I have no substantial arguments with MP3 - decoded and upsampled prior to listening, they demonstrate perfectly adequate levels of information are available, for satisfying playback.I wonder ... I'm listening to 256K mp3's I personally converted from my
1999 CD collection (my transition - last purchased media) .
They still are really quite good (my flacs are a little better). Original L3 codec from
germany was quite advanced for the 20th century.
But nothing has really changed ...besides sound card 16>24 bit dac's and win98
to 7 . No real "groundbreaking" improvements - AT ALL .
An amusing memory is that my first exposure to MP3 was via a burnt CD - I was given this to play, just for the experience of the particular music. Good modern material, I had no complaints - only later was I made aware that the sources were MP3, that had been turned into WAV files to create a playable CD. Now knowing this, they were still perfectly fine to listen to - what's all this whinging about then, I thought to myself ...
I'm really talking about the chasing of meaningless numbers of bits, levels of THD, power outputs of amplifiers, numbers of speaker drivers, of enormous sizes, etc, etc ...
My point exactly , most thoughtfully mastered music was pre - 24bit (pre-20th century) ,
listened through various Japanese consumer products.
Even as I (try) to have better than average equipment -
there is a point of "diminishing returns". I stand by the previous "mine is longer" assessment .
Edit - a very high power / low THD amp is a major step towards SQ.
OS
Last edited:
One problem with that is that the real output of the dac will not be as clean as that due to johnson noise in the converter and opamps.
Or a 21 year old who has a 'start up' with a couple of his mates, cos like facebook did it. I had an undergrad assigned to me who proudly announced he was a CTO and wondered why I rolled around on the floor.
Even worse is some of the CTOs that I come across in larger companies that know NO technology at all. So on balance scott you get the best of the CTOs in your field!
Almost poetry. A bit of work and you could manage a haiku. A nice start to the day to find a post that flows.
Are you confusing dither and noise shaping? Dither is generally added to the recording. Noise shaping to the playback device.
What I don't understand is that the -90dB graphs are only ever shown for 16 bit vs 24 bit, not 17, 18, 19, 20 to see how few bits you need to get a nice sine wave for those who think it matters.
Nicely linear DAC in those attachments anyway.
The biggest problem is acoustic impedances (transducer-air) mismatch (*)
For quartz/air (2π) interface the transmition coefficient is less than 0.00001.
This means that less than 0.001% of piezo vibrations are turned into air vibration right at the transducer/air interface. This percentage increases a bit (less than a decimal point) due to local compression of air at ultr freq.
Then add air travel attenuation
http://www.ktu.lt/ultra/journal/pdf_50_1/50-2004-Vol.1_09-A.Vladisauskas.pdf
And then, coupling piezo elements to an amplifier is not that straightforward
http://dspace.mit.edu/bitstream/handle/1721.1/33240/06370831.pdf
George
(*) This problem is to be met with acoustic frequencies transmition with ordinary speaker units too.
If you have noticed at the Audio Pixels LTD patent, they deal with it by having each element compressing the air in a tiny cavity, thus air density increases, acoustic impedance mismatch decreases, transmition efficiency goes up.
George
Last edited:
Of course, and why should-we forgo this.
But about 16 bits, at the end of many tunes, there is a fade out. Did-you notice something strange, like increasing distortion, or the music is just slowly going away ?
Yes the signal/HD ration increase. But the HD absolute level remain constant, and under audibility threshold.
It is so easy to prove to oneself that 16 bits is perfectly adequate, simply by listening to - or attempting to, because it is almost inaudible, - a conventional track that has been very heavily attenuated in the digital domain, without revving up the analogue gain to silly amounts. The only conclusion one can draw about why CD sound is often poor, is that the particular playback system implementation is faulty - and careful, comparative listening will reveal that this is what is going on ...
- a conventional track that has been very heavily attenuated in the digital domain, without revving up the analogue gain to silly amounts. The only conclusion one can draw about why CD sound is often poor, is that the particular playback system implementation is faulty - and careful, comparative listening will reveal that this is what is going on ...There is an SMPS in there (or even more
) so it has to be noisy by definition....Of course bespoke audiophile streamers with processors and memory wont be subject to noise such as SSN etc.......
Oh and the use of ferrite beads is the work of the devil...
Just think of ultrasound cavitation of fat, blood, other liquids inside the body
I'm thinking hearing damage.
When I first approached safety for the acceptable limits of 25 Khz U/S exposure, the safety guys found that there was NO standard. Pulling out the ol' radio shack spl meter (or whatever they were using) was useless as the meter was not capable of reading the frequency.
