audio questions

[Davros]

Why not, your supposed to have a good ear (pity you've only the one)

 

[L233]

I'm actually most interested hearing from _xxx_. After all, he claims he can hear the difference between MP3s and CD on his car stereo with the sort of material I posted.

 

[Neb]

Whats this, no Hammerfall?

 

[_xxx_]

For that I'll first need internet at home, which will happen end of the week. Proxy won't let me dl it at work.

 

[Bambers]

Why not, your supposed to have a good ear (pity you've only the one)

I've never claimed to have a good ear, or just one for that matter.

I have picked out what I think are the 128kbps samples from the last 3 of the 4 samples, I must admit on the whole they all seem to have compressed better than the piano, then again most aren't my kind of music so I've no idea what to listen for as 'correct'.

Particularly the first one which all sounded horrible to me.

*runs from the deathmetal fans

 

[Albuquerque]

I like some older metal, but not much of the new stuff. Gimmie some Rob Zombie or some of the old Black album from Metallica...

 

[_xxx_]

Ok, now I could dl it.

The Amon Amarth example just sucks. Whatever the original, it already sounds crappy. In my ears zxlpw could be the original, wtsnc is 320 kbit. But as said, whoever produced it is a moron.

Nevermore: xcipw sounds like the original, zdqnr could be 320.

Didn't bother with the rest. Now let's see how it turns out

 

[_xxx_]

L233, where art thou?

And won't anybody else try his ears/luck? For comparison.

 

[Simon F]

And won't anybody else try his ears/luck?

Only if I was "within large concrete bunkers some thirty-seven miles from the stage"

 

[L233]

Ok, now I could dl it.

The Amon Amarth example just sucks. Whatever the original, it already sounds crappy. In my ears zxlpw could be the original, wtsnc is 320 kbit. But as said, whoever produced it is a moron.

Heh, it's actually my favourite Amon Amarth album It was mixed by Peter Tägtgren.

zxplw is the 192 vbr. wtsnc is the 128kb CBR. yjkdh is the PCM, iuzst is the 320.

Nevermore: xcipw sounds like the original, zdqnr could be 320.

This is weird. Again, xcipw is the 192 vbr and zdqnr is the 128 CBR... PCM is wrtkn and ybqlf is the 320 kbit one.

In both cases you thought that the 192 vbr is the original and that the most compressed is the least compressed.

This also means that the original and the 320 kbit sounded "worst" to you.

You couldn't identify the original as the original and the 128kb as the most compressed, which is surprising, since you said that, with that kind of material, "missing harmonics" and missing cymbals were apparent to you as a musician, even on a car stereo.

Didn't bother with the rest. Now let's see how it turns out

Try the Destruction sample, at least.

 

[Bambers]

Hmm I had xcipw as the 128, rest I couldn't tell apart.

the Destruction sample I have twhgv as the 128, again rest I can't pick between.

Overall for both those samples though they essentially all sound the same

The shadow gallery one:

ygtha is the 128, bit more confident on this one than the other samples

and on a much less confident note and some (lots ) guessing I'd put zwqsf as the 192, xlkgw as the 320 and vtrzx as the original

 

[L233]

Hmm I had xcipw as the 128, rest I couldn't tell apart.

the Destruction sample I have twhgv as the 128, again rest I can't pick between.

Overall for both those samples though they essentially all sound the same

The shadow gallery one:

ygtha is the 128, bit more confident on this one than the other samples

and on a much less confident note and some (lots ) guessing I'd put zwqsf as the 192, xlkgw as the 320 and vtrzx as the original

I'll send you the solution via PM, I don't wanna spoil it for _xxx_.

 

[_xxx_]

I'm sorry but I don't believe you. Either it's pun or you didn't simply compress it, but added some fx or boosts or whatever.

By the way I viewed it in Sound Forge today totally zoomed in to a millisecond and it definitely does look like what I heard. Just the level of detail is worlds apart, I can post screenshots as well.

 

[_xxx_]

Now take a look at this 20ms sample and tell me what's the original:



That was Amon Amarth. Either your compression proggie does some additional processing or it's just bogus.

 

[L233]

So you can't do what you claimed you can, in fact, you failed spectacularly by mistaking the lowest quality MP3s for the highest quality samples, ergo the samples must be bunk? Let's face it, you can't really hear much of a difference between CD quality and well-encoded MP3s, at least not with Metal.

Yes, I did compress it. I ripped the tracks with EAC, cropped the samples with Nero WaveEdit and added the fadeIn/fadeOut, converted it to MP3 with EAC (using the latest stable LAME) and then used the decompress option in EAC to convert them back to PCM. I edited the resulting .WAVs by cutting factions of a second from the very end of the track so the file sizes are roughly equal. Then I removed the WAV tags that Nero WaveEdit annoyingly adds without asking. That's it.

I've uploaded the source files (i.e. the pcm and the 3 compressed files) for the Nevermore sample here:
QTest4Source.zip

You should be able to exactly replicate it by using the uncompress feature of the latest EAC (v0.99 preBeta 4).

Apparently, your SoundForge shenanigans didn't help since you don't know which is which to begin with.




Edit:

This is for the Amon Amarth sample:

pcm: yjkdh
320: iuzst
192: zxlpw
128: wtsnc

So on your screenshot, the 3rd one from the top is the uncompressed one. I know jack shit about sound stuff, so I don't know what I am supposed to see on your screenshot. Even if the compression program did some additional processing... how would it add the cymbals and "dynamics" and "harmonics" that are supposedly missing with lossy compression? I'll upload the Amon Amarth source files, too.

