Some good things have come out of that thread that I don't see happening on AVS - some of the fence-sitters took Winer's test & confirmed that they could hear 20 passes & surprised themselves with the fact that ABX results proved they could differentiate 1 pass but couldn't pinpoint or didn't know what the differences were. One, open-minded, guy admitted that he realised his negative bias was stopping him from hearing any differences in the past & that this has now opened up a whole new vista for him - he learned a different way of listening & learned to have more confidence in what they were hearing & not concerned about having to describe the differences for others. This to me is the benefit of ABX - it gives results that can't be argued over so there's less/no anxiety in defending what is heard (which I think creeps into most blind listening)
Conclusive "Proof" that higher resolution audio sounds different
- Thread starter amirm
- Start date
Couple of problems though with them putting IM distortion as the cue.
A) hirez music in nearly every instance has ultrasonic content at least 40db below that of the fundamental musical note and usually it is 60db down, so the IMD would need a substantial boost somehow or require sounds/tones that do not reflect "normal" complex sounds or music.
b) music also has both odd and even harmonics further masking IMD.
c) IMD for audibility usually requires the good electronic component to either clip (might happen for DACs if the signal is 0dbfs but is that really clipping, or amps in a way we all know) or behave in a non-linear fashion.
d) Much equipment is measured for IMD, it is not a difficult test measurement and if they argue such as test is too simple then they must agree most of the other audio test measurements relying upon simple test tones are also flawed
Two caveats; SACD/DSD may (emphasis on may as not guaranteed and this was not part of the test anyway) have an interesting aspect on IMD due to the hump (so increased amplitude) of its ultrasonic noise (which is not the same as white noise from PCM).
The other is that very cheap converters (especially some older ones) can suffer IMD, but then this still needs to take into account its attenuation that would be well below that of the fundamental tone.
Sorry if wrong but I am sure Amir also proved a difference even with some music samples created and not just jangling keys (which does have an unusually high ultrasonic content but even so one would need to identify/prove a DAC or other component behaves non-linear which tbh can be measured.
Their argument and way to suggest Amir is hearing IMD is unfortunately a really bad way of proving it IMO, but I get the feeling it is more about creating doubt to invalidate the results of Amir and others (when there are other reasons they should investigate such as the transparency including downsampling Bit-Sampling Rate/recording of the hirez files/etc or less obvious but even possibility of stopband rejection/impulse response and filters used with specific sampling rates).
This ignores bit depth focus-scope of testing though and potential areas this can be influenced.
Anyway the onus should not be on Amir, but those who proposed and created the test should investigate why in a methodical and thorough way; kudos for Amir carrying on with the various testing and taking time to do that and also explain aspects such as the fact and why his system does not have IMD (nor did JA's).
As always just IMO.
Cheers
Orb
Last edited:
Yes, running through Audio Diffmaker shows an offset of 13.53usec between Arny's 24/96 & the downsampled tracks. This is the equivalent of about half a sample - I doubt that it's audible although I do know that the just noticeable Interaural time difference is around 4usecs. But what we're talking about here is a shift of both stereo channels by 13.53usecs, not shifting one of the stereo channels.
I'm confused here about what files you are referencing - I thought the AVS files were ArnyK's files, no?
BTW, be careful with AudioDiffmaker - it seems to have some problems:
- it seemingly aligns the two files getting rid of the offset between them but introduces some shift anomaly in the first 0.177secs of the aligned file. This causes a reduced reported null value in Diffmaker.
- I brought the aligned file (16/44) from Diffmaker into Audition, inverted it & nulled it against the original 24/96 file. What should be seen is a good null for all frequencies up to 22K but you don't. What you see in the first 0.177secs is a bad null fluctuating -60dB, -70dB pretty much jumping all over as you move the pointer into the file & look at the frequency analysis plot. If you select the whole file & scan it for freq plot, you get about the same null figure as Diffmaker reports, -80dB. However, if you skip the first 0.177secs & scan the rest of the file you get a much better null of about -100 or -110dB.
