Sweep VS. Pink noise

[stsal...@gmail.com]

I have taken both a 2 min sweep measurement from listening position freq 20 to 20000 hz, as well a pink noise measurement. The results are shown in the attachment. The picks and valleys are virtually the same but look how skewd are the two curves... The question is then, which measurement represents perceived reality of sound?

[Julian Bunn]

That's a head-scratcher for me! It must be related to the energy density of the two signals: in the sweep, at any moment in time, only one frequency has energy, whereas in Pink at any moment in time the energy is distributed over the whole spectrum and falls as a function of frequency. But I can't quite see how that results in the curves you've measured :-) Perhaps if you make a curve using White Noise, it will look similar to the Sweep data? 

Julian

On Fri, May 22, 2020 at 5:05 AM <stsal...@gmail.com> wrote:

I have taken both a 2 min sweep measurement from listening position freq 20 to 20000 hz, as well a pink noise measurement. The results are shown in the attachment. The picks and valleys are virtually the same but look how skewd are the two curves... The question is then, which measurement represents perceived reality of sound?

--
You received this message because you are subscribed to the Google Groups "AudioTool Discussion Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to audiotool-discussio...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/audiotool-discussion-group/8af569cb-affb-44e4-8180-4fc8d69f1a52%40googlegroups.com.

[Sotiris Tsalkitzoglou]

Thank you for the attention. I run the white noise too. As predicted is closer to sweep data. Just in case, I run another pink noise test from the Stereophile cd and the results are very similar to the pink noise shown in the graph!

My ears tell me that the "correct" representation is the pink noise, but still... Any ideas? explanations? 

On Friday, May 22, 2020 at 7:57:15 PM UTC+3, Julian wrote:

That's a head-scratcher for me! It must be related to the energy density of the two signals: in the sweep, at any moment in time, only one frequency has energy, whereas in Pink at any moment in time the energy is distributed over the whole spectrum and falls as a function of frequency. But I can't quite see how that results in the curves you've measured :-) Perhaps if you make a curve using White Noise, it will look similar to the Sweep data? 

Julian

On Fri, May 22, 2020 at 5:05 AM <stsal...@gmail.com> wrote:

I have taken both a 2 min sweep measurement from listening position freq 20 to 20000 hz, as well a pink noise measurement. The results are shown in the attachment. The picks and valleys are virtually the same but look how skewd are the two curves... The question is then, which measurement represents perceived reality of sound?

--
You received this message because you are subscribed to the Google Groups "AudioTool Discussion Group" group.

To unsubscribe from this group and stop receiving emails from it, send an email to audiotool-discussion-group+unsub...@googlegroups.com.

[Julian Bunn]

Hopefully someone on the list more experienced with sound perception will chime in. In any case, these are very nice data!

Julian

On Fri, May 22, 2020 at 12:02 PM Sotiris Tsalkitzoglou <stsal...@gmail.com> wrote:

Thank you for the attention. I run the white noise too. As predicted is closer to sweep data. Just in case, I run another pink noise test from the Stereophile cd and the results are very similar to the pink noise shown in the graph!

My ears tell me that the "correct" representation is the pink noise, but still... Any ideas? explanations? 

On Friday, May 22, 2020 at 7:57:15 PM UTC+3, Julian wrote:

That's a head-scratcher for me! It must be related to the energy density of the two signals: in the sweep, at any moment in time, only one frequency has energy, whereas in Pink at any moment in time the energy is distributed over the whole spectrum and falls as a function of frequency. But I can't quite see how that results in the curves you've measured :-) Perhaps if you make a curve using White Noise, it will look similar to the Sweep data? 

Julian

On Fri, May 22, 2020 at 5:05 AM <stsal...@gmail.com> wrote:

I have taken both a 2 min sweep measurement from listening position freq 20 to 20000 hz, as well a pink noise measurement. The results are shown in the attachment. The picks and valleys are virtually the same but look how skewd are the two curves... The question is then, which measurement represents perceived reality of sound?