Edit 2:

Amon Amarth source files: http://www.file-upload.net/download-1202515/QTest1Source.zip.html


Edit 3:

Ok, I think I figured out, what's going on. I might have cut a few milliseconds from the _beginning_ of some of the WAVs in order to make them all the same file size. So the one you identified as the "original" (the 192 kbit one) is simply offset by a few milliseconds.

And I guess when you tried to cheat and looked at the files with Sound Forge, you noticed one of them looking quite "differently" from the others and you figured that it just _had_ to be the original because it's the only non-compressed one while those that look very similar to each other must have compression in common. BUSTED. Looking at the WAVs in Sound Forge mislead you to the extend that you posted an almost comically bad guess and then you try to convince me that Sound Forge "confirmed" what your ears supposedly heard (but probably never did). This is quite amusing.

 

[thehulk]

Another wannabe audiophile busted. lol

 

[N00b]

Now take a look at this 20ms sample and tell me what's the original:



That was Amon Amarth. Either your compression proggie does some additional processing or it's just bogus.

You forget that compressors have become a whole lot better in the last years. The latest lame is able to overcome most mp3 deficiencies (pre-echo, etc.) by clever encoding, so it simply produces very, very few artifacts anymore.
BTW, if you would know a bit more about audio, you would have seen that the last window shows a different sample position and not more "level of detail". It's really extremely obvious.

 

[_xxx_]

It makes no sense to compare anything if the compressed stuff sounds BETTER to a listener not knowing the original source than the actual original due to additional signal processing. That is not compression, but sort of remastering. Boosted stuff doesn't count at all, that's cheating.

And nope, I reinstalled SoundForge yesterday just to make sure because it bothered me.

What you can see in the screenshot is the amount of actual data in each sample, the samples with more compression have less information inside and the waveform shows that quite clearly, regardless of delay. So as it happens, your tools processed the signal and boosted some stuff - it simply doesn't resemble the original.

I'll take your source files and convert them PROPERLY in SoundForge, without any additional signal editing and upconverting or renaming and post them here for comparison.

 

[L233]

It makes no sense to compare anything if the compressed stuff sounds BETTER to a listener not knowing the original source than the actual original due to additional signal processing. That is not compression, but sort of remastering. Boosted stuff doesn't count at all, that's cheating.

Nothing was "boosted". This is getting quite sad right here. First you complain that it's the wrong material (Albuquerque's sample) and demand Metal. I go out of my way and provide you with what you want and then you complain that the production is "crappy" and Peter Tägtgren, one of the most sought-after Metal producers, is a "moron".

Of course, I anticipated the usual Golden Ears' shenanigans, so I posted not one, not two, but four different samples. Still, you cannot seem to hear all these differences that are supposed to be apparent in the compressed audio samples, so you accuse me of cheating by "additional signal processing".

Yet, I do wonder how "additional signal processing", or whatever I supposedly did to fool you, could possibly restore all the "missing cymbals" and the "missing harmonics" that supposedly get lost during compression. You know, the stuff that is so glaringly obvious to you that you claim to hear the difference on your car stereo.

Anyway, there is no "additional signal processing" involved.

And nope, I reinstalled SoundForge yesterday just to make sure because it bothered me.

Right, so you just happened to identify the second most heavily compressed one as the uncompressed original and it also just happened to be the one offset by a few milliseconds when looking at it in Sound Forge, confirming what you claim you "heard".

What you can see in the screenshot is the amount of actual data in each sample, the samples with more compression have less information inside and the waveform shows that quite clearly, regardless of delay. So as it happens, your tools processed the signal and boosted some stuff - it simply doesn't resemble the original.

No, again, what it shows is simply a different sample position. I'm not quite sure why I continue to waste time on this, since you're clearly not willing to concede that even your musician's Golden Ears cannot hear a clearly noticable or relevant difference between the uncompressed original and the compressed audio but here it is:



When you look at "zxlpw - Kopie", the sample is simply offset by a bit more than 00:00:00.025 (I actually _added_ a faction of a second to the beginning, I did that because the file sizes gave away the answer). As you can see, I was able to find exactly the correct position and the best thing: you can do it, too, and see for yourself.

So where is this "less information" that the waveform supposedly shows so clearly now, huh? Stop grasping at straws. The zxlpw sample is the 192 kbit VBR one and even that rather stongly compressed audio seems to be good enough for you to mistake it for uncompressed audio.

I'll take your source files and convert them PROPERLY in SoundForge, without any additional signal editing and upconverting or renaming and post them here for comparison.

There was no "upconverting" or anything like that involved and since I added a fraction of a second at the beginning of that particular WAV, it's unlikely that you can replicate the exact position. As I've said and shown, it merely starts a bit more than 00:00:00.25 later and you can look at it yourself.

 

[_xxx_]

Adding a fraction at the beginning IS a delay, what's wrong with that?

As for the rest, you can only hear the missing cymbals if they're actually being played. In case of Amon Amarth he merely plays stupid 4/4 full power with zero subtleties to begin with.

And whatever your opinion, it is simply impossible that the compressed audio produces more highs and more "air" than the uncompressed one without any additional processing. That's a simple fact, no need to even talk about it. That is clearly audiable and that's why I chose what I chose. Before I viewed it in SoundForge as mentioned, not that it has any relevance.

I'm not saying that you manually processed anything, but probably the software you used did it automatically. That is not a simple conversion, that is processing by definition.

And that producer may be as big as anything, he is either a moron or had a very bad day there. Sometimes everyone has a bad day, see latest Metallica mastering debacle with Rick Rubin who is otherwise a real guru. Obviously that has nothing to do with the creative quality of the music itself, I'm only talking about the sound.