Jump to content

Analysis of the HTC-Soundbug (tsss-noise)


Guest bazzo

Recommended Posts

I've made a further test. I wanted to test if the noise appears during a call.

So i called my SDAmusic with my girlfriends mobile and broadcasted a 3000Hz-sinetone to the SDA.

The noise did NOT appeare!!!

regards bazzo

Link to comment
Share on other sites

Guest Rui_Costa

the worst thing is i understood most of the stuff bazzo was going on about. looks like that university education was usefull then!

but seriously i think it is more prominent with different genres of music - in particular Bloc Party and Gorilaz (currently on my phone) not som prominent with faster rockier type music but is still audible on stuff like the killers.

solution: only listen to loud rock music!

but seriously this "tss" is really gonna get on my nerves cus now i'm gonna listen out for it!

Link to comment
Share on other sites

Guest Exploited

I had the *tsss* on my phone at the beginning. Then it didn't happen for several weeks only to re-appear again. To me this proves that it should be possible to write a software-patch that permanently removes the bug.

Once you start to use the phone as an MP3-player (like I've been doing these past few months) it starts to be REALLY irritating. :cry:

Link to comment
Share on other sites

Guest conflagrare
I had the *tsss* on my phone at the beginning. Then it didn't happen for several weeks only to re-appear again. To me this proves that it should be possible to write a software-patch that permanently removes the bug.

Once you start to use the phone as an MP3-player (like I've been doing these past few months) it starts to be REALLY irritating.  :cry:

<{POST_SNAPBACK}>

Thank you for doing such a complete and elaborate test on the phone! This is almost like one of those labs I do in university. :lol: I hope we can eventually get to the bottom of this...

thank you for your hard work!

Link to comment
Share on other sites

Guest PayableOnDeath

Great Work maybe we can get it fixed now :lol:

Test 9: Does the noise also a occur when playing back via internal loudspeaker?

A: Yes.

I havnt Noticed this on Loudspeaker on my C500

edit

jsut seen your test file post will try it now

Edited by payableondeath
Link to comment
Share on other sites

Guest alexis_winters17

Since you all seem to be in the know about sound issues, how about the one myself and The_Undertaker have been having? i.e. our phones have gone quiet! Despite a hard reset... :exclaim:

Link to comment
Share on other sites

Guest joelb69

Fantastic work bazzo! Have you forwarded the info to HTC? I never thought anyone would do such a comprehensive test. its odd it doesnt happen when taking a call? Im amazed that the SDA music has it aswel!

once again, truely impresive research

Link to comment
Share on other sites

Guest bazzo

hi, thanks for your appreciation;)

and Yes, i forwarded the Analysis to HTC, to Microsoft, to Dangaard and to T-mobile. I got no feedback so far...

I only found a PR emailaddress of HTC, so no idea if it was forwarded to a technican.

gerards bazzo

Link to comment
Share on other sites

Guest cold_fire

Nice work Bazzo !!! I've heard the noise as well and it's heard different depending of the kind of music played , but it has the same period aprox. 17 sec with anyplayer or file type (audio or video) . I have an Spv c500 orange in Romania and the guys at Orange Data said they'd tested it but not heard a thing and it may be a faulty file. I hope that HTC will solve the problem . Thx again man nice job !

Link to comment
Share on other sites

Guest bazzo

hi mates,

i got the following eMail today, from T-Mobile Service. I sent theme a link to my german analysis.

E-Mail von T-Mobile

Sehr geehrter Herr  xxxxxxx,

vielen Dank für Ihre E-Mail an T-Mobile.

Die von ihnen angesprochene Beeinträchtigung ist bekannt, und dem Hersteller mitgeteilt worden. Auch andere Auffälligkeiten, die bei uns auftauchen werden an den hersteller gegeben. Leider wird ein Update etwas dauern, aber wenn es verfügbar ist, finden sie es auf unserere Homepage unter www.t-mobile.de.

