Hi Mike,
that wasn't actually the test I had in mind... though interesting (my first thought was, hmm, tests for channel crosstalk...).
What I had in mind is, if you roll off frequencies which by themselves are inaudible to you as a pure sine wave, does this "soften" regular, sharp transients as encountered in music? Or produce some other audible effect?
For this, complete removal of say, >16 kHz content is not necessary, just a steep rolloff. The most important thing being that level and phase below the cutoff should be as untouched as possible. Say if you hear until 14 kHz, make a cutoff at 16 kHz which at 14 kHz has less than 0.1 dB deviation. If possible. Phase will likely start shifting way below cutoff but that is part of the effect to be tested.
Am I making sense ... ?
that wasn't actually the test I had in mind... though interesting (my first thought was, hmm, tests for channel crosstalk...).
What I had in mind is, if you roll off frequencies which by themselves are inaudible to you as a pure sine wave, does this "soften" regular, sharp transients as encountered in music? Or produce some other audible effect?
For this, complete removal of say, >16 kHz content is not necessary, just a steep rolloff. The most important thing being that level and phase below the cutoff should be as untouched as possible. Say if you hear until 14 kHz, make a cutoff at 16 kHz which at 14 kHz has less than 0.1 dB deviation. If possible. Phase will likely start shifting way below cutoff but that is part of the effect to be tested.
Am I making sense ... ?
Hi MBK,
With a FFT you can cut off exactly without any phaseshift to other
frequencies. My last experiment was only a quick check if IMD in
the ear would make HF audible. I was tricked by a highly distortive
speakersystem. Right now i will continue with experimenting with
this signal, this seems very useful. As the signal itself is silent, any
distortion created by an amp should get audible by creating this
easily audible 1khz signal.
Programs like cooledit can use fft on music, it should be able to make
such a "perfect" cutoff. In theory you should not hear any difference
by cutting of all HF, but the sound might get softer as the amp
reproducing gets less HF, creating less problems ?
(depending on the quality of the amp)
Mike
With a FFT you can cut off exactly without any phaseshift to other
frequencies. My last experiment was only a quick check if IMD in
the ear would make HF audible. I was tricked by a highly distortive
speakersystem. Right now i will continue with experimenting with
this signal, this seems very useful. As the signal itself is silent, any
distortion created by an amp should get audible by creating this
easily audible 1khz signal.
Programs like cooledit can use fft on music, it should be able to make
such a "perfect" cutoff. In theory you should not hear any difference
by cutting of all HF, but the sound might get softer as the amp
reproducing gets less HF, creating less problems ?
(depending on the quality of the amp)
Mike
Hi Mike,
I am not "sattelfest" enough to comment on the FFT method, I have a hunch it could lead to other artefacts, but it's surely worth a try.
As for the effects I would "predict" a softening of the sound, not because of lighter load to the amp, but because of "rounded" transients. In addition I would predict a more uniform / less intelligible timing of transients on musical program material, again because of the softening of transients.
This whole test would be even more useful on material recorded in analog that was never in the digital domain - any music that had gone through A/D conversion already had a lowpass and a digitization applied to it and the whole test might get less conclusive. Say assuming a 16/44 A/D converter would apply a steep 22 kHz lowpass, how does the material compare before and after the filter, and how does it compare before and after the filter plus A/D plus D/A? I am sure this has been tested ad nauseam in the past 25 years of CD, but it would be interesting to actually hear for oneself... at the expense of having to dig out a tape machine, a good mic, and building say, an analog 48 dB/oct LP filter.... and no one would be able to ecxchange thew files ove the net... true DIY...
I am not "sattelfest" enough to comment on the FFT method, I have a hunch it could lead to other artefacts, but it's surely worth a try.
As for the effects I would "predict" a softening of the sound, not because of lighter load to the amp, but because of "rounded" transients. In addition I would predict a more uniform / less intelligible timing of transients on musical program material, again because of the softening of transients.
This whole test would be even more useful on material recorded in analog that was never in the digital domain - any music that had gone through A/D conversion already had a lowpass and a digitization applied to it and the whole test might get less conclusive. Say assuming a 16/44 A/D converter would apply a steep 22 kHz lowpass, how does the material compare before and after the filter, and how does it compare before and after the filter plus A/D plus D/A? I am sure this has been tested ad nauseam in the past 25 years of CD, but it would be interesting to actually hear for oneself... at the expense of having to dig out a tape machine, a good mic, and building say, an analog 48 dB/oct LP filter.... and no one would be able to ecxchange thew files ove the net... true DIY...
A good A/D converter samples at a much higher samplerate and
downsamples digitally afterwards. This way you don't need a filter
at 22khz that would impact also lower freqs.
The problem with your test could be, that the 48db/oct filter might
introduce artefacts, i have no idea how to be absolutely sure !
The best would be to use equipment working with 192khz samplerate,
and apply filer again digitally ? But who has such equipment...
btw, a correct filter would not really "round" the transients, it looks
more like symetrical ringing.
Mike
downsamples digitally afterwards. This way you don't need a filter
at 22khz that would impact also lower freqs.