- so my conclusion is that if there is an offset between files, Diffmaker makes a mess of aligning it which results in a lower null.
- one other thing is that there seems to be timing drift between the files which if not turned on in Diffmaker gives a much worse null of -30dB or so
To be fair, Orb, the jangling keys audio clip does have a very high level of energy in the ultrasonic frequencies, much more than any music recorded (right up to 45KHz). This & the fact that the recording is at low volumes has been used as an explanation for possible IMD artifacts appearing in the audible output (the volume has to be turned up above normal which may stress the audio devices more than normal -close to clipping). The other factor is that a seminal presentation by Griesinger has already analysed the jangling keys test with the conclusion that it causes IMD in amplifiers. So, I would favour testing for IMD
From the presentation:
However, the Hydrogen audio positive ABX results from 2010 were done with music files (unfortunately I can't find the files) so I'm thinking that this might not be the key.
From the presentation:
However, the Hydrogen audio positive ABX results from 2010 were done with music files (unfortunately I can't find the files) so I'm thinking that this might not be the key.
Last edited:
Thanks Orb, I forgot about Scott's files & haven't analysed them. So Scott' files also show offset differences too, according to esldude. I wonder why they have an offset because when I do a downsample from 24/96 using Audition or R8Brain I get no offset - hmmm, interesting!
I'll have a look at them - I only analysed ArnyK's files
For sure and no disagreement but scope and context are critical, I did mention jangling keys having unusually high ultrasonic energy but that does not reflect "normal" musical sounds, even so this is no guarantee to cause IMD unless the electronics are pushed beyond linear behaviour.
Didn't Griesinger also state that to hear the IMD one had to remove ALL content below 20khz first.
For normal operation where no manipulation has happened, then one must force the electronic equipment to be non-linear/clip.
I will double check though.
Cheers
Orb
Last edited:
Ok Just loaded up the presentation of his I have; it is the "Perception of mid frequency and high frequency intermodulation distortion inloudspeakers, and its relationship to high-definition audio".
He does state:
On his Sopranino analysis
Maybe I am wrong but I think a lot of the test was back in 2003, therefore IMD/non-linear behaviour was more prevalent for cheaper internal components.
Did he do a follow up to this presentation-study I have of his?
Also back then digital signals louder than -3dbfs (due to waveform) could behave in a non-linear way.
OK..
Double checked that is the same one I have so his study is 2003, which explains some aspects of possible IMD with very cheap components, but caveats I raise still stand.
Key point from his tests that to hear IMD he had to remove sub 20khz content and the ultrasonic signal had to be played at high levels; in essence stress testing an amp beyond its normal operating spec.
Worth noting jangling keys were still inaudible on a high quality amp in context of the above sentence and the unusual stress test scenario.
Cheers
Orb
He does state:
On his Sopranino analysis
Maybe I am wrong but I think a lot of the test was back in 2003, therefore IMD/non-linear behaviour was more prevalent for cheaper internal components.
Did he do a follow up to this presentation-study I have of his?
Also back then digital signals louder than -3dbfs (due to waveform) could behave in a non-linear way.
OK..
Double checked that is the same one I have so his study is 2003, which explains some aspects of possible IMD with very cheap components, but caveats I raise still stand.
Key point from his tests that to hear IMD he had to remove sub 20khz content and the ultrasonic signal had to be played at high levels; in essence stress testing an amp beyond its normal operating spec.
Worth noting jangling keys were still inaudible on a high quality amp in context of the above sentence and the unusual stress test scenario.
Cheers
Orb
Yes, he did say that audibility (JND) of ultrasonic IMD was in the absence of signal <20KHz - still no harm in bearing it in mind.
I also forgot to mention that Amir also posted positive ABX results for Scott's high-res music files so the conditions that might apply to the jangling keys files (high energy ultrasonic frequencies recorded at low volume) probably don't apply to Scott's files.
Just tried to run two of Scotts files through Diffmaker to check for offset, timing drift, etc but it reports our of memory. Any other way to do this?
