HydrogenAudio

It's a new version of previous poll (http://www.hydrogenaudio.org/forums/index.php?showtopic=93660).

This poll has a separate options. It's "20 kHz" , "21 kHz"  and "22 kHz" options  instead of single "20 kHz and higher".

Also sampling rate of tone files is 48 kHz now.

Please perform the test at your normal loudness level that You commonly use to listen a music.

Also, make sure you have your software and OS configured so that it does not resample the audio from its native sample rate.  If your hardware has trouble playing files at 48 kHz, make sure you use a resampler that does not alias.  See this image as an example in Windows 7 and Vista (which apparently cause aliasing when set to 44.1kHz):
http://i.imgur.com/snX6u.png (http://i.imgur.com/snX6u.png)


Download the test samples here (6.74 MB):
Hearing test (http://www.hydrogenaudio.org/forums/index.php?act=attach&type=post&id=6966)

If you are unsure whether you can hear a particular frequency or not, ABX it against the silent track.  Be sure to play each sample in its entirety all the way through for each and every trial in order to avoid audible clicking that will give away the identity of the non-silent track.

This poll has two sections. You should check two options.

I  barely hear 17 khz, it's when I  compare with 18 khz that I  notice that it's not silent.

Also I  wish I would have been able to do an Abx  Test with higher frequencies and silence, but there's an obvious pop when you start to play, and you know immediately which file has high frequency content. Oh, I  tried to put a file with insane frequency ( like 40 khz) so that I  get a pop too  while being "silent" for normal ears , but then I  notice that the pop can sound different.

At a normal loudness level I can hear 13 kHz only (aged 62). I didn't read carefully about loudness level during the last test that's why I got at a 16 kHz there.

I'm 54 (well, nearly 55), and can just hear 16kHz at normal levels. I can hear 17kHz if I turn it up.

About 20 years ago I was easily able to hear 18kHz. I suppose losing 2kHz in 20 years isn't bad.

I was just listening to music at the usual (rather moderate) level with headphones on my iPhone, then without touching the volume I upload the files and ran the test: 15kHz (and I'm 43).

Actually, perhaps I think I perceived something at 16kHz but I'm not so sure I really heard it or just a placebo effect of the lasting click of precedent tone, so I voted conservatively.

You can even create pure sin tones in foobar2000 by typing "tone://16000,10" in add location dialog. Before listening set in Preferences>Advanced>Decoding your native soundcard sample rate (most cheap sound cards works at 48000 Hz)

About topic - sometimes I can hear 18 kHz (especially after sleeping) but not always. Always I can hear 17 kHz so I chose 17 kHz in poll.
BTW. It can be more interesting if you add age to poll... I'm 24 years old.

I'd like to vote (18-19 kHz) but it says that I already have voted.

By making them 48khz you only shifted the problems to those that had properly set Windows and their souncards to 44khz. I suggest providing another set of files in 44khz, so everybody can doublecheck.

I have only heard of sound cards resampling from 44 to 48. I have never heard of one that resamples 48 to 44.

14 kHz with this samples played with foobar2000
15 kHz (barely...) with foobar's tone://15000,10

Different levels?

Age 57 years.

I have only heard of sound cards resampling from 44 to 48. I have never heard of one that resamples 48 to 44.

It gets resampled to whatever you set in Windows in addition to what your soundcard does. http://i.imgur.com/snX6u.png (http://i.imgur.com/snX6u.png)

Ok, so change it there.

Has anyone perfomed tests on the particular resampler involved?  I'm betting it's transparent.

Ok, so change it there.

Has anyone perfomed tests on the particular resampler involved?  I'm betting it's transparent.

Using Audition 1.5 and Windows set to 44khz, I can clearly hear 19khz, sounding more like 16khz. If I set Windows to 48khz the sample is silent according to my hearing. Resampling the file to 44khz has the same effect.
If you set Audition's resampler to low quality it also produces lower frequencies along the original signal (visible in spectrum view), so I guess Windows is trading quality for performance too.

Setting Windows to 44khz should be prefered if you listen to regular music on said system, instead of working around that problem.

