Test your soundcard for clipping, with this sample Options

[grbmusic]

did you read the warnings posted by garf and pio?

Yes but sadly after to fried my tweeter, if you take a look, I was the first in made the test and repply in this thread.

[LoKi128]

Well, tested with Winamp and my shitty on-board soundcard (Yamaha DS1)... got tons of weirdness at 44.1kHz... sounds fine when I resample to 48kHz with out_ds_ssrc.

So... does that mean that my soundcard has a shitty upsampler? Is it windows (i run 2k)?

In any case, resampling makes my CPU usage go from 1% to a pretty steady 30%. I have a Celeron 400. So I won't be running the resampler all the time... Still good to have a test sample that actually SHOWS people the effects of clipping and or resampling.

[KikeG]

I tested again at home with my Audiophile, and heard same thing as with the motherboard sound card I used previously: A ultra-high pitched tone aside from the telephone dialing sound. It seems that the mb card I used before (SoundMax, in a Compaq computer) is not that bad.

The strange thing is that doing some more additional testing at home, I could hear (or better say "feel" inside my head) tones up to 20.5 KHz or so, if played at enough loud level. I could swear that in previous tests I could not hear anything past 18.5 KHz. Maybe it's because of my strong hay fever causing congestion and ear side effects, or maybe some of the older equipment I used at my previous tests was rolling off these frequencies.

Again, if anyone wants to try this, I suggest that you use headphones, they are much less prone to fry, and even less if they are high impedance headphones. Also, use short, fade-in-fade-out, test tones.

...said that, yesterday at home, when playing these tones at loud levels, something started smelling burnt, and I saw some white smoke near the amp., I quickly switched it off. I think it was someting in the amp that overheated (or burned?), but I opened it and saw nothing wrong, and the amp seems to work fine. I have to check it more carefully.

This post has been edited by KikeG: May 28 2003, 13:33

[Pio2001]

THIS is a killer sample

I could swear that in previous tests I could not hear anything past 18.5 KHz.

Nor me above 16 kHz.
IMHO, it's just the playback level that is insane. The dial tones are so weak that we set the volume higher than we should.

Maybe it would be wise to record a 4000 Hz full scale sine just before, to prevent people from playing it at full volume

[lucpes]

M-Audio Delta 410 card... dial tone + very high pitched noise... the udial sample was the best way so far to test the fuses that protect the tweeters in my Infinity RSII speakers (have 2 tweets/spk and would cost around USD 250-300 to replace all, IF I can find them). Fuses got burnt, everything's ok, besides my ears...

[2Bdecided]

In fact, there isn't a hard cut off limit in human hearing. What there is, is a steep rise in threshold. Below 12kHz, it's typically ~0dB SPL. Above 20kHz, it's typically ~100dB SPL. There's usually a steep rise from (approx) 20dB to 80 dB around (approx) 16-18kHz. This varies greatly with age and individual.

It's usually somewhere along that steep rise that we conclude we cannot hear any more. But, in fact, it's just the equipment that doesn't produce the sound loud enough. We can hear at that (high) frequency, but with a drastically reduced sensitivity.

I found myself hearing 18kHz the other day, after being convinced that I couldn't hear above 16kHz. As Pio says, it's the loudness that counts. Listen to a 1kHz tone at the same loudness, and then decide if it's useful hearing!

Cheers,
David.

P.S. Do a web search for ultra sonic hearing and you might find some info on hearing up to 50kHz and beyond by bone conduction. There's even a technique for making deaf people hear, by modulating speech at ~50kHz (IIRC), and touching the transducer to their skull: this works for anyone, and is perceived as "normal" pitch speech. It still works in people with some types of deafness, so can be very useful.

[Moitah]

When using Winamp's DirectSound plugin (v2.2.6, included with Winamp 2.90), I hear the weird 'alien' sounds. This is effected by changing the sample rate conversion quality in Control Panel, so this means Windows is resampling? When I use the Wave Out plugin it sounds fine. Also, the DirectSound output in foobar2000 sounds fine (as well as playing it thru Windows Media Player 9, which I assume uses DirectSound). Why is this happening with the Winamp DirectSound plugin?

