Exposure to high level ultrasonics on hearing over extended periods…and other things down the chain like amps, tweeters
Don't forget the zero-order hold filtering. Most of the power in music is usually at low frequencies, the images of which get suppressed reasonably well by the notches of the sin(πf/fs)/(πf/fs) shaped response of the zero-order hold. It also causes some 3 dB of treble loss at the higher audio frequencies if the sample rate is low.
On the other hand, there are the imperfections of the reconstruction filter when you do use one. For a non-oversampling DAC with reconstruction filter, that will mostly be its poor phase response, for an oversampling DAC, clipping on intersample overshoots, pre- and post-echoes and quantization artefacts.
Of course you can design a digital interpolation filter such that it doesn't clip on intersample overshoots, has extremely small pre- and post-echoes and that the only quantization artefacts are those of a dithered rounding stage after the whole filter, but who knows what was done in the oversampling filters of the DACs NIXIE62 compared his DAC to?
@NIXIE62 Do you use digital volume control? If so, all I wrote about intersample overshoots doesn't apply unless you play very loud.
Last edited:
Post #1332 of the NOS DAC thread is the demonstration of what a NOS DAC with only zero-order hold filtering would sound like to a cat when playing 44.1 kHz-sample-rate recordings:
You say: "...because it happens only after the ear converts the acoustic signal to a bio-electrical one."
I would have thought that it is during the conversion to a bio-electrical signal that the ultrasonics are lost.
Marcel, what has a cat to do with it? Maybe that cat hears that phenomenon but is the reaction as usual with cats. Indifferent.
I wonder if non filtered NOS DACs are popular only with older population having some HF hearing loss?
Research shows that 10 % of young population can hear above 20 kHz, if sound level is high enough.
https://pubs.aip.org/asa/jasa/artic...Hearing-thresholds-for-pure-tones-above-16kHz
This coincides with my experience. In group of 24 people, two were able to hear powered on ultrasonic alarm on the lab bench.
Research shows that 10 % of young population can hear above 20 kHz, if sound level is high enough.
https://pubs.aip.org/asa/jasa/artic...Hearing-thresholds-for-pure-tones-above-16kHz
This coincides with my experience. In group of 24 people, two were able to hear powered on ultrasonic alarm on the lab bench.
Well then, you have yourself some test subjects to audition NOS vs OS techniques with
😉
I'd expect that even if you were capable of hearing up to 28 kHz, it would get masked by the rest of the music. Especially given the fact that for those higher frequencies to be audible, 110 dBspl was needed (yikes!).
😉
I'd expect that even if you were capable of hearing up to 28 kHz, it would get masked by the rest of the music. Especially given the fact that for those higher frequencies to be audible, 110 dBspl was needed (yikes!).
A whining SMPS with 45 kHz injecting in the signal was easily noticed even at normal listening levels. Probably not the 45 kHz itself but what it caused.
Nope. That was looong time ago and I was one of those two. Now, 16 kHz is out of my reach.
28 kHz and 45 kHz is quite low.
Band limiting non linear device is possible to demodulate the signal to lower frequency.
https://arxiv.org/abs/1708.09537
Band limiting non linear device is possible to demodulate the signal to lower frequency.
https://arxiv.org/abs/1708.09537
What I mean is, somewhere in the process of conversion to a bio-electric signal. I have no objection to your viewing it as occurring during the conversion.
No and yes.
What does "to sound" really mean? Human brain is constatly imagining things, we are hallucinating all day every day from the time we are babies till the day we die. If you do A/B test that is not blind and one dac is cheap, small and ugly and other big, expensive and looks luxurious, the big one is going sound better to large percentage of people due to the placebo auditory hallucination effect.
Realisticly since noisefloor in average listening room is like 20-30db and maximum spl is 90-100db, we only need signal to noise and thd ratio of 60-80db for noise and distortion of dac to be completly inaudible which means 99.9% of dacs will sound same in blind test from the first TDA1541 r2r dacs to modern state of art AKM,Sabre, Cirrus delta sigma dacs.