Oh yeah for sure, but the risk is that some will ignore the context-scope-limitation as we are seeing on AVSF, and possible others may not appreciate its lack of importance in the scheme of things compared to some of the other variables touched upon by many of us in the past.
Worth noting Amir and JA checked their system for IM, but in reality it is more of an issue for power amps with loudspeakers rather than headphone ABX played at normal and not extreme high levels - again IMO and we have IM distortion measurements by such as JA/Paul Miller/etc for reviewed products.
Cheers
Orb
Couple of problems though with them putting IM distortion as the cue.
A) hirez music in nearly every instance has ultrasonic content at least 40db below that of the fundamental musical note and usually it is 60db down, so the IMD would need a substantial boost somehow or require sounds/tones that do not reflect "normal" complex sounds or music.
b) music also has both odd and even harmonics further masking IMD.
c) IMD for audibility usually requires the good electronic component to either clip (might happen for DACs if the signal is 0dbfs but is that really clipping, or amps in a way we all know) or behave in a non-linear fashion.
d) Much equipment is measured for IMD, it is not a difficult test measurement and if they argue such as test is too simple then they must agree most of the other audio test measurements relying upon simple test tones are also flawed
Two caveats; SACD/DSD may (emphasis on may as not guaranteed and this was not part of the test anyway) have an interesting aspect on IMD due to the hump (so increased amplitude) of its ultrasonic noise (which is not the same as white noise from PCM).
The other is that very cheap converters (especially some older ones) can suffer IMD, but then this still needs to take into account its attenuation that would be well below that of the fundamental tone.
Sorry if wrong but I am sure Amir also proved a difference even with some music samples created and not just jangling keys (which does have an unusually high ultrasonic content but even so one would need to identify/prove a DAC or other component behaves non-linear which tbh can be measured.
Their argument and way to suggest Amir is hearing IMD is unfortunately a really bad way of proving it IMO, but I get the feeling it is more about creating doubt to invalidate the results of Amir and others (when there are other reasons they should investigate such as the transparency including downsampling Bit-Sampling Rate/recording of the hirez files/etc or less obvious but even possibility of stopband rejection/impulse response and filters used with specific sampling rates).
This ignores bit depth focus-scope of testing though and potential areas this can be influenced.
Anyway the onus should not be on Amir, but those who proposed and created the test should investigate why in a methodical and thorough way; kudos for Amir carrying on with the various testing and taking time to do that and also explain aspects such as the fact and why his system does not have IMD (nor did JA's).
As always just IMO.
Cheers
Orb
I don't think the IMD argument invalidates Amir's results at all, but it may define what it is Amir is hearing. The IMD argument, as I understand it, is not that IMD in the supersonic range can be heard, but that distortion in both amplifiers and transducers tends to increase dramatically at the lowest and highest frequencies, and that their attempt to reproduce ultrarsonic content along with the audible content, any nonlinearity will shift some of the ultrasonic content down into the audible range as uncontrolled IMD. The result is a slight, but audible effect. Is this what Amir is hearing? I don't know, but I'd like to. It wouldn't be the first time a very low-level but audible distortion was perceived as detail.
Above: Illustration of distortion products resulting from intermodulation of a 30kHz and a 33kHz tone in a theoretical amplifier with a nonvarying total harmonic distortion (THD) of about .09%. Distortion products appear throughout the spectrum, including at frequencies lower than either tone.
Inaudible ultrasonics contribute to intermodulation distortion in the audible range (light blue area). Systems not designed to reproduce ultrasonics typically have much higher levels of distortion above 20kHz, further contributing to intermodulation. Widening a design's frequency range to account for ultrasonics requires compromises that decrease noise and distortion performance within the audible spectrum. Either way, unneccessary reproduction of ultrasonic content diminishes performance.
--xiph.ort
Tim
Sure we all have a tendency to ignore evidence that doesn't suit our pre-conceived conclusions.