(Turtle Beach Santa Cruz w/ latest beta drivers, Windows 2000 SP3)

EDIT: Found out from reading Halcyon's post that it's the "Allow hardware acceleration" checkbox in the plugin. To allow the card to do the resampling, this has to be checked, and the "Hardware acceleration" slider in Control Panel must be on at least the 3rd notch (Standard or Full acceleration).

This post has been edited by Moitah: May 29 2003, 19:45

[JeanLuc]

I tried the same sample with my SBlive 5.1; The best setting (not surprisingly) was with SSRC at 48000Hz but even at this setting I could still hear some weird ambulance in the background (not like the super annoying ultrasonic I've experienced with SoundMAX) but weird all the same; With both cards I can't make it to sound like tigre's mpc.

The same here with my SB 512 PCI and with my onboard (SIS 710) sounds worst.
Still I'm crying for my speakers . Thanks of God I can get brand new tweeters (JBL Titanium) in my country but it's expensive yet, with a bit of luck I will can buy them the next month, only 2 of 5 weren fried

Sorry for your speakers, man ... if I were you, I would mount the two remaining tweeters into the front main system or re-arrange your speakers (if front & rear are the same types) as a start ... that way you can still enjoy two-channel music ...

[jues]

Soundcard: Delta 410.

Christ that hurt - yeah, put something nice and loud at the start so that people don't crank this too high - at first I couldn't hear the sine wave, but I sure did when I reached for those faders.

I feel sorry for everyone who has blown their tweeters on this test.... :|

[DickD]

I have a Win NT4.0 machine with a 'Pro16V-A Pnp' soundcard at work which is rather limited. I already worked out that it can only work at 44.1, 32, 22.05, 16, 11.025, 8 kSa/s sampling rates.

With Microsoft Sound Mapper, sampling rates like 24 kHz sound like a 45 record played at 33 rpm as all the sound comes out slower. It purports to work at 48 kHz, but it's clearly downsampling to 44.1 kHz. "How well?" is the question.

Just to check how bad Microsoft Sound Mapper is, using FB2K I resampled to 48000 S/s, (even in Slow mode, 64-bit), and set the playback waveOut device to MSM. The sound broke up terribly after the first three tones, with a dirty sounding loud siren sound. This shows that there's practically no filtering going on in the downsampling, so as I suspected, I'm much better off using FB2K's Resampler to play 48 kHz content.

Listening on a plain WAV player or FB2K at 44.1kHz to the waveOut/'Audio Playback' device that doesn't support 48000 S/s, I hear the dialing tones plus a reasonably quiet and pure high pitched frequency-modulated varying tone over the four longer tone-dial tones. The frequency-varying tone sounds like an american cop show siren on helium, unrelated to the tone dialing frequencies. Musically, I guess it might have been about 9 semitones of variation from peak to trough, or a frequency ratio of 1.7.

Looking at the spectrogram in a WAV editor, it's 19-21 kHz frequency modulation and very intense, so I shouldn't be able to hear it all with the volumes my soundcard can reach, if any at all. So I guess my cheap soundcard isn't properly filtered and I'm getting beating between the 19 kHz and its image above the Nyquist limit, at 25.1 kHz, coming out at a 6.1 kHz difference frequency, with the 21kHz part of the sweep beating with a 23.1 kHz tone, causing a 2.1 kHz beat frequency.

Hmm, that doesn't sound quite right as it's more like 18.5 semitones - a ratio of 2.9 to 1. But alternating Add Location... tone://6100,0.3 and tone://6200,0.3 (both with volume set to about -45 dBFS by editing replaygain values) sounds plausibly about right for the extremes of the tone sweep at similar loudness, so I suspect that the brickwall filter is imperfect, but perhaps it's good enough to reduce aliased frequencies in the 23.1 to 25.1 kHz range to below -45 dB or so (very roughly).

I'm glad I don't have clipping in my signal chain, and I'd say for real music -45 dB intermodulation with aliases isn't going to be an audible problem (certainly not with the noise floor of this sound card and the noise of the computer fans), so I'm pretty darned happy with it.