You can even create pure sin tones in foobar2000 by typing "tone://16000,10" in add location dialog. Before listening set in Preferences>Advanced>Decoding your native soundcard sample rate (most cheap sound cards works at 48000 Hz).

I like this method best; it works great and it allows one to easily setup even intermediate frequencies such as "tone://14500,10", so one can quickly jump between different half steps. Bravo!

I'm still a little confused by everything. A few question from a foobar2000 newb, trying to understand how to avoid SRC, if you don't mind:

A. How do I determine the native sampling rate of my (outboard USB) Behringer UCA-202?
B. When I set "Preferences>Advanced>Decoding" in foobar2000, as you advise, does this obviate what I have set in my Windows7/64bit, Playback devices, "Speakers 2- USB Audio CODEC" Properties>Advanced>Default Format ("when running in shared mode") of "16bit, 48000Hz"? Or do I need to (or should) change that too?
C. How should I set the two "Exclusive Mode" boxes, below this Win7 setting in question B, above, regarding "control" and "priorities"? I don't understand what they mean.

Thanks.

Using Audition 1.5 and Windows set to 44khz, I can clearly hear 19khz, sounding more like 16khz. If I set Windows to 48khz the sample is silent according to my hearing.

Ok, thanks for this.

Resampling the file to 44khz has the same effect.

For the sake of clarity, which effect is this, 19kHz is silent according to your hearing, or does it sound more like 16kHz?

Setting Windows to 44khz should be prefered if you listen to regular music on said system, instead of working around that problem.

This does not follow from the information you provided, which suggests that the resampler in Windows aliases.

Also, I don't believe this justifies creating 44.1k samples for the test.

Personally  I  just use SOX, to set at a sample rate supported by my soundcard. I  use 192 khz. If sox resampling is as transparent as it is claimed, then this shouldn't be a problem.

For the sake of clarity, which effect is this, 19kHz is silent according to your hearing, or does it sound more like 16kHz?

This does not follow from the information you provided, which suggests that the resampler in Windows aliases.

Also, I don't believe this justifies creating 44.1k samples for the test.

Proper resampling (quality set to 999 in Audition) makes it silent.

Windows only resamples if input and output samplingrate differ. So if your music is 44khz, and your soundcard can output 44khz and Windows is set to 44khz, no resampling will be applied.

[...] and your soundcard can output 44khz [...]

You left this part out earlier, hence my concern.

That still doesn't address pdq's point.  What sound card resamples 48 to 44?

If you change any of the 44khz in the above chain into another sampling rate, Windows will resample in order to match them again. Even if your soundcard can output 44 and 48khz properly, any regular music file with 44khz will be resampled to 48khz by Windows if you leave it at the default setting. Not implying that it's audible for music.
If you never want Windows to resample, you either have to use exclusive mode in your player or change the output rate anytime you play file with different samplingrate (for example DVD Audio with 48khz).

There probably are some older soundcards that can't playback 48khz either, so they would also need to apply resampling down to 44khz. Worst case scenario would be, a 44khz Audio CD, resampled by Windows to 48khz will then be resampled again by the soundcard because it only accepts 44khz maximum.

Since regular music is mostly 44khz, that's what I set my Windows to, otherwise useless resampling would be applied. My soundcard accepts any samplingrate.


Only if you have one of those infamous soundcards that only accept 48khz, you would leave the Windows setting at 48khz and then either live with the quality degradation or use a software resampler in your player.

I appreciate the helpful input, but recommendations about general playback is a bit off-topic.

Samples were created @48kHz in order to avoid the potential problems with the previous poll.  I wish they were faded in and out, hopefully to help avoid clicking, but that's a different matter.  If there were to be two sets we would invariably see people falsely claim the ability to hear higher frequencies, either because they chose the wrong set or chose the highest result from either set.

To my knowledge, all old soundcards that could play 44.1k but not 48k did not resample, they just refused to play 48k.

As such, I think providing samples only at 48k was the proper decision.

Well in my particular setup it did produce aliasing, so while I might be in the minority here, that could happen to other people too. I didn't meant to start a discussion about it as I thought my first comments were enough to get my point across.

You're absolutely right.  This information is invaluable in getting proper results.  I simply contend that verifying the sample rate setting in your OS and/or player is the proper solution, rather than creating two sets of samples.

