difference of 1 dB...

[DrWho]

Ok guys, it's time to bust out them headphones for a short listening test.

Attached is a zip file containing two wave files, both are a 500Hz tone recorded at 44.1kHz 16bit (basic red book cd quality). Both are identical except that one of the tones has a part in the middle where the volume drops to -1dB for awhile and then it comes back up to 0dB sometime before it stops playing.

I have posted a really tall picture so feel free to post your opinions of which file is which (A or and then later on down the road I'll tell y'all the real answers. Just make sure you listen to the two tones before seeing what other people have written. I'm curious how many people might get it wrong...to be honest, I had a hard time telling which is which even when I knew exactly when it'd happen...I think the transition between -1dB and 0dB is more noticeable, but that's just me.

speakertone.zip

[colterphoto1]

I think B has the drop. After 8 listens.

Michael

[mr_ayo]

B, the drop is more noticeable to me...

[johnyholiday]

i have tuned so many guitars using harmonics, an set the bridge intonation of the strings ,that i must withhold my vote ,i suffer from true pitch ear, GREAT TEST

[colterphoto1]

Johny, what does 'pitch ear' mean?

Michael

[Born2RockU]

I'll say "B".

It's quite obvious on my Mackie studio mixing monitors.

[LarryC]

It was hard for me to hear a drop, but I thought B was ever so slightly louder at the outset, maybe a db?

[jacksonbart]

It sounds like B drops then comes back up, while A remains constant. I did not have headphones handy so I used by my 'awesome' cambridge sound work cubes attached to my PC. Perhaps that voids my test?

[dodger]

Webtv only downloads one tone, mos likely thaat tone that is downloaded is a. That stays constant.

B has the change if the above is true.

Can you do 2 downloads - test ability and memory. Call # 1 A, # 2 B

Thanks,

Win

[sfogg]

Guys,

Better way to do this test.

Download the software from here....

http://pcabx.com/

Then load the above two files into it.

Then using the above program run a test and see if you can identify if 'X' is A or B and do this at least 10 times to get a good enough sample. The program will keep score for you and tell you the percentages at the end. To 'pass' the test you need to answer enough questions correct that the possibility of that happening by random is below 5%. For 10 trials that means about 8 correct answers. If you score less then that you have failed to identify X a signifigant number of times. That does not mean there is no difference, just that the person under testing could not hear it that time. (Can't 'prove' a negative, only a positive)

You can do this with all sorts of files. If you want another SPL level difference test make two files of pink noise and lower the SPL of one of them by however amount you want to test. The limits of human hearing on pink noise is around 1/3 of a dB... see how many can tell them apart using the above test.

Shawn

[colterphoto1]

ah, good a statistician in the group. I have always known that this was the problem with many of the A/B tests done for magazines, etc. Also to be valid Mike, your test needs to have 30 or more participants, preferably randomly selected, not self-selected.

In other words , to be able to say if the average human can tell 1 db differnece, you should randomly sample from a population 30 or more persons, then do te 10 trial double/blind test that Sfogg has suggested.

We are 'trained ears' or at least most of would like to think so, and we select our selves to participate or not. ALso those with Rat shack meters can cheat. (ok, I'll admit, I checked after my test and before I posted- didn't want to be 'wrong') But it's a fun exercise nonetheless.

I assign to you to set this up and randomly select some of your classmates (there will be an age bias here, but at least not the dreaded 'stereophile' bias) and chart the results. Same circumstances, same headphones, etc. Have em round to your place. (good way to get the girls over- 'say, I'm running this little experiment, would you like to participate?')

Funny, I was always told that humans needed +/- 3 db (double power) to tell any appreciable difference in loudness. Looks like you're on your way to disproving that theory. Good man! Nice thread Mike

Michael

[Trey Cannon]

B drops out and comes back.

I used the Promedia 2.400 system at my desk at work.

Easy to hear.

[johnyholiday]

true pitch ear means i can listen to a song, an tell you what key it is in, a note an tell you what note it is,a scale an tell you what mode it is ,ie that song is in the key of C, using an ionian scale , with the melody ,or solo tonel center(root note),also know as ear training ,yes you can train your ears, it's even a class were you repeatly pound the same note on a piano, 3 units i think ,or it's a god given gift also. pitch;the first seven letters of the alphabet are used to repersent the pitch of notes that comprise the western worlds music system these are C-D-E-F-G-A-B there are historicaly reasons why the musical alphabet starts with C instead of A ,non english speaking countries these are DO-RE-MI-FA-SO-LA-TI to name the pitch of each note,playing all the white an black keys of a piano in one octive is a chromatic scale the basis of western music, if i ask you what key is this in an you say key of A ,pentonic scale, i know what the chord progression is ,or visa versa ,an the key tells me the tonal center, it's a closed loop ,a few scale modes are, mixolydian mode(scale), one of the most common scales in blues an jazz,or locrian mode(scale), least used in western music, but forms an importanted part of japanese an hindu music, so as you can see your ear is trained somewhat already, because jazz or blues might seem pleasant to you, where japanese or hindu will make you head for the exit . lets jam,

[ygmn]

But would not the playback source determine if the SPL is 1dB.....

i mean my crappy computer sound system...may not be able to take the source you have and reproduce the actual 1dB drop....??

some food for thought....

maybe get a SPL meter to check peeps who have taken the test....

[sfogg]

"But would not the playback source determine if the SPL is 1dB....."

Nope.

The playback will determine if the two play at 88 and 89 dB or at 95 and 96 dB or whatever but there will still be 1dB of difference between them no matter where you have your overall playback level set.

Shawn

[dodger]

----------------

On 3/3/2005 11:14:17 AM sfogg wrote:

"But would not the playback source determine if the SPL is 1dB....."

Nope.

The playback will determine if the two play at 88 and 89 dB or at 95 and 96 dB or whatever but there will still be 1dB of difference between them no matter where you have your overall playback level set.

Shawn

----------------

I

I consider source as a PC, Receiver, Pre-amp, amp etc..

Some may note that speakers have a + - db range freqency difference, That same rule that Shawn noted would apply. The drop should be noticable if one can hear it as the plus minus through the frequency range would put the same frequency at the same spot on a plus/minus chart.

[auto_xer]

B at ~3 seconds it drops and comes back at ~6 seconds.

[minn_male42]

definitely "b"

[tpg]

Ya... B. I thought it was pretty noticeable.

[colterphoto1]

My el cheapo Labtec computer speakers played back 'B' with a precise 1 db drop and regain. The system shouldn't matter unless the media (um) has any compression to it.

So this goes contra to the belief that most folks can only hear differences of 3 dB, or do we have better ears (and brains) than the general public hmmmmmmm...

Michael