Bitte haben sie Verständnis dafür, das wir nicht mit Daten aus privat erstellten Foren arbeiten können, da diese Angaben für uns sehr zeitaufwendig wären und teilweise auch nicht immer auf einen Endgerätefehler deuten, sondern auf eine Fehlbedienung. Eine Trennung ist da schwierig, da man keinen direkten Zugriff auf den Ersteller und das jeweilige Gerät hat. Anmerkungen unserer Kunden geben wir selbstverständlich umgehend weiter.

Mit freundlichen Grüßen

xxxxxxx

T-Mobile Kundenservice

Technisches ServiceCenter

TRANSLATION:

Dear Mr. xxxxxxx

Many thanks for your eMail to T-mobile.

The bug you mentioned is known and forwarded to the manufacturer. Although other noticeable problems were forwarded to the manufacturer. Im sorry that the update will take a while, when its available, you can download it at our homepage (www.t-mobile.de). Please have appreciation, that we can't work with data from private forums.....

Best regards

X.XXXXXX

T-Mobile customer service

Technical ServiceCenter

regards bazzo

Link to comment
Share on other sites

Guest underwriter

Hi, I think I've found a really simple way of hearing the tss noise (damn it to the bowels of hell) plug yer headphones in, go into media player and go to the playlist. Using the joystick scroll down the list of tracks to the bottom, you'll notice the ticky noise mplayer makes as you do this doesn't stop when you reach the end of the list. Keep the joystick pressed to continue the ticky noise and every 15 or so seconds the acursed noise will be heard.

Hope this works for you, this however suggests that the tss noise (wince) is not connected to any variables of MP3/WMA encoding/decoding as the ticky noise is a built in part of mplayer and when the sound card is active (actually producing a audable sound) the noise is processed by the sound card hardware. I read in older threads it could be when the phone comunicates with a base station, from what I understand (I'm no expert) mobiles constantly (10-20sec intervals?) check the nearest base to acertain coverage etc.

I've only had my c500 for 12 days supplied with latest update and was gutted when I realised the tss was there for ever. I understand Orange have a 14 day return? can anyone confirm as I don't see any point in keeping it as the noise just ruins the whole experience of listening to music on the c500, as the huge screen and MP3 player were the reason I bought it, so I didn't have to carry a mobile and mp3 player as two devices.

Reading the threads it looks unlikely that a fix is anywhere near.

I cringe every time I hear that bloody noise.

Link to comment
Share on other sites

Guest jonnegri

Sorry I don't mean to be rude but, what a waste of time. I hardly think that anyone is really that interested.

Again, sorry it this offends.

Analysis of the audio-bug on HTC Smartphones

...also known as "tsssss-noise"

Summary of the test results

Test 1: Which relevance has the audioformat/compression?

A: It is not relevant.

Test 2: Are there regularities in the apperance of the bug?

A: Yes, the bug appeares all 16,680s and lasts 110ms.

Test 3: Which influence does the frequency of the input signal have on the level of the noise?

A: The level of the noise increases with increasing input frequency logarithmically proportionally.

Test 4: Which influence does the Freq. of the input signal have on the spectrum of the noise?

A: There are characteristic frequencies developed, additionally a whitenoise is superposed.

Test 5: Which influence does the level of the input signal have on the noise?

A: The level of the noise increases equivalent with the level of the input signal.

Test 6: Which influence does the betaPlayers Equalizer have on the noise?

A: It does not have an influence.

Test 7: Which influence do other settings of the test-smartphone have?

A: I found no relevant settings for the bug.

Test 8: Which influence does the quantization of the audio data have?

A: It does'nt have an influence.

Test 9: Does the noise also a occur when playing back via internal loudspeaker?

A: Yes.

The detailed test

The aim of this analysis is the more accurate identification of the, from many users discovered, "TSSSSS" noise with the mp3-playback on HTC Smartphones. This noise makes the music unenjoyable. Affectet devices are only usable als Audioplayer in a streetcar.

Test-setup:

The test equipment was a SDAmusic (aka "HTCamadeus"), it is attached to the line-in of the computers soundboard. As test software I used a 30day-trial of Adobe Audition. The test-smartphone was always in flightmode to exclude interferences by the GSM module.

Execution:

All tests were accomplished as follows. First, an audio test-file was generated with the AdobeAudition noise generator. This file was then transferred in each case to the test-smartphone and played back with the betaPlayer. Its Equalizer was always switched off, if explicitly not mentioned, and the volume of the test-smartphone was adjusted to maximum.