Your post is worth repeating, and with IgorC's approval I can include something about it in the initial post:

It gets resampled to whatever you set in Windows in addition to what your soundcard does. http://i.imgur.com/snX6u.png (http://i.imgur.com/snX6u.png)

greynol, go ahead.

Done.  Feel free to PM me to make any additional changes.

17kHz right ear, 16kHz left ear (had some damage during military service)
age 42

FIIO E7 at vol 10, full scale windows
The "tone approach" in foobar sounds louder than the files sampled at 48kHz .

13kHz. I think I hear something at 14kHz but it's so faint I can't be sure. 53 years old.

This is more in line with my age than the 1rst test. Now I know I was hearing aliasing. Too bad! 

13kHz. I think I hear something at 14kHz but it's so faint I can't be sure. 53 years old.

This is more in line with my age than the 1rst test. Now I know I was hearing aliasing. Too bad! 

I  guess all senses are affected by aging  ...

15 kHz definitely, 16 kHz barely, 17+ kHz nothing .
Noisy PC, Windows XP, Realtek audio driver, foobar2000 with Kernel Streaming output .
Over 40 years old.

I posted a warning before but I'll post again. Is it really a good idea to play a 20Khz sine wave into an amp and speakers an then turn the volume up. The tweeters will be destroyed.

You're NOT supposed to turn the volume up.

Wow, am I the only one who can hear 22kHz ?

I`m 31 years old.

The 22kHz sample is very quiet, but i can still hear it.

It sounds similar to the tone that i hear if i'm on a very silent environment, its just a bit louder.
Maybe I have some kind of tinitus, but it's so damn low in volume that i only hear it in very silent places.

Now I am interested if other peoples hear high frequency tones too in places with very low noise. 

Did you perform an ABX test between the clip of silence and the clip of 22kHz?

16 khz 32 years old.

Did you perform an ABX test between the clip of silence and the clip of 22kHz?

ABX between silence.wav  and particular tone is useless.  xkHz.wav has a loud click at begining and in the end of file while silence has not.

when I get home I will make a new silence.wav with the same clicks and inbetween silence .

The click should go away if you your tone samples fade in and out.

I mentioned this in the original poll:

There are clicks when the samples start and stop which could probably be eliminated by creating ones that ramp up and down.

...as well as in this one:

I wish they were faded in and out, hopefully to help avoid clicking, but that's a different matter.

Hm, yes, the fading is a better solution.
If You have time to make a new files then I will be only glad as I can't in this moment. Thank You for previous add to OP.

I think we're getting closer to get  more  usable results than all previous polls like this one.

It's done.

I also changed the contents of the readme file to:

Note:
Please perform the test at your NORMAL listening level.  These samples are already FAR louder than they would occur in regular music.

Again, feel free to PM me for any additional changes.  While there may possibly be a few incorrect votes, I think we can continue rather than start over.  If enough people feel that the poll should start over again then that is easily handled.

Well, with latest files,  I  managed to ABX 18 khz from silence, which is better than the 17 khz I've initially selected on the poll.
However with 19 khz I give up, I'm not even trying to guess on the ABX  test, it's just the same.

The volume was set at a level , I  usually listen to music.

To be perfectly clear, all that changed in each sample was the length the fade in and the fade out.  The level and frequency remained unchanged.  Feel free to verify the new files against the old ones if you like.

yes,  the fade in/out is important ,  to avoid the obvious pop when starting to play , and to do a fair ABX  test.

Wow, am I the only one who can hear 22kHz ?

I`m 31 years old.

Can You perform ABX test with 22kHz file and silence.wav (updated files*)  at normal volume and with precaution  for sampling rate (see OP (http://www.hydrogenaudio.org/forums/index.php?s=&showtopic=93733&view=findpost&p=787815)) ?

* the only update is fade in and out of the tones to avoid the clicks.

