I hate to get out of HA..., What would you tell this guy O_o Options

[ilikedirtthe2nd]

he, i just wanted to say: i really like this place here... audio discussion on other forums makes me headache

http://de.europe.creative.com/support/foru...&foru=12&page=1

regards; ilikedirt

This post has been edited by ilikedirtthe2nd: Jun 17 2003, 19:50

[Atlantis]

"I don't need to read anything and I am not wrong."

[ilikedirtthe2nd]

"I don't need to read anything and I am not wrong."

you see what i mean, he?

[ilikedirtthe2nd]

this is even better:

I've read the links, and I've read similar before.

The problem is that listening tests are subjective and the result varies on the pair of ears listening. A true scientific test (a waveform comparison), balanced with scientific information about frequency response in the ear is the only common sense way to test.

i don't think, he actually read the links (and understood what he saw...)

how can one alone be *that* ignorant

[ff123]

Nice reply to the thread on the other board, JohnV.

ff123

[Atlantis]

"I don't need to read anything and I am not wrong."

you see what i mean, he?

Yeah, the Dark Age of audio compression

[Lev]

HUGE thumbs up to JohnV for posting that... Thats the sort of thing I imagine most people could never be bothered to do.

Seriously, that amount of effort is very commendable.

[indybrett]

Nice reply to the thread on the other board, JohnV.

ff123

So good in fact that I'm going to task myself with memorizing (and learning) the entire text of it.

[userXYZ]

So good in fact that I'm going to task myself with memorizing (and learning) the entire text of it.

yepp, me too. Besides, nice avatar, indybrett. Did you make i yourself?

[ilikedirtthe2nd]

JohnV: Thanks for the post. I have read a lot on the topic before and do understand a lot of the background to this.

I still think that a waveform comparison, albeit skewed for frequency response and obviously taken over time, is a much more accurate test.

Of course MP3 is a lossy format. The critical thing is how much redundant data can you remove, but still create a waveform that is as similar as possible to the original (with frequency skew). Listening tests are interesting, but suffer from all sorts of issues statistically speaking.

omg

[nyarlathotep]

JohnV: Thanks for the post. I have read a lot on the topic before and do understand a lot of the background to this.

I still think that a waveform comparison, albeit skewed for frequency response and obviously taken over time, is a much more accurate test.

Of course MP3 is a lossy format. The critical thing is how much redundant data can you remove, but still create a waveform that is as similar as possible to the original (with frequency skew). Listening tests are interesting, but suffer from all sorts of issues statistically speaking.

There isn't much more we can do...
So sad

[indybrett]

Besides, nice avatar, indybrett. Did you make i yourself?

Nope. But I did fix it up a little.

[Pio2001]

Something we can do ? Post some sceen captures of waveforms and ask him which one looks best

[bluewer than blu...]

And I was wondering where's Roel been lately...now I know!

[ViPER1313]

Will someone please tell these people to use LAME v3.92 and test the following two command-lines against the original .wav, and then tell them to eat their words -

"-V0 -q9 -b128 -B128 -md -p --noshort --notemp --nores --strictly-enforce-ISO --ns-bass 2 --ns-alto 12 --ns-treble 9 -k"

and then....

"--alt-preset 128" (Both without quotes, of course - Just looking out for people on the creative forum)

The first command line LOOKS so much better in the spectral view . That's the one I should use, right? Oh wait - it sounds like crap! I wonder why? Could it be because there is NO lowpass? And no bit reservoir? And dual stereo? And other various "tweaks" ? Isn't it amazing that --alt-preset 128 sounds so much better, yet "looks" worse? Go figure.........

[AgentMil]

OMG!! If they were right that means all the test that have been performed in the world would be invalid then?

I hope posting that link on there doesn't bring those types of people here onto Hydrogen Audio. I think WE all know what sort of trouble that brought last time. I personally only come here to browse and sometimes participate due to heavy work and education commitments I don't the time to do all the nice interactive things that HA is known for!! But all the hard work of the new members and old members are very much appreciated by myself and most probably a lot of other people! I been using MusePack ever since I was introduced to the format via this forum. I also use all the other codecs on a need to use basis especially for car stereos only can playback MP3s.

