ffc in #project_mayhem provided the following samples which he says are affected in stereo image using --alt-preset extreme:

c1.wav and c2.wav (http://ff123.net/samples.html)

problem areas are in c1, first 2 seconds, and in c2, 11 through 18 seconds.

description of problem
the end of at least one tone instead of being portrayed from the right speaker where it originated is protruding from the center with the ap file; as a result, some of the "sting" is lost

ffc performed 7 ABX trials and identified 6 correctly (94% confidence that he actually heard a difference).

ff123

Edit: ffc's preferred command line, which he says sounds better than --alt-preset extreme on these samples is:

-q2 -V0 --athlower 1 --lowpass 20 -b160 -p

Also, his listening tests were performed with speakers, shown here:

http://www.axiomaudio.com/products.cfm?productID=M60

I don't notice anything wrong with these samples using --alt-preset extreme, using my headphones.

ff123

I don't know it exactly, but I could imagine, that listening to music via headphone and via big speakers is a difference regarding stereo image.

I think that with headphones your image is not so wide.

If your speakers are properly set up, eg. 4 to 7 meters distance, listener is sitting at best place of stereo-triangulum, then image will be wider ?

Well, it was the first time since long i burned CD to listen at my speakers
With ap extreme decoded with Lame the first 2 seconds of c1.wav don´t show a difference.
I can´t imagine these axiom are offering something special that only them can reveal.

When you like to listen a real artifact, try sophia.pac via speakers, her voice suffering is really easy to hear.

Wombat

Nonsense IMHO. Headphones give _full_ stereo seperation, something that speakers never can give.

--
GCP

I can hear this loss of sting too and ABX the clip on it, however, this is simple preecho, and has nothing to do with stereo imaging.

[Edit: Some joker messed with my headphones. I'm retesting the stereo image]

[Edit2: Result unchaged. As far as I can tell, there is nothing wrong with the APS stereo image]

--
GCP

Is this some kind of sick joke? The above line gives worse results than aps at a 25kbps bitrate increase.

The preecho gets a lot worse and is now very easy to notice.

--
GCP

Unfortunately it seems very likely the guy who provided the samples is a troll using nicks like "ffc", "pchore" etc.
IPs seemed to be very similar.

Heh, not going to quote the irc discussions from today, but obviously he had something very much against Dibrom...
Obviously believing Dibrom has tricked everybody to hear funny listening test results..

I'm not defending ffc's command line (because I think it's inferior based on what I've heard so far), but a two speaker setup does have a couple of things over headphones in the area of imaging. Headphones are poor at giving a realistic impression of whether a sound is coming from in front or behind you. Same thing with height. Even with binaural recordings, which are made especially for headphone listening, depth and height are not rendered well.

The sort of imaging which speakers are better than headphones at was especially dramatic on some material (mostly minimally-miked acoustic music) when I used to use an inter-aural canceller, like Carver's Sonic Hologram. I could get an impression of sounds with realistic width, depth, height and size (the first which exceeded the boundaries normally set by the speakers).

I didn't normally turn on the Carver, though, because most music I regularly listened to sounded absolutely horrid with it on (too echoey and with added hiss from the unit). The problem was that recordings are typically made with normal two-speaker playback in mind.

ff123

I am far from convinced. I'd like to see an example where a recording that wasn't specifically made for a fixed 2 speaker setup has worse imaging on headphones than on speakers.

I agree the imaging is different, since with speakers you have room reflections, which headphones will lack. However, the posts here were about the *wideness* of the image. I don't buy that.

I know stuff like you describe exists. I think I even have a CD like that here. But they are special cases. i.e. if your speaker setup is not as described on the CD case, the imagining will be off.

--
GCP

I'd have to concur, having been to the channel earlier today and having talked to him directly.. heh. He really seemed to think I was out to ruin LAME somehow..