Finished with the new files 

(http://sv1.imagefire.net/image/10450babe1e3.jpg)


(http://sv1.imagefire.net/image/8191242685b4.jpg)

foo_abx 1.3.4 report
foobar2000 v1.1.7
2012/03/02 17:17:29

File A: C:\Users\Martin\Desktop\Hearing_maximum_frequency_test\22kHz.wav
File B: C:\Users\Martin\Desktop\Hearing_maximum_frequency_test\Silence.wav

17:17:29 : Test started.
17:18:23 : 01/01  50.0%
17:18:52 : 02/02  25.0%
17:19:58 : 03/03  12.5%
17:20:49 : 04/04  6.3%
17:24:00 : 05/05  3.1%
17:24:50 : 06/06  1.6%
17:26:20 : 07/07  0.8%
17:26:47 : 08/08  0.4%
17:27:32 : 09/09  0.2%
17:28:34 : 10/10  0.1%
17:29:47 : 11/11  0.0%
17:49:47 : Test finished.

 ---------- 
Total: 11/11 (0.0%)

All DSP's were disabled and i has set my soundcard (Envy24HF) to 48kHz to avoid resampling.

Soundcard: Audiotrak Prodigy 7.1LT
Reciever: Denon AVR2105 (connected via SP/DIF, Volume at -34dB)
Speaker: Kef IQ7se

If you're running Windows Vista or 7, did you make sure that it was set at 48kHz, per the note in the initial post? Or is there something in fb2k to override it as there at least used to be with WinXP?

Also, I do have the ability to change the results of the poll, so if anyone wants to change their vote, send me a PM.

28 years old in a noisy environment (PC noise, some traffic) with open headphones:

X        = "Normal" music listening volume.
           Can add 6dB or more depending on mood and music type
           (that is "normal" is X+6dB in those cases).
X + 18dB = Not unusual that I listen to tracks that I like at this volume

X +  0dB: 18 kHz
X +  6dB: 19 kHz
X + 16dB: 20 kHz
X + 18dB: 21 kHz
X + 28dB: 22 kHz (10/10 ABX)

What should I vote? :-p

Part of the request not to boost the volume is to save your speakers/headphones, and most importantly, your hearing.  The levels in those files are already insanely high compared to normal music, perhaps 20dB louder if not much much more.

After exposure to grossly high levels of these samples, you might find that you can't hear them anymore.

I'll leave it up to others as to how you should vote.

The levels in those files are already insanely high compared to normal music, perhaps 20dB louder if not much much more.

I don't think so. They are about +5dB to -1dB louder than other tracks on my playlist. According to ReplayGain anyway, and I think that's about right.

<sigh>

It should read, "compared the levels of those frequencies in normal music."

Pure tones and music are two completely different things.  Have you tried a high pass filter 8 kHz on normal music and listened to the result? If so did you find the level to be comparable to these test samples?  Try filtering again at 12 kHz and then at 14 kHz.

Finally, ReplayGain is not an appropriate metric to gauge the loudness of high frequencies since the algorithm was designed to closely mimic the human ear which is more sensitive to mid-range frequencies than it is upper-treble frequencies.

I was hoping the initial warning would have appealed to common sense.

The test files have some artefacts due to improper fading; also the silence file is not dithered but the others are.

Here's some shell-script to create a set of artefact-free files:

for f in 0 $(seq 8 23); do
  sox -n -b16 ${f}kHz.wav synth sin ${f}k vol .5 fade h .25 5 .25
done

or just run sox manually.  0kHz.wav is dithered silence.

also the silence file is not dithered but the others are.

I  might miss something, but I  don't think silence needs a special encoding 

artefacts due to improper fading

Sorry, but I  thought that fading was a basic operation. I didn't know it would need special care  (at least regarding artefacts)

After exposure to grossly high levels of these samples, you might find that you can't hear them anymore.

Could you please approximately compare the "grossly high levels" to some sound in real life?
Maybe about as loud as in movie theatres, or less?

Not sure because at times I listen to music quite really loud. 
I'm afraid if I'm already deaf by now after tried listening to 17khz like 6-8 times and 18khz more like 10 times and some few more

18Khz btw, age 25. 
wasapi exclusive mode, no resampling.

I  might miss something, but I  don't think silence needs a special encoding 

It probably doesn't; however, to introduce any unnecessary uncertainty into an experiment is foolhardy.