What was, & still is interesting about the thread I started on another forum "sorting out evidence-based Vs faith-based" is the amount of denial that is evident by the very same posters who were the first to jump on any subjective reports with demands for blind listening tests. They prove they are faith-based by their very posts & they now find themselves between a rock & a hard place.
Will their faith get them through........... join us for the next episode of "when pet (theories) go bad & bite you on the bum"
What was, & still is interesting about the thread I started on another forum "sorting out evidence-based Vs faith-based" is the amount of denial that is evident by the very same posters who were the first to jump on any subjective reports with demands for blind listening tests. They prove they are faith-based by their very posts & they now find themselves between a rock & a hard place.
Will their faith get them through........... join us for the next episode of "when pet (theories) go bad & bite you on the bum"
Yes, Tim but that is somewhat of a distraction unless you are saying that playing music files on Amir's system with content > 20KHz, will cause IMD?
It's a possibility that his system & all the others that have return positive ABX results on both the "jangling Keys" files & the music files from Scott (& HyrogenAudio) are suffering from IMD. I can think of one or two possibilities how this could be the case but not willing to make my theory public until I've amassed some more data.
Wherever the evidence leads us & whatever the conclusions, I'm neutral on the outcome but find it all very interesting
It's a possibility that his system & all the others that have return positive ABX results on both the "jangling Keys" files & the music files from Scott (& HyrogenAudio) are suffering from IMD. I can think of one or two possibilities how this could be the case but not willing to make my theory public until I've amassed some more data.
Wherever the evidence leads us & whatever the conclusions, I'm neutral on the outcome but find it all very interesting
No, I am not going by any delay and such. I listen to a complete segment, and then the other segment. I am not doing instantaneous switching as the whole track plays. So any delay does not play a role in my findings.
I should say that I have done no mechanical analysis of the files prior to listening to them. And even now, have not done much investigation.
John, while I have you here, can you please post your listening results here that you wrote on AVS? Like to have this much shorter thread have such data in it too. THanks.
Tim,
you miss the point it is not about the IMD in ultrasonic region BUT about the attenuation of that back in the sub 20khz.
Read my post again and you will see that I mention it will be at LEAST 40db down and probably closer to 60db down, so music that is usually -20dbfs may have IM if the system is stressed that is at around -80dbfs.
You need to appreciate some of these test are at 0dbfs or close to this for ultrasonic signal and that is just insane and wrong in real world.
In reality those signals Xiph shows at 30khz would be around -60dbfs and -80dbfs for good hirez recording, well there would not be IMD there unless pushing the product into being non-linear; but we are talking about dac-headphone setup anyway and less able to stress the setup.
Read the link Jkenny provided earlier and I discussed with him, again you see clear examples of what I am talking about.
Jangling keys or using ultrasonic tones at very high dbfs are an abnormality and tbh why the heck Arny chose keys for this test is beyond me, ah so he then can go on about IMD causing cues and invalidating results; however it is shown by Amir/JA/the article myself and JKeny discusses that this still cannot be that important in the scheme of the other variables for valid reasons I outlined earlier and the fact IMD is not an issue for good products these days unless you deliberately push it beyond operating spec/stress test.
Furthermore stressing a system is not part of ABX, and from what we know Amir did not play the files extremely loud.
EDIT:
OK I also notice Xiph mentions in his chart "in a theoretical amplifier", so his values are not necessarily real world, well definitely not real world for DACs/headphone product.
Cheers
Orb
Last edited:
Hot and distracted by TdF cycling, great race today
I should say this as my 1st sentence is a bit wrong in last post.
With music instruments ultrasonic content is usually 40db below the level of the fundamental, and for recordings analysed by HiFiNews it is usually 50db to 60db down.
SO the ultrasonic content is at around -60 to -70dbfs, then IMD created back in the sub 20khz range will be at LEAST a further 60+db down, meaning it will be at least -120dbfs at worst....
But that is for music rather than jangling keys.
Anyway here is the IMD measurement for Asus Xonar Essence One DAC/headphone product 19+20khz tone at 0dbfs.