--
You received this message because you are subscribed to the Google Groups "AudioTool Discussion Group" group.

To unsubscribe from this group and stop receiving emails from it, send an email to audiotool-discussio...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/audiotool-discussion-group/8af569cb-affb-44e4-8180-4fc8d69f1a52%40googlegroups.com.

--

You received this message because you are subscribed to the Google Groups "AudioTool Discussion Group" group.

To unsubscribe from this group and stop receiving emails from it, send an email to audiotool-discussio...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/audiotool-discussion-group/1d70b15d-766e-469c-8854-14c13a6d0d04%40googlegroups.com.

[Jean-Baptiste ROMANET]

Hi.

Was it log or linear sweep ? 

Log sweep energy equals pink noise whereas linear sweep should result to white noise. 

Human perception is following a log scale, see articles about Weber-Fechner. 

This is true for noise intensity. 

Also for frequencies, our ear gives more resolution in low frequency than higher ones. This is due to the layout of the sensitive hair cells in the cochlea. 

The 1/n oct. analysis is following the same rule with a resolution growing opposite to the frequency bands center frequency. 

So, the natural hearing response should be a flat 1/n spectrum obtained from pink noise or log sweep. 

Thinking mechanically and not physiologically, a perfect system is giving a flat FFT (narrow bands) spectrum when playing a white noise or a linear sweep. 

I always prefer using 1/n and pink for acoustic measurements. 

Then, if a physical problem must be identified, FFT and white noise is helping. 

Sorry for my Fr-english! 

Jean-Baptiste 

To view this discussion on the web visit https://groups.google.com/d/msgid/audiotool-discussion-group/CAAxkZQtGcBJ%2B%2BW0TzS%2Bt4mRQUng%2B0nqCBYZG0Bq_jhVeqehaFQ%40mail.gmail.com.

[Brian Greenwood]

The answer to you question is, it depends. The professional trainers in acoustical and system design and tuning at SynAudCon/prosoundtraining.com will tell you square wave sweeps will always be more accurate than pink noise.  Pink noise is most useful for setting signal gain and amplification levels in systems where all those components are separate. Sweeps are what the top professionals use for measuring the audio frequency response of a room. 

Pink noise generators aren't perfect, even the best of them.  At any given frequency their output will fluctuate +/- 3dB (and that's the good ones). While a sweep is producing the exact same amplitude for every individual frequency it's producing.  Pink Noise however is designed to produce the same power per 1/3 octave.  Here is a quote from hyperphysics.com on pink noise:

Whereas white noise is defined as sound with equal power per Hz in frequency, pink noise is filtered to give equal power per octave or equal power per 1/3 octave. Since the number of Hz in each successive octave increases by two, this means the power of pink noise per Hz of bandwidth decreases by a factor of two or 3 decibels per octave.

The result is the power output at low frequencies is not the same as higher frequencies.  

However, you asked which is more representive of perceived sound.  That is a far more complicated question than which is more accurate.  Because humans don't hear frequencies equally across our range of hearing. We're more sensitive to higher frequencies than we are to lower one's.  For most people our sensitivity peaks at around 4,ooo Hz, which is why that is usually perceived as the harshest frequencies when playing back pure tones (square waves).  However, the exact differences of perception to measurable accuracy isn't the same from one person to the next.  Thus my original supposition that the answer to which is more representive of perceived sound is.. it depends. 

I believe a better question is which one is more accurate, and which one is better to use for the purpose of adjusting for room accustics (the colorization the room has in the audio reproduced by the speakers).  And that answer is without question the Sweep.  The sweep is an accurate representation of the combination of the direct sound from the speakers and the indirect sound of reflections and reverberations of the room. 

 In small room measurements (which I'm assuming you are working in) those indirect sounds are well within 50 milliseconds in delayed time from the direct sounds. In which case measuring in the time domain isn't necessary and you can make EQ adjustments directly from the measurement you posted above.  In large room acoustics (auditoriums, concert halls, etc) any indirect sound beyond 50 milliseconds after the arrival of the direct sound has to be filtered out because long reverberations cannot be corrected by EQ. 