Sorry, but I  thought that fading was a basic operation. I didn't know it would need special care  (at least regarding artefacts)

Science always requires extreme care. How 'basic' the subject of a mistake is, has no relation to the significance of its consequences (see, for example, here (http://news.bbc.co.uk/2/hi/science/nature/462264.stm)).

Designing an experiment around perceiving the presence or otherwise of audio, and unnecessarily introducing artefacts such as these is again, foolhardy:

(http://i42.tinypic.com/14xdphi.png)

I'm with bandpass here. Sorry if I'm too repetitive and educative, but we're on HA.

Why don't you guys add white background noise to the test samples? I.e. let the listener compare low-level noise and low-level noise plus a tone with fade-in and -out? That way both a) the artifacts bandpass showed and b) aliasing tones due to bad resampling at the listener side could be made irrelevant. In other words, we could have saved ourselves pages of speculation whether certain 22-kHz-hearing poll participants had configured their PCs properly because it wouldn't matter.

And by the way, one more argument for background noise is that it would also mask any low-level aliasing {or non-linear distortion} introduced by the human auditory system or the transducers in case of ultrasound stimuli (if such aliasing occurs; don't know if it does, but with background noise we're on the safe side). I think that's why some Japanese scientists in an article described adding background noise in their tests for audibility of up-to-28-kHz tones. I'll try to dig out the paper. {Here it is (http://www.jstage.jst.go.jp/article/ast/27/1/27_12/_article).}

So, in conclusion, I suggest you go ahead with bandpass's approach, but also add low-level white noise (preferably the exact same noise) to every sample, including the silence.

Chris

P.S.: Yes, I mentioned the background noise issue before (http://www.hydrogenaudio.org/forums/index.php?s=&showtopic=93660&view=findpost&p=787628).

More than fair enough. As for me it's not a problem to re-open a poll one more time. But previously  let's create the files and discuss possibe issues. Feel free to upload new files.  I would do it right now but I won't be home anytime soon.

How does this look?

for f in 0 $(seq 8 23); do
  sox -R -n -b16 ${f}kHz.flac synth sin ${f}k noise remix 1p-6,2p-46 fade h .25 5 .25
done

which:

• has 6dB of head-room
• puts masking white-noise 40dB below the test tones
• -R ensures that the noise is generated the same each time
• uses flac to save some space

Sounds good overall, but some remarks:

• I'd leave out the 23 kHz sample, since it's beyond 22050 Hz. I think it wasn't in the test anyway.
• Are the resulting files mono? If not: they should be so that we don't have stereo noise and mono tones.
• I think we should let the tones start and end a few msec into the file, in case a listener's player chops off some samples (which might lead to clicks again)
• To save space maybe it's better to use 8-bit WAV (without dither, the noise already dithers). FLAC is still not supported by every player.

Sorry, I tried to upload my own samples here, but it doesn't work at the moment.

Chris

I can hear up to 18 KHz. But with ABX test I can  identify  19-22KHz.  foobar Wasapi- Asus U1- Sennheiser HD 555. What is wrong with the setup or samples?

• I'd leave out the 23 kHz sample, since it's beyond 22050 Hz. I think it wasn't in the test anyway.
• Are the resulting files mono? If not: they should be so that we don't have stereo noise and mono tones.
• I think we should let the tones start and end a few msec into the file, in case a listener's player chops off some samples (which might lead to clicks again)
• To save space maybe it's better to use 8-bit WAV (without dither, the noise already dithers). FLAC is still not supported by every player.

Yes, they're mono.
Worried that 8bit might not have enough dynamic range, so sticking with 16-bit wav and addressing the  other points gives:

for f in 0 $(seq 8 22); do
  sox -R -n -b16 ${f}kHz.wav synth sin ${f}k noise remix 1p-6,2p-41 fade h .25 5 .25 pad .25 .25
done

I also have problems uploading samples, so please go ahead anyone who doesn't!

Designing an experiment around perceiving the presence or otherwise of audio, and unnecessarily introducing artefacts such as these is again, foolhardy:

(http://i42.tinypic.com/14xdphi.png)

Humm, I'm a bit surprised. According to the graph and colors, the artefacts are in the  -80 db  -100 db region... (blue-pink colors) .
So at  least this shouldn't affect a "normal use" of the fade, i.e in music.
Have you tried  to "analyse" the fade generated by sox, you don't see such artefacts  ?