Notice the level of IMD tones are waaaay down.
So if jangling keys is even at -20db for ultrasonic content, the IM distortion should be around -120db or lower.
I appreciate would be great if Amir/JA could do the IM test measurement using tones around 30khz mark to see if it differs greatly from 19+20khz for a couple of products; integrated headphone/DAC product being definitely one as it fits more in line with ABX testing.
Jkeny, when ready look forward to your thoughts on what may be causing audible IM as I appreciate I may be missing something here myself, but these measurements suggest IM is actually less of an issue than stopband rejection and impulse response behaviour of filters and sampling rate, let alone the potential transparency of the downsample process employed for bit depth and sampling rate done by the person - IMO of course
Cheers
Orb
I should say this as my 1st sentence is a bit wrong in last post.
With music instruments ultrasonic content is usually 40db below the level of the fundamental, and for recordings analysed by HiFiNews it is usually 50db to 60db down.
SO the ultrasonic content is at around -60 to -70dbfs, then IMD created back in the sub 20khz range will be at LEAST a further 60+db down, meaning it will be at least -120dbfs at worst....
But that is for music rather than jangling keys.
Anyway here is the IMD measurement for Asus Xonar Essence One DAC/headphone product 19+20khz tone at 0dbfs.
Notice the level of IMD tones are waaaay down.
So if jangling keys is even at -20db for ultrasonic content, the IM distortion should be around -120db or lower.
I appreciate would be great if Amir/JA could do the IM test measurement using tones around 30khz mark to see if it differs greatly from 19+20khz for a couple of products; integrated headphone/DAC product being definitely one as it fits more in line with ABX testing.
Jkeny, when ready look forward to your thoughts on what may be causing audible IM as I appreciate I may be missing something here myself, but these measurements suggest IM is actually less of an issue than stopband rejection and impulse response behaviour of filters and sampling rate, let alone the potential transparency of the downsample process employed for bit depth and sampling rate done by the person - IMO of course
Cheers
Orb
I'm neutral too, John. I don't hear it, so it makes neutrality pretty easy for me. xiph.org is saying that playing files with supersonic content - on Amir's system, or any system, will cause very low level IMD. I'm just saying I'd like to know what the difference is, and acknowledging this possibility.
Orb - 40dB down from zero? Looks like more than that. Inaudible to most of the people, most of the time. Of course you can say that of whatever is the source of these differences. That doesn't mean it's not IMD. Besides, audiophiles hear all kinds of things that are masked, all the time. Just get in the right thread with the right challenge to a conventional wisdom and they'll tell you all about it.
Of course here, we're confirming a conventional wisdom, so masking could very easily eliminate IMD as a cause.Tim
Hot and distracted by TdF cycling, great race today
I should say this as my 1st sentence is a bit wrong in last post.
With music instruments ultrasonic content is usually 40db below the level of the fundamental, and for recordings analysed by HiFiNews it is usually 50db to 60db down.
SO the ultrasonic content is at around -60 to -70dbfs, then IMD created back in the sub 20khz range will be at LEAST a further 60+db down, meaning it will be at least -120dbfs at worst....
But that is for music rather than jangling keys.
Anyway here is the IMD measurement for Asus Xonar Essence One DAC/headphone product 19+20khz tone at 0dbfs.
View attachment 16615
Notice the level of IMD tones are waaaay down.
So if jangling keys is even at -20db for ultrasonic content, the IM distortion should be around -120db or lower.
I appreciate would be great if Amir/JA could do the IM test measurement using tones around 30khz mark to see if it differs greatly from 19+20khz for a couple of products; integrated headphone/DAC product being definitely one as it fits more in line with ABX testing.