Once you have corrected for the rooms reverberations using EQ you can then shape the sound to your or your client's preferences (perceived sound).  Attempting to do so before or without correcting for the acoustics of the room will result in an inability to achieve your goals using EQ to shape the sound to your desired goal.  The EQ simply won't be effective until those corrections are first performed. 

I hope that helps!! 

Brian Greenwood 

Former Audio Engineer and Systems Designer

To view this discussion on the web visit https://groups.google.com/d/msgid/audiotool-discussion-group/CAAxkZQtE8LV%3Dp-kE5iU3ctCbmxpz61%3Dkdy12OQ-X%3DgbqJJuAYg%40mail.gmail.com.

[Sotiris Tsalkitzoglou]

Thank you very much for the detailed and thorough answer, which however produced some further questions... 

Some background first. I have a fairly large for domestic standards room measuring 9.75 x 6.63 x 3.85 meters (32 x 21.75 x 12.63 feet). My system is just a stereo system - no equaliser, no subwoofer and generally considered as "hi-end". Speakers are very heavy and relatively difficult to move... although I have some limited flexibility for their exact positioning. Being rather a perfectionist, I try to get the best possible sound, through not equalisation, but through changing if necessary, speaker positioning (unfortunately, I have no ability to apply other form of acoustic treatment to the room...).

Now, coming to the questions I mentioned earlier: When I look at the two curves I measured in the room, i.e. the sweep and the pink noise, and I was asked to describe, how I imagine the sound to be for both, I would say that in the case of sweep, that I would expect a rather "bright" sound with relative lack of bass. Based on the pink noise however, my description would be exactly the opposite! However, in real life, my ears tell me that the later is correct.

If, in my case try to "correct" the frequency response by say moving the speakers closer to room boundaries, in order to boost those lacking bass frequencies and tow-out the speaker to compensate for the brightness, I sense I would achieve the complete opposite result!

One could seriously say, "finally" trust your ears". But, I am very curious as to why this huge (+10 - 15 db) difference in the lower frequencies, between the sweep signal and the pink noise, What would you do in my case in order to "better" the sound in the room? (Remember, I have no EQ, no subwoofer).

I would greatly value an answer! 

On Saturday, May 23, 2020 at 8:42:38 PM UTC+3, Brian Greenwood wrote:

The answer to you question is, it depends. The professional trainers in acoustical and system design and tuning at SynAudCon/prosoundtraining.com will tell you square wave sweeps will always be more accurate than pink noise.  Pink noise is most useful for setting signal gain and amplification levels in systems where all those components are separate. Sweeps are what the top professionals use for measuring the audio frequency response of a room. 

Pink noise generators aren't perfect, even the best of them.  At any given frequency their output will fluctuate +/- 3dB (and that's the good ones). While a sweep is producing the exact same amplitude for every individual frequency it's producing.  Pink Noise however is designed to produce the same power per 1/3 octave.  Here is a quote from hyperphysics.com on pink noise:

Whereas white noise is defined as sound with equal power per Hz in frequency, pink noise is filtered to give equal power per octave or equal power per 1/3 octave. Since the number of Hz in each successive octave increases by two, this means the power of pink noise per Hz of bandwidth decreases by a factor of two or 3 decibels per octave.

The result is the power output at low frequencies is not the same as higher frequencies.  

However, you asked which is more representive of perceived sound.  That is a far more complicated question than which is more accurate.  Because humans don't hear frequencies equally across our range of hearing. We're more sensitive to higher frequencies than we are to lower one's.  For most people our sensitivity peaks at around 4,ooo Hz, which is why that is usually perceived as the harshest frequencies when playing back pure tones (square waves).  However, the exact differences of perception to measurable accuracy isn't the same from one person to the next.  Thus my original supposition that the answer to which is more representive of perceived sound is.. it depends. 