I'll redo the hearing test later...

Why don't you guys add white background noise to the test samples? I.e. let the listener compare low-level noise and low-level noise plus a tone with fade-in and -out? That way both a) the artifacts bandpass showed and b) aliasing tones due to bad resampling at the listener side could be made irrelevant. In other words, we could have saved ourselves pages of speculation whether certain 22-kHz-hearing poll participants had configured their PCs properly because it wouldn't matter.

I'm curious. Why would the noise prevent aliasing of high frequency tones?

It wouldn't prevent it. It would mask the aliasing.

Chris

Seriously I  didn't hear any "artefacts" in the fade.  The fade in / fade out are relatively fast (250 ms) ,  you don't even have time to realize there's a fade in / fade out by the way.
I avoided aliasing by configuring soundcard output , or just using sox.
And you guys are over-complicating things.

I don't think my samples are going to cause any false positives, but If they can be replaced with better ones then that should be done.  Regarding the idea that noise will mask SRC artifacts, it better be a lot of noise, since some implementations can be really bad (older creative cards come to mind).  As such, I think 48k sampling rate should still be used and other caveats should still be included.

Anyway, feel free to upload new and improved samples to the upload section under a new topic and I'll handle the the rest. I will not be generating them myself.

Could you please approximately compare the "grossly high levels" to some sound in real life?

Sorry, I can't.  We are talking about pure tones at the edge of human hearing.  This is not typical of what you might encounter in everyday life, except for something like the tone coming off a CRT.

Wait, the poll says "Total Votes: 74 - You have already voted in this poll", where neither of this is true. I count only 56 votes.

Anyway, let's see how the last poll ended: total number of listeners: 122, listeners voting 18 kHz and above: 29 (26.6 %)

The results of the current poll so far are: total number of listeners: 56, listeners voting 18 kHz and above: 18 (32 %)

So we have been very successfull in eliminating hypothetical false positives  I'll let the others decide if it's worth opening another poll...

Chris

listeners voting 18 kHz and above: 18 (32 %)

So we have been very successfull in eliminating hypothetical false positives

...and I can arbitrarily cherry-pick data and be sarcastic about how we haven't eliminated hypothetical false positives.

Sorry, wasn't my intention to criticize your and Igor's work. If I create a group "19kHz and above" instead of "18kHz and above", that group percentage-wise hasn't decreased in the current poll either. So the previous poll seemed good enough - or at least the listeners educated enough. That's what I tried to get across.

Chris

Anyway there is still sharp drop from 18 kHz (10 votes) to a single(!) vote for 19 kHz.  And there are still more votes for 20 kHz than for 19 kHz as I've talked before (http://www.hydrogenaudio.org/forums/index.php?s=&showtopic=93660&view=findpost&p=787728)
Considering some young guys (20-26 years old. I know them from previous public tests) who have reported <19 kHz as max. frequency it's hard to beleive that some members actually hear a pure tones at 20 kHz and above (plus those members are somewhat older.  30-40 years old).  That's my personal conclusion with which somebody is free to be agee or not.

Anyway there is still sharp drop from 18 kHz (10 votes) to a single(!) vote for 19 kHz.  And there are still more votes for 20 kHz than for 19 kHz as I've talked before (http://www.hydrogenaudio.org/forums/index.php?s=&showtopic=93660&view=findpost&p=787728)

I can understand the strong drop from 18 to 19 kHz - drops above 16-17 kHz have been reported in the literature, and it might be smoothed a bit by more participants. But the "peak" at 20 kHz is strange indeed. But maybe that's simply coincidence. I'd guess the participants in both polls are (will be) roughly the same.

Congrats to 1000 posts

Chris

Congrats to 1000 posts

 
Thank You. I haven't noticed that. What a waste of moment.     

I think it's important that we do our best to eliminate doubt.

If the results don't change does that make the exercise a waste of time?

Well, I would try one more time again with the proposal of Chris.  I guess it's way to learn.

It wouldn't prevent it. It would mask the aliasing.