The audiosignal was recorded and analyzed, over the soundboard, with Adobe Audition. The input of the soundboard was adjusted, so that the recorded signal has nearly the same level as the generated input file. Further details on execution are to be found with the appropriate tests.

In first pre-tests, it turned out that a very high sinetone is well suitable to provoke the tssss-noise.

Test 1

Which relevance has the audioformat/compression?

Testtone: 16000Hz sinetone (level -3dB) - 44,1kHz/16b/stereo

This testtone was stored in different files with different compressions.

Testfile formats:

mp3(lame3.90.2 --alt -preset standard)

wma (128kbps CBR)

wav (PCM uncompressed)

Result:

The noise occurs, with all testfiles, with same level. Since it is audible with the uncompressed wav-file, I assume the compression does not have influence on the error. A software fault of the mp3-decoders is to be excluded therefore.

Test 2

Are there regularities in the appearance of the bug?

Testtone: 16000Hz sinetone (level -3dB) - 44,1kHz/16b/stereo

Testfile formats: wav (PCM uncompressed)

I used "Auditions" notch-filter to eliminate the input sinetone in the recording. Otherwise no further manipulations were accomplished.

(for the screenshot I have deleted everything before the first noise, therefor the beginning of the bug-noise starts at 0ms on timeline.)

* "notch filters" - eliminates extremely narrow-band a certain frequency

1.The length of the noise

noise no.      1 |  2 | 3  |  4 |  5  |  6  | 7  | 8  |  9  | 10 |

length(ms) 114|110|109|112|111|116|110|112|114|110|

Based on 10 mesurements, there is a duration of approx. 110ms. It is remarkable that the noise appeares 55ms on the right channel and only thereafter 55ms on the left channel (see fig.1)

abb1.GIF

fig.:1 Waveform of the disturbance after notch filtering of 16kHz

2. Spacing between the noises

noise no.    1-2    |  2-3  |  3-4  |  4-5    |  5-6  |  6-7 |  7-8  |  8-9  |  9-10 |

intervall(s)|16,679|16,681|16,679|16,680|16,680|16,680|16,681|16,680|16,681|

I've mesured the intervall between the beginning of a noise and the beginning of the respective next noise.  The interval is on an average of 16,680s.

abb2.GIF

fig.:2 intervall between 4. and 5. noise (consider time code in the field ("Sel.";"Begin") = position of the marker)

Result:

The noise appears all 16,68s, when continuous playing the test-file, and lasts approximately 110ms(55ms left, then 55ms right).

During the interrupted playback, thus with intermediate pausing of the playback, the intervall between the noises changed. There are no more regularities recognizable. Within continuous played passages, the intervall between the noises was however again 16,68s. The length, of 2x55ms, of the noise did not change.

Test 3

Which influence does the frequency of the input signal have on the level of the noise?

Test tones:

16000Hz sinetone (level -3dB) - 44,1kHz/16b/stereo

8000Hz sinetone (level -3dB) - 44,1kHz/16b/stereo

4000Hz sinetone (level -3dB) - 44,1kHz/16b/stereo

2000Hz sinetone (level -3dB) - 44,1kHz/16b/stereo

1000Hz sinetone (level -3dB) - 44,1kHz/16b/stereo

500Hz sinetone (level -3dB) - 44,1kHz/16b/stereo

250Hz sinetone (level -3dB) - 44,1kHz/16b/stereo

50Hz-22kHz whitenoise (level -3dB) - 44,1kHz/16b/stereo

Testfile format: wav (PCM uncompressed)

In these resulting recordings, I eliminated the sinetone by using the notch-filter with the particulare frequency. The maximum levels were mesured in the waveform, in each case 3 noises and averaged afterwards.

input freq.(Hz) 16000 8000 4000 2000 1000 500 250

noiselevel (dB) -2, -6, -12,-16,-24,-27,-37

abb3.GIF

fig.: 3 level of the noise over the logarithmical inputfrequency

translations:

"Eingangsfrequenz"= input frequency

"max. Pegel der Störung"= maximum level of the noise

Result:

It becomes clear that the level of the noise rises linear with the logarithm of the input frequency. It is particularly remarkable that the interference exceeds the level of the input signal (-3dB) starting from a frequency of 16kHz. The peaks of the noise are to be already recognisable before notch-filtering the sinetone.