I believe a better question is which one is more accurate, and which one is better to use for the purpose of adjusting for room accustics (the colorization the room has in the audio reproduced by the speakers).  And that answer is without question the Sweep.  The sweep is an accurate representation of the combination of the direct sound from the speakers and the indirect sound of reflections and reverberations of the room. 

 In small room measurements (which I'm assuming you are working in) those indirect sounds are well within 50 milliseconds in delayed time from the direct sounds. In which case measuring in the time domain isn't necessary and you can make EQ adjustments directly from the measurement you posted above.  In large room acoustics (auditoriums, concert halls, etc) any indirect sound beyond 50 milliseconds after the arrival of the direct sound has to be filtered out because long reverberations cannot be corrected by EQ. 

Once you have corrected for the rooms reverberations using EQ you can then shape the sound to your or your client's preferences (perceived sound).  Attempting to do so before or without correcting for the acoustics of the room will result in an inability to achieve your goals using EQ to shape the sound to your desired goal.  The EQ simply won't be effective until those corrections are first performed. 

I hope that helps!! 

Brian Greenwood 

Former Audio Engineer and Systems Designer

On Fri, May 22, 2020, 11:57 AM Julian Bunn <jjb...@gmail.com> wrote:

That's a head-scratcher for me! It must be related to the energy density of the two signals: in the sweep, at any moment in time, only one frequency has energy, whereas in Pink at any moment in time the energy is distributed over the whole spectrum and falls as a function of frequency. But I can't quite see how that results in the curves you've measured :-) Perhaps if you make a curve using White Noise, it will look similar to the Sweep data? 

Julian

On Fri, May 22, 2020 at 5:05 AM <stsal...@gmail.com> wrote:

I have taken both a 2 min sweep measurement from listening position freq 20 to 20000 hz, as well a pink noise measurement. The results are shown in the attachment. The picks and valleys are virtually the same but look how skewd are the two curves... The question is then, which measurement represents perceived reality of sound?

--
You received this message because you are subscribed to the Google Groups "AudioTool Discussion Group" group.

To unsubscribe from this group and stop receiving emails from it, send an email to audiotool-discussion-group+unsub...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/audiotool-discussion-group/8af569cb-affb-44e4-8180-4fc8d69f1a52%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "AudioTool Discussion Group" group.

To unsubscribe from this group and stop receiving emails from it, send an email to audiotool-discussion-group+unsub...@googlegroups.com.

[Jean-Baptiste ROMANET]

Again.
Lin sweep = white noise energy = flat spectrum in narrow bands "full res"

Log sweep = pink noise energy = flat spectrum in 1/n oct. bands (1/1 1/3 1/6 1/12...)

If your goal is to have a flat response for listening, best is pink noise or log sweep and 1/n oct. bands analysis.

Jean-Baptiste

[Matt Mathis]

I'm curious why you don't EQ:  The phase distortion introduced by carefully placing speakers (e.g. timing multi path echos against the primary signal) is on a par with the phase distortion introduced by digital EQ, which does a similar transformation to the signal electronically/mathematically.

Thanks,

--MM--
Matt Mathis  Home & Travel voice: 412-654-7529  SMS and Daytime: 650-417-3029

Email is best, please leave a message if you must call.
-------------------------------------------
Evil is defined by mortals who think they know "The Truth" and use force to apply it to others.

--
You received this message because you are subscribed to the Google Groups "AudioTool Discussion Group" group.

To unsubscribe from this group and stop receiving emails from it, send an email to audiotool-discussio...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/audiotool-discussion-group/445d9773-0309-467a-a8b9-531396f06724%40googlegroups.com.

[Sotiris Tsalkitzoglou]

Thank you for the suggestion. Any EQ equipment, adds coloration and distortion, which maybe outweighed by the benefits realized in a home theater setup, but not in a stereo music hi-end one. Hence you see very few high quality music systems incorporating digital EQ.