Jkeny, when ready look forward to your thoughts on what may be causing audible IM as I appreciate I may be missing something here myself, but these measurements suggest IM is actually less of an issue than stopband rejection and impulse response behaviour of filters and sampling rate, let alone the potential transparency of the downsample process employed for bit depth and sampling rate done by the person - IMO of course
Cheers
Orb
IF you're right, that's way down there, Orb. Here's the article: http://people.xiph.org/~xiphmont/demo/neil-young.html. Again, whatever it is, I can't hear it. A few good masters would do my listening pleasure a lot more good than a hard drive full of hi-res files. A new, quality re-mastering movement would transform it. I hope, for Amir's sake, that when he's just listening to music for pleasure, he doesn't hear this stuff either. That would be cursed ears, not golden ones.
Tim
Carrying on thinking this through with regards to my thoughts on filters and stopband rejection, it err...umm.....
Ironically these may be able to influence IM due to aliasing
But this does require a weaker stopband rejection (so requires good filter design when not going with "traditional" steep brickwall), and would be applicable to both normal and hirez music formats so again still no identifiable cues in Amir's ABX.
Removing NOS DACs from the discussion as these are normally CD orientated.
There are several products I can use that show this up as an anomaly (I should note that these are probably corrected now), but also need to stress there are other products that manage to do slow rolloff/some form of minimum phase design without these extreme behaviour.
By default most products will default to a steep filter with strong rejection.
Anyway 1st example showing how IM can be influenced by filter selection/stopband rejection:
Fig.9 dCS Scarlatti, balanced output, Filter 1, HF intermodulation spectrum, 19+20kHz at 0dBFS peak into 100k ohms:
Fig.12 The Filter setting had no effect on the results of the HF intermodulation spectral analysis when the data were upsampled to DSD or 176.4kHz PCM. However, whereas F1 offered textbook performance with non-oversampled 24-bit data (fig.9), the other three filters introduced major spectral changes. Fig.12, for example, shows the behavior with F4, which, with its slow rolloff, gives the lowest rejection of ultrasonic image energy and the greatest degree of "leakage" into the audioband. Many aliasing products can be seen below 20kHz; while this test is very much a worst-case situation, and unlikely to be encountered with music, it is instructive to see the tradeoff this filter requires for better time-domain performance. F3 is better in this respect than F4 and F2 is better still, though neither approaches the spectral purity offered by F1.
Bolded to ensure it is put into context.
A clearer example of what is happening can be seen if I use Ayre's QB9.
Fig.12 Ayre Acoustics QB-9 set to Measure, HF intermodulation spectrum, 19+20kHz at 0dBFS peak into 600 ohms, 24-bit data (left channel blue, right red; linear frequency scale)
Set to Measure, Intermodulation products from the 19 and 20kHz tones are all below –80dB (0.01%), and while the reconstruction filter lets through a couple of images, at 24.1kHz and 25.1kHz (44.1 minus 20kHz and 44.1 minus 19kHz, respectively), these lie at or below –92dB.
Fig.13 Ayre Acoustics QB-9 set to Listen, HF intermodulation spectrum, 19+20kHz at 0dBFS peak into 600 ohms, 24-bit data (left channel blue, right red; linear frequency scale).
Set to Listen (fig.13). While actual intermodulation products are still very low in level, the less steep rolloff that goes with the optimized time-domain performance allows many more aliased images to "leak" into the audioband. The primary images, at 24.1 and 25.1kHz, now lie just 9dB below the original tones, and other images are visible in the audioband itself. Fortunately, these all lie close to –100dB or below,
So as I thought stopband rejection design can potentially influence performance but did not think it through fully to being Alias-->IM, but again it needs to be emphasised these test fundamental signals are 0dbfs and not at real world -60dbfs or lower, and most products as shown with the Asus utilise more traditional filters so would not have this effect.
So IM may be able to contribute to cues in context of filter/stopband rejection, but it would require stopband alias rejection to be weak and/or poor design of DAC filter, along with the fact the ultrasonic signal would need to be extraordinarily high in comparison to the real music sub 20khz content (which according to Griesinger needs to be removed for consistent audibility).
Edit:
That said the DCs would be the same for both hirez and normal music in terms of IM so no cues, BUT it might help to create cues if comparing filters in the DAC (again same caveat as before).