When playing the file with the whitenoise, the disturbance wasn't measurable. The "noise filter" left no disturbance. If there was a noise at all, it went down in the backgroundhiss.

Test 4

Which influence does the frequency of the input signal have on the spectrum of the disturbance?

Test tones:

16000Hz sinetone (level -3dB) - 44,1kHz/16b/stereo

8000Hz sinetone (level -3dB) - 44,1kHz/16b/stereo

4000Hz sinetone (level -3dB) - 44,1kHz/16b/stereo

2000Hz sinetone (level -3dB) - 44,1kHz/16b/stereo

1000Hz sinetone (level -3dB) - 44,1kHz/16b/stereo

500Hz sinetone (level -3dB) - 44,1kHz/16b/stereo

250Hz sinetone (level -3dB) - 44,1kHz/16b/stereo

Testfile format: wav (PCM uncompressed)

abb4.GIF

Fig.: 4 spectrum of the noise at 250Hz

abb5.GIF

Fig.: 5 spectrum of the noise at 500Hz

abb6.GIF

Fig.: 6 spectrum of the noise at 1000Hz

abb7.GIF

Fig.: 7 spectrum of the noise at 2000Hz

abb8.GIF

Fig.: 8 spectrum of the noise at 4000Hz

abb9.GIF

Fig.: 9 spectrum of the noise at 8000Hz

abb10.GIF

Fig.: 10 spectrum of the noise at 16000Hz

The figures(4-10) show the spectrum of the noise (RED) compared, with that spectrum of the continuous backgroundhiss(BLUE)  which is superposed all recordings. I assume that the peaks recognisable in the noise and in the hiss, result from an overdrive of the smartphones amp or the soundboards line-in.

This clipping simply cuts the peaks off of the sine function and produces higher frequency portions (furrier-progression-components of higher order) which are not effected by the notch-filter. These peaks are normal in this case and have nothing to do with the bug.

It is remarkable that into far divide the spectrum of the noise is congruent. It seems that the bug simply amplifies the hiss, but thats not the truth. The hiss is not added until the recording. Therefore is to be assumed, that a part of the noise is a "whitenoise", which increases accordingly with increasing frequency (test 2).

Another much strange part of the noise is characterised by sinetone newly developed. Considering e.g. the pair of peaks in (fig.: 4) around 9500Hz, some similar peaks are also with 6000Hz, 15500Hz and with 19000Hz clearly recognizable. These peaks seem to be pushed apart with increasing frequency of the input signal (see fig.:5-10) around exactly the amount of the frequency of the input signal. In (fig.:8) the pair of peaks is now expanded from 9500Hz to 5500Hz and 13500Hz. The peaks moved thus in each case around 4000Hz, the frequency of the input signal. This connection is up(fig.: 9) to be clearly recognized.

The notch in the spectrum of the noise, with the respective input frequency, results from the notch-filtering.

Test 5

Which influence does the level of the input signal have on the noise?

Test tones:

16000Hz sinetone (level - 3dB) - 44,1kHz/16b/stereo

16000Hz sinetone (level -10dB) - 44,1kHz/16b/stereo

16000Hz sinetone (level -20dB) - 44,1kHz/16b/stereo

16000Hz sinetone (level -30dB) - 44,1kHz/16b/stereo

Testfile format: wav (PCM uncompressed)

Level input signal (dB) -3,-10,-20,-30

Level noise (dB) -2,-12,-23,-33

The level of the disturbance decreases approximately equivalent with the level of the input signal. The level of the input signal has apparently no influence on the spectrum of the noise.

abb11.GIF

fig.: 11 spectrum of the noise for the input signal with a level of

(-3dB-green, -10dB-red, -20dB-blue, -30dB-yellow)

Test 6

Which influence does the betaPlayers Equalizer have on the noise?