DACs in current times are semi snakeoil, while there are measurable differences and in extreme cases like driving ultra sensitive IEMs and hearing noisefloor, the differences are below treshold of audibility and the reviewers are either lying to profit of affliate links or are hallucinating themselfs. In 2024 the build in dacs in motherboards, laptops or the 9€ Apple 3.5mm dongle are all audibly flawless.
What does "to sound" really mean? Human brain is constatly imagining things, we are hallucinating all day every day from the time we are babies till the day we die. If you do A/B test that is not blind and one dac is cheap, small and ugly and other big, expensive and looks luxurious, the big one is going sound better to large percentage of people due to the placebo auditory hallucination effect.
Realisticly since noisefloor in average listening room is like 20-30db and maximum spl is 90-100db, we only need signal to noise and thd ratio of 60-80db for noise and distortion of dac to be completly inaudible which means 99.9% of dacs will sound same in blind test from the first TDA1541 r2r dacs to modern state of art AKM,Sabre, Cirrus delta sigma dacs.
DACs in current times are semi snakeoil, while there are measurable differences and in extreme cases like driving ultra sensitive IEMs and hearing noisefloor, the differences are below treshold of audibility and the reviewers are either lying to profit of affliate links or are hallucinating themselfs. In 2024 the build in dacs in motherboards, laptops or the 9€ Apple 3.5mm dongle are all audibly flawless.
I agree. And amps are heading the same way at modest listening levels. Class D amps nowadays can measure extremely well for not much money. This means high-detail hifi listening as cheaper than it ever was. Loudspeakers and room acoustics are the biggest factors, with room acoustics being the hardest for most people to improve as most of us have limited living space and have to squeeze our hifi in somewhere.
I agree, that there is no inherent reason why NOS should sound better than OS. It seems like they should barely sound different at all. An exploration of why they do sound different in practice, however, was the purpose of that 2021 NOS thread which Marcel linked. What was indicated (not scientifically proven) is that the primary reason for the subjective character difference between OS and NOS is due to the, somehow, insufficient implementation of typical, DAC co-resident, silicon resourced digital interpolation FIR filters. I say, somehow, as Marcel had designed a listening test to explore what appeared to be the most likely candidate cause (per, Lagadec), but that test was hampered by test participant/set-up control difficulties.
By the conclusion of that NOS thread, what I personally found most interesting was that NOS and sophisticated PC s/w interpolation, sounded much more alike than either of them did to typical, DAC silicon co-resident, half-band, equirpple FIR interpolation filters. In other words, typical, DAC co-resident silicon FIR filters were implicated as the culprit in producing the subjective difference between OS and NOS, and not FIR filtering itself. Which, presumably, is due to their utilization of silicon processing architecture designed to conserve chip area. That’s only conjecture, though. So, for many of us DAC hobbyists, bypassing the DAC’s on-chip FIR filter results in a subjective character near that of a sophisticated s/w interpolator, and superior the DAC’s co-resident digital filter.
Last edited:
The reason I am a bit pedantic here is because I think that the frequency limiting parts of the hearing are the parts that move because they have inertia, not the actual conversion to electrical signals.
Same here, though my hearing nowadays runs out at somewhere between 10 kHz and 12 kHz, and according to a test I did is typical for my age. I guess all those moving parts inside the ears are also getting a bit stiffer with age.Nope. That was looong time ago and I was one of those two. Now, 16 kHz is out of my reach.
When I was young I coud hear all the way up to the 20 kHz test tone on my 99 track Denon test CD. So it stands to reason the curve had only started to roll off and extended a bit further up the spectrum.
My interpretation of your question was whether people able to hear well above 20 kHz would dislike NOS. rs Perhaps there are some youngsters here who could chime in?
In reality, the differences in low pass filters usually result in the most noticeable variations in sound, but that doesn’t really show up in common measurement values.
- Status
- Not open for further replies.