Yeah... this particular person seemed to have quite a few misconceptions, one being that vbr-new (the actual non-mtrh old version) was the highest quality vbr mode of all (it was scrapped in 3.90 because of being much lower quality than mtrh and vbr-old) and that 3.89 actually gives better results than 3.91!

It's really a pity these type of people's claims are even given this much credence, especially when it's fairly obvious from the start what their intent is. Anyway, this particular situation really shows why community verification of claims is very necessary and that just someone stating they scored such and such on an abx test isn't enough.

@Garf

Well about imaging whith headphones, the best headphones i listened on
with really good recordings leave me very unimpressed and bored.

You can´t tell me you can show even nearly the same feeling with headphones
than with a good Hifi. That feeling comes mostly from the imaging in front of you
and lets call it the "Good Vibrations" that your body feels directly from the speakers.

You don´t need extraordinary recordings for that.

I only can advise people to go to a shop where real upper class speakers are presented under
good conditions and listen. Me, i can´t discern the depth of a room with the first beats
of a recording, even with my Sennheiser 590s while on my speakers i can immediately and easy.

Maybe i am to dump to listen to headphones or haven´t got enough imaging brain parts, but
this feeling of a recording is directly according to the stereo image and is very different
to hear on speakers for me.

Maybe when i were used to listen much with headphones in the past i would listen different.


Wombat

You are talking about a different thing here, i.e. the feeling of the sound. There are two big differences between headphones and speakers there:

a) speakers will generate additional room reflections

b) a good (sub)woofer will additionally be able to transmit vibrations to your skin, which a headphone can't

c) a headphone has better stereo seperation (note that this doesn't necessarily improve the feel of the music)

All of these will give an improved feeling of music to a listener when he uses speakers. However, I repeat, that wasn't my point. The statement was that using headphones will make you unable to detecting certain stereo imaging problems. I don't believe that.

I don't contest that they generate a different feeling of space; that should be obvious from the above three differences. Since the one from speakers is closer to what you are naturally used to, you'll have less problems linking this to a certain dimension in your brains. For headphones, it will be different and perhaps not so natural, but it will just as much be there.

--
GCP

the biggest differences between headphones and speaker, exept the separation that is full for headphones and low for speakers, especially at low frequencies (we hear both speaker with both ears), is that headphones move with our head, while speakers stay at the same place when we move our head.
Our brain is very good at analyzing dynamic information coming from the head's movements. With headphones, it realizes that the source of the sound is moving when the head is moving, and therefore concludes that the sound is coming from an object put on our skull, and not from something in the room.
Another difference is that the wavefront comes parallel to the ear, and, for high frequencies (~10000 Hz complex noise, impulses...) the brain imediately knows that the source is to the left or to the right, and not in front, because it can analyze the holes in the spectrum that are characteristics of a plane wave reflecting into the ear when it comes from a given angle.
The same with speakers put completely to the left and to the right of the head : medium and low frequencies, as well as pure high frequencies (sines) are OK, but you'll never hear complex high frequencies or transients coming from in front of you, even if they are played in mono. The brain can't be fooled, it perfectly localize the two tweeters to the left and to the right.

Notice that it completely depends on the signal used for the test.

A funny game is to put a speaker in front of a friend and another behind him, and bet that he won't be able to tell from which speaker the sound is coming.
Just use a 1000 Hz pure sine, but beware to playback the sine in background, and switch front/rear speakers with the volume muted.
Then quickly rise the volume with an analog volume control, for a very short time. He won't be able to tell from which speaker the sound came ! Or he will mistake one time out of two.

Don't play the sine too long, because if he turns his head a little while the sound is playing, he will be able to tell, also, don't use digital volume controls, nor on/off switches to play the sines, because the noise they make changing the level step by step (i.e. the "click" it makes) has a spectrum wide enough for the brain to localize the good speaker Just turn the volume control of the ampli.

Edit : explained better

Very interesting post, thanks!

--
GCP