Hmm. Aliasing sounds like a lower frequency. So let's say 20khz turns into 18khz. How would noise do any good here? I can only see it masking anything by being much louder, so you can't hear any high freuquencies at all.

Wait, the poll says "Total Votes: 74 - You have already voted in this poll", where neither of this is true. I count only 56 votes.

I don't remind, but there might be a "null option" for people that only want to see results.

Hmm, I checked what happens during bad resampling from 48 to 44.1 kHz. See attachment. Tones from 18 to 22 kHz are from left/green to right/pink, respectively. The white arrow points to the strongest aliasing component, which belongs to the 20-kHz tone. This might explain why more people voted "20" than "19" or "21"?

But onkl and greynol, you're right, masking such strong aliasing requires very loud noise, with a level equivalent to dither noise when going to 4-bit WAV. Not very nice.
[attachment=6967:48_to_44...sampling.png]
Chris

Okay, I've uploaded some new samples taking into account the above comments. In fact, I opted for masking noise at -30dB (instead of the -40 I suggested earlier).

@onkl, aliasing effects, unless a playback resampler contains no filtering at all, should be sufficiently small to be masked by this level of white noise.

Cheers,
Rob

I think it is appropriate to add noise to the silent sample that it be identical to that used to dither the tones.  I think noise should also adequately mask the modulation introduced due to volume ramping as well as the small burst that occurs when the ramping up ends and the ramping down begins.  Perhaps this is what is happening in the samples just uploaded by bandpass.

EDIT: I assumed the noise would be LSB dither, this is overkill.  Regardless, if the samples were at 44.1k and I tried them on something like an SB Live! I have my doubts that would be enough to mask resampling artifacts introduced by the card.  I also am concerned about people claiming to hear 22kHz because they hear something, not unlike the person claiming to hear a tone because it was "rumbling."  Honestly, I think the previous samples were better, but am more than happy to see how these go over.

Well, I  found the background noise too loud for being able to ABX 18khz from silence (I  was able before) , but I  can still ABX  17 khz.
And I  had to use convert mono to stereo , in dsp chain otherwise I  have playback problems, and I  prefer to use both ears.

Just to clarify:

If I have found that 15k seemed louder than 14k, does that mean my test was invalid? (aliasing, or other problems?) [I'm on Windows XP and I couldn't find an option to fix the sample rate].

By the way, I could hear up 17k. [I'm 20 years old]. I'm not going to vote until I'm sure the test was properly realized.


Another thing: listening to the samples was pretty uncomfortable. I set foobar2000 volume control to -15dB when listening regular music with Replaygain. I had to set to -60dB during the test. Is it supposed to be THAT loud?

Honestly, I think the previous samples were better, but am more than happy to see how these go over.

With the new samples, the signal to noise is much better controlled than in the previous samples; specifically, they are:

1 less dependent on playback volume
2 less susceptible to playback aliasing
3 not subject to the spectral splurge (and other unexplained artefacts) caused by using a linear fade instead of half-cos-wave fade

The amount of masking noise (-30dB) is just sufficient to mask aliasing from the pulseaudio (linux) playback chain, but insufficient to mask aliasing from for example, cubic interpolation. I've uploaded another set of samples with the noise 20dB lower, so the first set I uploaded allows listening tests of relatively 'quiet' (i.e. small signal to noise) tones and the second set relatively loud (higher signal to noise) tones.  If you're sure that your playback chain does not resample 48k, or that it does but has half-decent resampling, then it should be okay to use the second set; otherwise, stick with the first.

The amount of masking noise (-30dB) is just sufficient to mask aliasing from the pulseaudio (linux) playback chain.

Your original upload from yesterday (-30dB, Hearing_maximum_frequency_test2.zip) also has sufficient noise to mask aliasing in Windows Vista and 7. Still, of course, we should keep the user instruction to configure the OS to avoid resampling.

Chris

With the new files I can easy hear 18 KHz but this time I can't ABX 19-22 Khz.

I'd say same result with both last tests, 17 khz is the last frequency I  manage to ABX  "easily".

If I have found that 15k seemed louder than 14k, does that mean my test was invalid?