Using the Equalizer to apply -20 dB to the 14 and 20 kHz bands, is enough to make it inaudible, even though I'm pretty convinced it's well below 10 kHz, so this backs up the idea.

By the way, to hear the effect of clipping on this sample in FB2K turn your Windows Volume down low and remove Advanced Limiter (or Soft Clipping Limiter if you use it) from your DSP list, then turn on ReplayGain but turn off clipping prevention. Alternatively, just listen to the musepack samples already posted (with clipping prevention turned on - perhaps those samples should have been manually edited in FB2K to set ReplayGain values to 0.00 dB).

Incidentally, a possible safety idea when posting potentially equipment-damaging files (e.g. sounds too quiet so you turn it up and fry your tweeters with the ultrasound) is to post the files in a password-protected zip file with a password such as 'warned' so the user has to acknowledge they've been warned before playing the sound and can't just play it with FB2K's archive reader either.

People should also be careful that perceived loudness is reasonably high before hiding messages such as text in the spectrogram's 18-22 kHz region, as some people were discussing after the Aphex Twin pictures in music thread. (This will also protect people's pets from hearing damage to a degree)

Do many speakers or headphones come with a fuse, circuit breaker or limiter to protect them from currents in excess of their design rating? (I've seen that lucpes' Infinity RSII speakers do).

Incidentally, the same effect of gradual reduction of sensitivity is true of vision. Red light is normally reckoned to be no more than about 670 nm, I can see a 790 nm infraredlaser (I used safety precautions and did this in a laser lab), but only if it's focussed to a small very intense spot (major dazzle on an IR viewer, way brighter than sunlight would be) will it be visible as a dull red glow on a piece of card in a reasonably dark lab. Also, the reason weak red light plus weak blue on a computer looks like indigo/violet is that the red cone response versus wavelength has a second peak, but at lower sensitivity beyond the green and towards the tail of the blue at short wavelengths.

[Pio2001]

This is a picture of the aliases that can appear when the sample is resampled from 44.1 to 48 kHz. Here with SoundForge 4.5, quality 1, no antialias.

[fragtal]

I just downloaded the correct.flac and I'm really quite shocked:

After the first three or four dialing beeps there is horrible high-frequency noise/sound coming up that, according to correct.flac, shouldn't appear. Some of you (Creative-haters) will feel confirmed when I say that I've got an Audigy2. Reducing the master-volume with the creative mixer doesn't change anything. Also changing from ds ssrc 2.2.6 to waveout doesn't affect the quality (in this case I shouldn't speak of quality ) in any way.

I guess I should download actual drivers... Mines are datet 5.2.2003

and grbmusic, sorry for your speakers. That's something that really deeply hurts

[ViPER1313]

Sounds fine to me on my SB-Live value w/ 48khz SSRC resampling and master volume set to 50% in windows mixer (all other volume sliders set to max.) Yes, this is the same effect that the maximized sweep.wav file tests for (although the sweep is more through and wont hurt your ears quite as much ) Peace.

P.S. - Equalizations done in the windows mixer can have drastic effects on this sample. The further I turn up my treble, the further I have to turn down my volume.... etc....

[atici]

M-Audio Delta 410 card... dial tone + very high pitched noise... the udial sample was the best way so far to test the fuses that protect the tweeters in my Infinity RSII speakers (have 2 tweets/spk and would cost around USD 250-300 to replace all, IF I can find them). Fuses got burnt, everything's ok, besides my ears... 

I realized the very high frequency sound in the udial but neither my fuses burnt nor the tweeters in my Infinity Alpha 40s. How do you do a damage check? I tried test tones up to 20k and it sounds as it used to. I hope it's alright. lucpes, do my speakers have any fuses too? How come they don't put a safety fuse to any speaker more expensive than $300? My amplifier AudioSource Amp Two has a peak limiter which I keep on. It says in the manual : "The peak limiter modulates high frequencies at high volume, protecting your speakers from damage or distortion. A red LED also indicates activation of this feature.".I don't know what saved them but I'm really glad.