The problem with your test could be, that the 48db/oct filter might
introduce artefacts, i have no idea how to be absolutely sure !
The best would be to use equipment working with 192khz samplerate,
and apply filer again digitally ? But who has such equipment...
btw, a correct filter would not really "round" the transients, it looks
more like symetrical ringing.
Mike
Attachments
I can actually hear both left and right channels, although the 20kHz was faint. When I played the signal back on my laptop through the built-in speakers I got a nice 1kHz tone alright. Shifting the balance control to either side removed the 1kHz tone. Through the headphones I could only hear the constituent tones - even if I had both caps to one ear. Apparently the intermodulation is created in the output stage of my laptop.MikeB said:
Guess that explains why I find it tiring to listen to that for long periods of time...
Rune
Humhoom, i am over 30 and can hear above 13khz... 13.5khz ! 
(And i am smoking...)
But there are plenty of people in this age still hearing up to 16khz.
Rune, this signal is quite bad, looking to it's shape, it's simply asking
for problems. It looks like a 20khz modulated with 1khz, but it really
contains only 19 and 20khz !
I think, a bad powersupply is already enough to make the 1khz audible...
Mike
(And i am smoking...)
But there are plenty of people in this age still hearing up to 16khz.
Rune, this signal is quite bad, looking to it's shape, it's simply asking
for problems. It looks like a 20khz modulated with 1khz, but it really
contains only 19 and 20khz !
I think, a bad powersupply is already enough to make the 1khz audible...
Mike
Attachments
hummhoom said:
I'm rather 40-something, but I can still hear those tones - at least at a rather exaggerated volume. The key might be that I've never smoked more that the odd cigar, and stayed clear of discos and rock clubs, unless I was playing myself.
Now, as far as to whom is making conversation through his/her rear end... I think Forrest Gump had some phrase that captured that fairly well.
Rune
bocka said:
But the TV line frequency is a possitively bassy tone at 15625 Hz
But it can sure be a pain to walk into a room with a TV that has a slightly defect inductor somwhere inside. I have to confess to having sprayed more than a few TV-s with plastic spray internalt to shut them up (back when I was a TV repair guy)
Rune
MikeB said:
Me, i don't need even a filter, a can't hear the original signal anyway...
Of course, when scoping, it's getting necessary...
Mike
I can't also, but there could be some damage to the ears (?) from the high level of 19 & 20K on the phones, and I want to keep what's left of my hearing...
MBK said:
I must take exception to this. There was NO ridicule. There was difference of opinion, and that difference was clearly argumented, and most of the times those arguments were ignored and unanswered by Graham.
You either have not read and/or understood the discussion or you chose to misrepresent the arguments. Sorry.
Jan Didden
Jan,
I did not mean you, re: mudfights over Graham. You have always been polite and respectful in discussion.
But I do remember quite a lot of snide directed at Graham over the years I've followed this forum, from a variety of people. I agree with you that as far as I remember Graham has always put his ideas in words, not formulas, and so, by not countering "d(theta)+omega phi =g^2!" with "but d sin omega theta /2 = c^3!!!" , he has not helped himself with regards to credibility.
OTOH he has developed his own circuits and presented them here, supported by data and all. Something you can't say from a number of his detractors. Again, this is not aimed at yourself.
I did not mean you, re: mudfights over Graham. You have always been polite and respectful in discussion.
But I do remember quite a lot of snide directed at Graham over the years I've followed this forum, from a variety of people. I agree with you that as far as I remember Graham has always put his ideas in words, not formulas, and so, by not countering "d(theta)+omega phi =g^2!" with "but d
sin omega theta /2 = c^3!!!" , he has not helped himself with regards to credibility. OTOH he has developed his own circuits and presented them here, supported by data and all. Something you can't say from a number of his detractors. Again, this is not aimed at yourself.
MBK said:Jan,
I did not mean you, re: mudfights over Graham. You have always been polite and respectful in discussion.
But I do remember quite a lot of snide directed at Graham over the years I've followed this forum, from a variety of people. I agree with you that as far as I remember Graham has always put his ideas in words, not formulas, and so, by not countering "d(theta)+omega phi =g^2!" with "but d
OTOH he has developed his own circuits and presented them here, supported by data and all. Something you can't say from a number of his detractors. Again, this is not aimed at yourself.
OK maybe I reacted a bit too strong. But in the thread on the FSD I haven't really seen anybody ridicule him, maybe on other occasions I haven't seen. But strong opposition, yes.
And he surely is to be commended for putting his ideas in hardware, but it is often the ideas that are discussed and taken apart, and that is what happened with FCD.
Jan Didden
I'm one that disagreed quite blatantly over Graham's postings.
The reason is that this forum is read by lots of people that looks for information - and finding myths without any very clear contrary view take myths as truth.
At least, they will be able to see the two sides of the coin and decide.
The reason is that this forum is read by lots of people that looks for information - and finding myths without any very clear contrary view take myths as truth.
At least, they will be able to see the two sides of the coin and decide.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.