Anyways I love the replys given by both JohnV and ff123 brilliant rebuttal cause all they can say now is scientifically valid results held in a controlled environment and nothing else to support their claims and their seemingly objective method of testing (I am pretty sure if someone had the time they can create a waveform that can look pictorially similar to a track but sound completely different). SSR can go back and support and help people with their Nomads IMHO, it seems that all he seems to have time for, if he was inclined reading those links would of at least make him learn a few things about audio coding.

Heres to more proper test held by people that we know and trust.

Laters
AgentMil

[2Bdecided]

I see http links to images appear in-line without you asking on that board. That makes for a dramatic looking post that I didn't intend!

Cheers,
David.


PS - of course, we all know that we're wasting our time over there - it's a creative message board after all! A triumph of marketing over reality - why let facts get in the way?

This post has been edited by 2Bdecided: Jun 19 2003, 10:33

[JohnV]

PS - of course, we all know that we're wasting our time over there - it's a creative message board after all!

Yeah, but next time we can copy-paste that text we wrote for the next "genius"..

[2Bdecided]

Yeah, but next time we can copy-paste that text we wrote for the next "genius".. 

I had chapter 2 of my thesis on-line as html for these arguments, but lost it when I left uni. Must sort my web page out one day...

D.

[Gabriel]

If I remember well, your thesis is on mp3-tech.org (but in pdf)

[2Bdecided]

If I remember well, your thesis is on mp3-tech.org (but in pdf)

Yes, and I sent him there. But it's a 19MB download - heck - I wouldn't bother with that, so I don't see why he should!

Cheers,
David.

[Pio2001]

Will someone please tell these people to use LAME v3.92 and test the following two command-lines against the original .wav, and then tell them to eat their words -

Why don't you do it yourself ? Anyway, I don't think it would convince him, because JohnV already talked about spectral views, and he answered that waveform comparison was better than spectrum comparison. We should find a sample that sounds better when the amount of noise resulting from the substraction of the original from the copy is bigger (substraction = differences between the waveforms).
I'm not interested in doing this, because if he wasn't convinced by JohnV's explanation, he won't be with another example, and will answer that it's just an isolate case chosen on purpose, and that for normal music, the less difference there is, the better the quality. And actually this may be true quite often.

I hope posting that link on there doesn't bring those types of people here onto Hydrogen Audio. I think WE all know what sort of trouble that brought last time.

I don't agree, I don't think we should be disappointed by anyone coming into HA. OK, fruitless discussions are a pain, but people with different beliefs should not be. It should not cause any problem if we answer with accurate and pertinent arguments, strictly technical, and not wander into the personal domain.
Well I know sometimes that there are people who are not interested in discussing (so called trolls), but I never assume someone is a troll without having tried to discuss first.

In this case, it is sad that these comments are done here "in their back" instead of there. But I let this rest, since I can't handle several forums at once, and JohnV and him have both clearly stated their positions, and i think JohnV's one is more convincing, thanks to the arguments provided.

[ViPER1313]

Why don't you do it yourself ? Anyway, I don't think it would convince him, because JohnV already talked about spectral views, and he answered that waveform comparison was better than spectrum comparison. We should find a sample that sounds better when the amount of noise resulting from the substraction of the original from the copy is bigger (substraction = differences between the waveforms).
I'm not interested in doing this, because if he wasn't convinced by JohnV's explanation, he won't be with another example, and will answer that it's just an isolate case chosen on purpose, and that for normal music, the less difference there is, the better the quality. And actually this may be true quite often.

I didn't post it myself because the forum itself is in German.... . I just thought that an easy to hear, real life example that proved the exact opposite of what people were saying might have helped. I guess I will have to give my 'Benutzername" and "Kennwort" and sign up to post . If people say it's an isolated case, oh well, it still flies in the face of what has been said.

[2Bdecided]

If you compare a recent codec (e.g. AAC) with an older one (e.g. mp2), you will get a general rule, rather than an isolated case. You'll find that the newer codec can be transparent whilst adding much more measurable noise than the older codec.

It's easy to find bitrates where the measured noise added by each codec is similar, but AAC sounds basically transparent, while mp2 sounds terrible.

This shows the absolute folly of relying on waveform comparison.


I'm posting here (behind his back) because I don't hink Creative would be appreciative of all our comments! More importantly, I've had a terrible internet problem this week where HA and Google are about the only sites I can reach most of the time. Most others come up with "The page cannot be displayed" so quickly that I'm sure my PC isn't even trying! It's getting so bad that I'm actually going to have to do something about it!

Cheers,
David.

[penvzila]

I actually learned a lot from that post.