So further thinking, the topic of this post would not create cues for identifying hirez vs CD quality in ABX as it is pretty clear the effect will happen on both formats, cues may happen when trying to identify filters though.
Please note these measurements are from Stereophile and done by JA.
Cheers
Orb
Ironically these may be able to influence IM due to aliasing
But this does require a weaker stopband rejection (so requires good filter design when not going with "traditional" steep brickwall), and would be applicable to both normal and hirez music formats so again still no identifiable cues in Amir's ABX.
Removing NOS DACs from the discussion as these are normally CD orientated.
There are several products I can use that show this up as an anomaly (I should note that these are probably corrected now), but also need to stress there are other products that manage to do slow rolloff/some form of minimum phase design without these extreme behaviour.
By default most products will default to a steep filter with strong rejection.
Anyway 1st example showing how IM can be influenced by filter selection/stopband rejection:
Fig.9 dCS Scarlatti, balanced output, Filter 1, HF intermodulation spectrum, 19+20kHz at 0dBFS peak into 100k ohms:
Fig.12 The Filter setting had no effect on the results of the HF intermodulation spectral analysis when the data were upsampled to DSD or 176.4kHz PCM. However, whereas F1 offered textbook performance with non-oversampled 24-bit data (fig.9), the other three filters introduced major spectral changes. Fig.12, for example, shows the behavior with F4, which, with its slow rolloff, gives the lowest rejection of ultrasonic image energy and the greatest degree of "leakage" into the audioband. Many aliasing products can be seen below 20kHz; while this test is very much a worst-case situation, and unlikely to be encountered with music, it is instructive to see the tradeoff this filter requires for better time-domain performance. F3 is better in this respect than F4 and F2 is better still, though neither approaches the spectral purity offered by F1.
Bolded to ensure it is put into context.
A clearer example of what is happening can be seen if I use Ayre's QB9.
Fig.12 Ayre Acoustics QB-9 set to Measure, HF intermodulation spectrum, 19+20kHz at 0dBFS peak into 600 ohms, 24-bit data (left channel blue, right red; linear frequency scale)
Set to Measure, Intermodulation products from the 19 and 20kHz tones are all below –80dB (0.01%), and while the reconstruction filter lets through a couple of images, at 24.1kHz and 25.1kHz (44.1 minus 20kHz and 44.1 minus 19kHz, respectively), these lie at or below –92dB.
Fig.13 Ayre Acoustics QB-9 set to Listen, HF intermodulation spectrum, 19+20kHz at 0dBFS peak into 600 ohms, 24-bit data (left channel blue, right red; linear frequency scale).
Set to Listen (fig.13). While actual intermodulation products are still very low in level, the less steep rolloff that goes with the optimized time-domain performance allows many more aliased images to "leak" into the audioband. The primary images, at 24.1 and 25.1kHz, now lie just 9dB below the original tones, and other images are visible in the audioband itself. Fortunately, these all lie close to –100dB or below,
So as I thought stopband rejection design can potentially influence performance but did not think it through fully to being Alias-->IM, but again it needs to be emphasised these test fundamental signals are 0dbfs and not at real world -60dbfs or lower, and most products as shown with the Asus utilise more traditional filters so would not have this effect.
So IM may be able to contribute to cues in context of filter/stopband rejection, but it would require stopband alias rejection to be weak and/or poor design of DAC filter, along with the fact the ultrasonic signal would need to be extraordinarily high in comparison to the real music sub 20khz content (which according to Griesinger needs to be removed for consistent audibility).
Edit:
That said the DCs would be the same for both hirez and normal music in terms of IM so no cues, BUT it might help to create cues if comparing filters in the DAC (again same caveat as before).
So further thinking, the topic of this post would not create cues for identifying hirez vs CD quality in ABX as it is pretty clear the effect will happen on both formats, cues may happen when trying to identify filters though.
Please note these measurements are from Stereophile and done by JA.
Cheers
Orb
Last edited:
Similar threads