Testtone: 16000Hz -Sinuston (level - 3dB) - 44,1kHz/16b/stereo

Testfile format: wav (PCM uncompressed)

EQ-settings (betaPlayer):

preamp: neutral (mid)

60Hz: max

170Hz-14kHz: linear decreasing

14kHz: min

16kHz: max

Its obvious, that the EQ has no influence on the spectrum of the noise. The amplification of the 16kHz-band in the EQ results in a light increase of the noiselevel. This test shows, that the noise developes after the audiostream passed the EQ. If the noise would have been developed prior to the EQs influence, their noises spectrum has to be decreasing with increasing frequencies.

abb12.GIF

fig.: 12 noise spectrum with EQ (green: EQ off)

(blue: EQ on with mentioned settings)

Test 7

Which influence do other smartphone-settings have?

Testtone: 16000Hz -sinetone (level - 3dB) - 44,1kHz/16b/stereo

Testfile format: wav (PCM uncompressed)

1. StarUp depressed by renaming \storage\windows\StartUp

no influence

2. Flightmode on/off

no influence

3.Bluetooth on/off

no influence

4. Activesync-connection via USB est./not est.

no influence

Test 8

Which influence has the quantization of the audio-data?

Testtones:

16000Hz -sinetone (level - 3dB) - 44,1kHz/16b/stereo

16000Hz -sinetone (level - 3dB) - 44,1kHz/8b/stereo

16000Hz -sinetone (level- 3dB) - 22,05kHz/8b/stereo

23000Hz -sinetone (level -3dB) - 48,00kHz/16b/stereo (added on 02-15-05)

Testfile format: wav (PCM uncompressed)

There are no differences in results, same length, same intervall and same spectrum. The quantization has no influence.

Test 9

Does the bug also occur in playback via internal speaker?

Testtone: 16000Hz -sinetone (level - 3dB) - 44,1kHz/16b/stereo

Testfile format: wav (PCM uncompressed)

I've started the test with min. volume and increased it evry 20sec.by pressing the side-button. Already at 2/5-volume was the noise audible clearly through the internal speaker.

Conclusion:

Cause of the many articles about this bug in this forum, i would allmost claim, that this bug is detectable at all HTC-Smartphone-modells, where the bug is in at least one case known. I know that some of you would decline this, but i think the facts argue for themselves. People who claim that they don`t have this bug, IMHO simply don't hear it. It took a long time for me to notice the bug too, i've heard many many songs on my mobile without noticing the bug, but then i've heard Genesis "Watcher of the skies" and the bug was more than audible. Since this experience, i started to hear the noise in nearly every song i listened to concentrated. The regularity in the bugs appearance and the many many users who noticed that bug, make it hard to believe that there is any HTC-smartphone without that bug, these can't be isolated cases!

The tests have shown, that the appearance of the bug depends on the frequencies in the played audio-file, thus its clear, that the audibility and the annoy-potential of the bug is strongly related to the kind of prefered music. For instance a piano or string solo with isolated heights should make the bug more audible than a rock-song with fast drums and overdrived e-guitars. Even for "goldears" the bug should be hardly audible in this case. The more complex and broadbandy a song is, the harder the bug should be audible. Furthermore the noise goes under in the music itselves.

© by Bazzo 2005

i hope my translation is not to bad for understanding, sorry for bad english 8-[

best regards Bazzo :D

<{POST_SNAPBACK}>

Link to comment
Share on other sites

Guest PayableOnDeath
Sorry I don't mean to be rude but, what a waste of time.  I hardly think that anyone is really that interested.

Again, sorry it this offends.

<{POST_SNAPBACK}>

erm not really a Waste of time as this proves there is a bug and the HTC will need to look in to it

Link to comment
Share on other sites

Guest conflagrare
Sorry I don't mean to be rude but, what a waste of time.  I hardly think that anyone is really that interested.

Again, sorry it this offends.

<{POST_SNAPBACK}>

buddy... it's this kind of in-depth research that enabled people to design and build phones like this.

Apparently whoever took the time to design everything in your home wasted their time too.

Edited by conflagrare
Link to comment
Share on other sites

Guest joelb69
Sorry I don't mean to be rude but, what a waste of time.  I hardly think that anyone is really that interested.