Not really, such results are possible. The frequency responses of headphones are usually not flat and also depend on the shape of your head/ears and how you wear the headphones. But you should try again with bandpass's new set of files (http://www.hydrogenaudio.org/forums/index.php?showtopic=93824).

Chris

Quote from: C.R.Helmrich on 2012-03-03 17:06:47

Wait, the poll says "Total Votes: 74 - You have already voted in this poll", where neither of this is true. I count only 56 votes.

I don't remind, but there might be a "null option" for people that only want to see results.

That’s correct; see the button “View Results (Null Vote)”.

I can’t speculate about why it thought Chris had already voted if he hadn’t.

Now when there are new files are we all ok with the current poll or new one should be opened?

It will be worth if a members who have voted for >=20 kHz will perform the test with the latest files.
We also can go farther and create ABX sessions as for public tests (but probably will receive less results/votes).

I'd say just add the new files to the opening post and ask people that score 19khz and higher to doublecheck with these.

I'd say just add the new files to the opening post and ask people that score 19khz and higher to doublecheck with these.

The files have been updated a few days ago.


Important update:
New set of files was uploaded.  Now tones also have additional background noise to mask aliasing.
People shouldn't be shy  if  they can't hear theoretical 20 kHz.  It's absolutely normal.
It will be much more useful to inform real hearing abilities. Nobody will be accused of anything that's for sure. 
Maybe after all there are some individuals who can hear anything at 20 kHz and above.

I will ask  those of members who have voted for 20 kHz and higher (well, other as well) to run a test with a new set of files again if it's not much to ask.

Thank You.

Quote from: onkl on 2012-03-05 20:04:21

I'd say just add the new files to the opening post and ask people that score 19khz and higher to doublecheck with these.

The files have been updated a few days ago.


Important update:
New set of files was uploaded.  Now tones also have additional background noise to mask aliasing.
People shouldn't be shy  if  they can't hear theoretical 20 kHz.  It's absolutely normal.
It will be much more useful to inform real hearing abilities. Nobody will be accused of anything that's for sure. 
Maybe after all there are some individuals who can hear anything at 20 kHz and above.

I will ask  those of members who have voted for 20 kHz and higher (well, other as well) to run a test with a new set of files again if it's not much to ask.

Thank You.

Is It possible to change my vote?

When I read "AT NORMAL LOUDNESS LEVEL", I assumed the second part is for loud levels(read the last option carefully after voting) so I answered 22. With the new samples, I can ABX 17 but the best I can really easily hear is 16.

Is It possible to change my vote?

I don't know if administrators can see or change the votes.

I will open a new topic with ABX sessions (like ones for public test ) instead of poll.
The question is which version of noise we should adapt instead of providing two sets of files.

Also the readme.txt will be updated as the listener should start to perform the test at 8-12 kHz  at normal loudness level and then continue to increase the frequency without changing the volume.
Now it  will be more close to call it a valid experiment 

I'd suggest to say 'comfortable' level rather than 'normal' level since the test signals are not normal listening material.

If only one set of test files, I'd suggest my first set (...2.zip) since the risk of other effects is lower and testing at a particular signal to noise level is a perfectly valid and useful thing to do, so even if future tests are done at other levels the data gathered will still be useful as part of an overall picture.

Cheers,
Rob

I  managed to abx 18khz with difficulty at the end  ( Only 5 try on abx test, as this annoy me, result  3%)
Also I m using now  a different pc, with onboard soundcard instead of xonar stx.
I  thought that my realtek soundcard might add distortions that helped in the abx test.

Apparently, as reported in a separate thread, Microsoft offers a hotfix to improve its resampler in Windows 7 and later (http://www.hydrogenaudio.org/forums/index.php?s=&showtopic=86676&view=findpost&p=788882).
Edit: I tried it, and it works for me (no more audible aliasing). I wish they would have offered this hotfix before the poll(s) started

Chris

Apparently, as reported in a separate thread, Microsoft offers a hotfix to improve its resampler in Windows 7 and later (http://www.hydrogenaudio.org/forums/index.php?s=&showtopic=86676&view=findpost&p=788882).
Edit: I tried it, and it works for me (no more audible aliasing). I wish they would have offered this hotfix before the poll(s) started 

Chris

Thanks for notifying about that, installed it for my machine now.