[Pio2001]

the sweep is more through and wont hurt your ears quite as much

Do you mean that this sample hurts ? In this case it's not played properly. You should hear 7 quiet dialup tones, nothing else.

[kritip]

the sweep is more through and wont hurt your ears quite as much

Do you mean that this sample hurts ? In this case it's not played properly. You should hear 7 quiet dialup tones, nothing else.

I really couldn't figure why it sounded so bad on my PC, i finally figured out that if i dropped the attenuation in Foobar2000 to about -18dB it finally sounded ok.
I thought this amount of attenuation was far too much, then i noticed i had replaygain enabled and the track gain for the file was abou +18dB. I disable replaygain, and when resampled to 48000Hz and now all i hear is a quiet (in relation to the dial tone) high pitched fluctuating noise!

Also, it din't make my eardrums feel like they had been scewered with cocktail sticks!!!

I didn't notice anyone point this out earlier in the thread, so i thought i'd let everyone know!

Kristian

[Stuv]

Quite interesting sample this one.
All ASIO, Wave and DS sound normal on my Aureon/Sky (master & wave maxed).
(my SB Live put up something like a cop chase in a pouring rain ).

[ViPER1313]

the sweep is more through and wont hurt your ears quite as much

Do you mean that this sample hurts ? In this case it's not played properly. You should hear 7 quiet dialup tones, nothing else.

No, I hear the 7 quiet beeps, and nothing else. I'm just saying that when this sample DOES clip (i.e. - When I max out the volume of my card...) the seven laser attacks can be painful to your ears. And your speakers . The when the sweep.wav starts to clip, the laser attack sounds start soft (at least for me - my clipping is caused by the EQ settings of the SB-Live drivers) and increase as the freq. gets higher. Not to mention that a max volume sine wave is much more effective / thorough way to test for this sort of clipping, as it covers all frequencies, and allows for you to tweak your volume / EQ settings while the sample is on repeat play. This is what I was trying to say.

[Halcyon]

A quick note about Creative cards and audible artifacts with this sample:

If you think you are hearing this sample right, you may not be playing it back loud enough. I know, I know. You should NOT PLAY THIS SAMPLE LOUD. It can destroy your equipment. Be very careful.

However, I noticed that with resampling done "almost right" (in subjective terms) you may not get any ambulance / chirping / alien sounds, but you may get a static, lower frequency noise (or even hum) all throughout the duration of the 19.5 kHz tone. This for example, when you set your playback software to resample the output to 96 kHz.

If you play it at low enough volume, you will not be able to detect this noise, but trust me it's there on Audigy 2 Platinum eX and I'm quite positive that it is also there on any Creative Live/Audigy sound card. Once you've played it loud enough to notice it, it's much easier to spot it even at a slightly lower volume.

I don't really believe a single creative card passes this test at ANY volume setting (at least not on Winamp resampling plugins). I've tried it down to 10% in Creative Surround Mixer (for both master volume control and wave playback volume control). The distortion and/or hum (depending your choice of resampling) is always there, it's just harder to spot at lower volumes (you need to have quite a lot of clean / external amplification outside your sound card).

As a comparison. I've played this sample back at insane levels (considering the 19.5 kHz full amplitude tone) on using RME DIGI 96/8 PAD. I can hear a faint amount of noise at the background (either in the sample or in my playback gear), but it is so miniscule and only audible at insane playback levels that the noise WaveOut or DirectSound Resampling plug-in in WinAmp 2.x produces on Creative cards is imho, clearly more audible.

BTW, one additional comment about the DirectSound plug-in for Creative cards (in Winamp), which I learned myself only now:

With the WinAmp 2.xx DirectSound resampling plug-in I cannot make the chirping go away regardless of what sample rate/bit depth/dithering/noise shaping/mode I choose, UNLESS I uncheck the "Allow hardware acceleration" in the DirectSound plugin configuration. As such, I'd recommend leaving it unchecked for any critical listening, if you are using a Creative based sound card.

regards,
Halcyon

[KikeG]

You should hear 7 quiet dialup tones, nothing else.

Well, in my case the sample was properly played, and I could "feel" the ultra-high pitched tones, even when the dialling sound was not very loud.