Again, sorry it this offends.

<{POST_SNAPBACK}>

This post and related posts have had over 300 replys and 15,000 views, Quite clearly people ARE interested.

Your the waste of time not bazzos efforts.

And please dont try and be polite about your naivety, its not going to make it any less obvious.

Edited by joelb69
Link to comment
Share on other sites

Guest MadduckUK

well i care for one.. the hiss bug has stopped me from buying this phone so far. if a solution is found i might pick one up so ill be watching the thread closely.

Link to comment
Share on other sites

Guest MadduckUK

sorry for double post but on a more on-topic subect. we know that the sound is being produced both through the earphones and internal speaker. and it is not connected to the type of sound being produced (scrolling down list and mp3 playback are examples). but is this *sound* being produced when no other audio is and therefore the speaker/eaphone socket are not *active* for it to be heard?

if it could be proved that this *hiss* was happening ALL THE TIME, but only audiable at certain times (when the sound is active) it would rule out a lot of possibilities.

The only way i can think of to check for this would be by purpousfully trying to get feedback from the phone using radio/other electronic devices.

anyway thats probably all a load of pony, but just a thaught :D

Link to comment
Share on other sites

Guest bazzo

there is no hiss, when the device doesn't play some audio-content. i've made a test to check if the noise occures during calls. girlfriends nokia(3000Hz->SDAmusic->PC->Adobe Audition. there seems to be no hiss during calls.

regards bazzo

Link to comment
Share on other sites

Guest conflagrare
hi mates,

i got the following eMail today, from T-Mobile Service. I sent theme a link to my german analysis.

TRANSLATION:

regards bazzo

Dear Mr. xxxxxxx

Many thanks for your eMail to T-mobile.

The bug you mentioned is known and forwarded to the manufacturer. Although other noticeable problems were forwarded to the manufacturer. Im sorry that the update will take a while, when its available, you can download it at our homepage (www.t-mobile.de). Please have appreciation, that we can't work with data from private forums.....

Best regards

X.XXXXXX

T-Mobile customer service

Technical ServiceCenter

<{POST_SNAPBACK}>

By the way, did you email them back with a pdf copy of the report? or do those fools think a report is not valid because it is in a forum and is not printed on letterhead paper in a nice report folder?

Also, while they are at it, please tell them to check the Digital to Analog Converter. DAC. I suspect that thing is busted... If that's not it, check stray capacitance and inductance too.

If it's a DAC embedded Texas Instrument microprocessor, give them hell and file a lawsuit.

Link to comment
Share on other sites

Guest bazzo
By the way, did you email them back with a pdf copy of the report?  or do those fools think a report is not valid because it is in a forum and is not printed on letterhead paper in a nice report folder?

Also, while they are at it, please tell them to check the Digital to Analog Converter.  DAC.  I suspect that thing is busted...  If that's not it, check stray capacitance and inductance too.

If it's a DAC embedded Texas Instrument microprocessor, give them hell and file a lawsuit.

<{POST_SNAPBACK}>

i think there is nothing left to do for us else than waiting. as you can read the bug is known to t-mobile and reported to the manufacturer. so if the bug still will be fixed soon, everithing is fine. if not i'll change to SDA II music, which is announced for june in germany 8)

regards bazzo

Link to comment
Share on other sites

  • 2 weeks later...
Guest tha_oli

srry if im beeing stupid but the tsss bug has been around since the nice sharp edged first spv, so mayb they dunno how to fix it :s that would suck ass, lets all hope im wrong and that orange will bring out a update/patch

Link to comment
Share on other sites

Guest bazzo
...

Test 7

Which influence do other smartphone-settings have?

Testtone: 16000Hz -sinetone (level - 3dB) - 44,1kHz/16b/stereo

Testfile format: wav (PCM uncompressed)

1. StarUp depressed by renaming \storage\windows\StartUp

no influence

2. Flightmode on/off

no influence

3.Bluetooth on/off

no influence

4. Activesync-connection via USB est./not est.

no influence

...

<{POST_SNAPBACK}>

Link to comment
Share on other sites

  • 2 weeks later...

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.