up laptop spinning quite easily, since its so damn easy to use (and u get spiffy fx in most programs+sample playback) for which u soon need an extra cd deck in an old-style setup

 

Loosen up people, you drive many miles to god's forsaken place and you trouble yourself with such minor issue?

haha, splendid man!

 

i don't know everything about mp3's but isnt especially the bass cut off with the making of mp3's?

 

if so with classical music there is less bass i assume so might it be easier to find the difference between wave and mp3 in psy than in classical (the musical type they used in the article referred to above)?

mp3 encoding includes the 2Ohz-2Okhz cut + it also removes 'backdrop' noises so tracks become thinner and more linear, only these traits only become audible in the real low bitrates, let's say 128 kbps with an average encoder, but really these things do differ a lot from encoder to encoder, lame_enc (the most common) does a pretty good job and has even 128 kbps good for full playback, on a big installation this is nevertheless a no-go and 192 is the absolute minimum.

 

conclusion: it's better to trust your ears and gut feeling on judging quality, and be very carefull of comparing material with eachother (knowing it's encoding/origin)

[slyman2.0]

Sorry but no. The difference between ANY compressed and uncompressed format is blatantly obvious when viewed with a spectrum analyser. MP3s all have everything above a certain point cut off, the exact point and nature of the audio above the cutoff point being determined by the encoder and bitrate.

 

The only way a blind test can tell us anything is if it takes place under controlled conditions.

 

comeon man, if anyone does spetrum analysis in a contest thats just in good fun like this they seriously need to get laid.

if this is a concern though then do it in lame. lame takes raw audio, compresses it, then plays its back uncompressed without loss of data. if anyone wants to take the pepsi challenge they can hear raw audio being uncompressed then lets put some money on the line with paypal, but it would have to be like 50 files to get a good statistical sample and not just luck.

 

i'm no hippy. i make my living trading stock index futures, love to play poker, love any game that in involves real money and probility theory combined with human uncertainty. this is a joke odds wise, i'll bet against anyone who wants to take the lame vs wav challenge.

[Blaze]

Sorry but no. The difference between ANY compressed and uncompressed format is blatantly obvious when viewed with a spectrum analyser. MP3s all have everything above a certain point cut off, the exact point and nature of the audio above the cutoff point being determined by the encoder and bitrate.

 

The only way a blind test can tell us anything is if it takes place under controlled conditions.

+1

We have to make out some rules and hope people don't cheat.

Also, AFAIK, Exact Audio Copy is considered the best program for ripping files, but if anyone who understands this stuff knows a better one I have no problem.

And doing 10-20 sec. samples sounds wrong to me, it has to be full-length songs and from different genres, not only psytrance.

IMHO

[thorn726]

screw it , im bumping a month later.

 

i've done so many tests with BIG ASSS systems it is silly.

 

i haven't found anyone yet who can't tell the difference on my home system, i am routinely demonstrating why low bit mp3s suck to ignorants

 

all you need for a good demo is a decent subwoofer.

 

play 128, 192, 320, wav. they all sound different

 

hearing sensitivity can be an individual thing also.

still, disturbing that the majority here is still in favor of playing to a bone head mass audience, and determining what they can get away with rather than how to give the people the best possible experience

[AkumaSama]

any DJ playing mp3s in a club must be in a shit club with a shit sound system.

 

and a dj playing mp3s?>?? never even heard of it, vinyl yes, CDs sometimes seen

[sammyhkhan]

Me thinks MP3s are the shizzle

[Nexus]

screw it , im bumping a month later.

 

i've done so many tests with BIG ASSS systems it is silly.

You told me in another thread that you've never done a test before

What you are talking about is not testing.

 

i haven't found anyone yet who can't tell the difference on my home system, i am routinely demonstrating why low bit mp3s suck to ignorants

placebo, placebo, placebo, did I mention placebo?

Do this, invite a friend and play both lossless but do what you do normally when playing those files like one is mp3 the other one is not. Guess what, your friend will choose one over the other even though they are identical.

 

all you need for a good demo is a decent subwoofer.

demo != test.

 

play 128, 192, 320, wav. they all sound different

128 kbps has loss that most people can tell on decent equipment. Proper 192 is close to transparent on most trance music.

 

hearing sensitivity can be an individual thing also.

It IS individual, not just can be. Your 'I haven't found anyone who can't tell' tells me that your 'demo' has nothing to do with a real test.

Stop pretending and do a real test, rip properly with EAC, compress properly into V0 with lame and play using ABX (winabx,foobar,etc), try it on your friends, if you find anyone who can tell the difference much over chance then you have someone with EXCEPTIONAL ears and if they are still not in the music/sound business they are wasting their talent.

[Acido Domingo]

all you need for a good demo is a decent subwoofer.

Subwoofer, eh? Audio quality. Yup.

[frozen dream]

man, when i render unmastered tracks to 192 they sound very bad

 

in 320 they are ok

 

though mastered tracks have no difference in wav, 320 kbps and 192 kbps(lame_enc), and i listen them on very expensive monitoring headphones, and i also do very high volume tests with my crappy hi-fi speakers (up to +120 db)

 

i think 320 kbps (CBR) still is a good contender to wav, for home listening that is.

 

for the very high volmue levels at a party wav is your friend i would say