If someone hears something apart from the dialling tones, a high pitched tone at 19 or 20 KHz (the one that IS at the sample) sounds quite different from a 10 KHz or 15 KHz tone (possibly caused from intermodulation/resampling artifacts). In the first case, you really don't hear anything, but feel like somebody was pushing a needle inside your head, or some kind on nasty pressure inside your head. In the latter, you clearly hear someting, its an irritating high pitched note, but it's at your ears.

This post has been edited by KikeG: May 29 2003, 09:40

[DickD]

I really couldn't figure why it sounded so bad on my PC, i finally figured out that if i dropped the attenuation in Foobar2000 to about -18dB it finally sounded ok.
I thought this amount of attenuation was far too much, then i noticed i had replaygain enabled and the track gain for the file was abou +18dB. I disable replaygain, and when resampled to 48000Hz and now all i hear is a quiet (in relation to the dial tone) high pitched fluctuating noise!

I think this was mentioned by somebody who hadn't enabled ReplayGain clipping prevention. If you enable that it should sound fine.

I'd suggest that if anyone posts this test tone in future, they manually edit the ReplayGain values to 0.00 dB in FB2K or similar, unless of course, it's supposed to be used to flag up "badly" set up ReplayGain (i.e. with no clipping prevention).

It's great to have a sample to really work out what's the best set-up for a soundcard/driver you're unsure about.

Pio: I think your picture of the aliases is pretty much what I was hearing on my Pro-16v-Pnp card, except the aliases (cop cars) were much quieter, due to my brickwall filter being fairly good (just not perfect).

[mrosscook]

I have an SB Live card and Harman-Kardon desktop speakers, and I don't hear any artifacts in the udial sample using any of the audio players that I have installed (about 12 in all, though I mostly use Winamp 2.81, and Media Jukebox 8 to a lesser extent). I don't use DSP or equalizer plugins; I've played with them from time to time, but I prefer to have them always turned off by default.

I hear only the seven quiet digital-phone dialing tones -- no sirens, ambulances, etc. My ears are not golden by any means, but my wife can't hear any artifacts either.

Halcyon suggests that at sufficiently high volumes, I should hear artifacts, but I'm not going to test that out. The only way that I know to "be careful" when increasing volume would be to do it gradually and to stop when I can smell my tweeters, or when the desktop bursts into flame, whichever comes first.

[LPTB]

KikeG, Halcyon
You're both correct this is exactly what happened with my SBLive (reviewed in one of the earlier posts), even when using SSRC with foobar I heard the high freq. wave, it was there at any volume setting but at lower volumes I could only feel it (uncomfortable feeling) and at higher volumes also hear quite painfully so I don't think it reaches 19kHz but it's close.

EDIT: The test was done using: SBLive-Marantz-HD580

This post has been edited by LPTB: May 29 2003, 15:04

[Pio2001]

I remember people saying their SB live was incredibly improved by the SSRC plugin, and other saying it made no difference. It depended on the OS, and the "sample rate conversion quality" of Windows settings. People having a bad sounding SB live got the same result setting Windows' sample rate conversion to "best", or enabling SSRC playback.
Thus this sample might also sound very different from an SB live setup to another.

[JonPike]

Huh.. interesting.

On XP here, can't remember if I had to adjust the XP controls for maximum quality, or they were by default,
soundcard is an Audigy I, Winamp 2.91 with DS 2.2.3 SSRC set at 16b x 48Khz, sounds near perfect, only very low
level hiss. Sliders don't seem to cause a sudden increase in noise, just make things louder.

Of course is ist death at 41.1Khz, and many others.. but 16b and 96khz is pretty quiet as well.

Oh, a fair amount of noise is caused by the good 'ol Winamp EQ being on.. but we expected that, right?

Hmmmm.... now Fubar .60 is wierd.. I get heavy "laser blast" sound effects on every one I tried, (using DS and resampler, and no other DSP) except 32Khz!! Even Audigy native 48Khz rate had bad noise. What's going on here?

Guess I'll be irritating and say, "Hey, I'm sticking with Winamp, for better sound!!" };-) Though it is Peter's earlier SSRC code.