So, no standard of exposure limits, and no way of measuring it. Just ducky.
The shorter wavelengths are capable of delivering far more energy to the inner ear. Even the US navy understood that back when I first read it, sometime in the late 60's (dad was a boilerman on a destroyer). There, they were concerned with the release of high pressure steam with the resultant high frequency rich sound blast.
I've not researched this since mid-90's, so don't know if OSHA actually addressed the issue.
jn
Hearing impairment in US service personal is a big problem that is being worked on in many areas, for land based forces the new AN-VIC 5 and noise cancelling headphones from the likes or Racal are areas where research and design has been going on. Again headphones and hearing protection for airborne (pilots etc) is being dealt with. I would imagine the same is going on for Navy personnel...
Some possible sources of information...Sources: 1“VA Service and Compensation
Costs:
Supporting Readiness For Life” by the National Center for Rehabilitative Auditory
Research (NCRAR) and the
American Tinnitus Association;
2Protecting Crew Members Against Military Vehicle Noise” presented by NATO and OTAN
at the RTO AVT Symposium on
“Habitability of Combat and Transport Vehicles: Noise, Vibration and Motion,” Prague, Czech
Republic, October 2004, and
published in RTO-MP-AVT-110.
Mainly lower frequency stuff though...
Some possible sources of information...Sources: 1“VA Service and Compensation
Costs:
Supporting Readiness For Life” by the National Center for Rehabilitative Auditory
Research (NCRAR) and the
American Tinnitus Association;
2Protecting Crew Members Against Military Vehicle Noise” presented by NATO and OTAN
at the RTO AVT Symposium on
“Habitability of Combat and Transport Vehicles: Noise, Vibration and Motion,” Prague, Czech
Republic, October 2004, and
published in RTO-MP-AVT-110.
Mainly lower frequency stuff though...
The dither is added to the source not in the player. Dr. Lipshitz took them over his knee and spanked them all a long time ago. You're an AES member? you can get all the papers for free. I still consider it an embarrassment to the audio engineering community that this was not all worked out before they cut the ribbon.
I can't imagine any CD produced in decades that is not dithered. The "properly" would refer to the fact that with noise shaping you don't have to lose the whole 3db of SNR. I can not hear the noise floor of 16/44.1 in any remotely normal listening situation.
AFASIK the spotlight did not produce it's effect at the transducer but along the focused ultrasound beam. That is the SPL of the ultrasound all along the beam has to be sufficient to exercise the non-linearity of the air. I was surprised to find that this effect was known in the 60's and used for sonar. To get 90dB SPL of reasonably low distortion audio by the modulation I would think the SPL of the ultrasound is >> than this. I'm not sure we are talking about the same thing exactly. They did try to make audio speakers this way in the 70's but failed miserably.
EDIT - A little quick search on some piezo-transducer companies turned up a number of 120db SPL at 200mW input for a single transducer. So I assume we are talking >140dB SPL in any case. Yes we did "point the thing right at our heads", reminds me of the early radar research.
jn is right there was never even a mention of considering the actual safety of the radiated power at the ultrasonic frequency.
Dimitri - the coupling to the body from free air is very low, medical ultrasound needs lots of impedance matching goop.
Last edited:
Yes that is the crux of the matter! Select your grounding well and there's no noise on it - referred to signal.
Jan
I have a HSS speaker similar to this (two transducer version) in the shed somewhere.
The effect is interesting in that sound appears out of thin air and is similarish to wearing headphones.
There is no localisation to speak of..ie the sound seems to originate directly at the ears, a bit like running open air headphones in mono.
When the box is pointed at a hard surface (wall, window etc) the sound appears to eerily emanate from a relatively small area of the hard surface.
Overall the effect is interesting but not particularly useful, the acoustic level attained is not very high, and the transducers are prone to arcing damage when overdriven...no inbuilt limiting/protection.
hypersonic-sound-technology
wiki/Sound_from_ultrasound
parametric/index_e.php
Last edited:
In my previous post I added to your comment that the needed electric power to drive the transducers has to be high (actually high Voltage with minimal current), whether ultrasonic carrier only or modulated ultrasonic emition and I spotted the three areas where energy loss occurs.
Now you are pointing to the demodulated acoustic signal which is to produce the acoustic effect, correctly predicting that the acoustic power of this demodulated acoustic signal has to be less than the acoustic power of the modulated ultrasonic beam.
Now you are refering to the demodulated acoustic signal which is to produce the acoustic effect, correctly predicting that the modulated ultrasonic beam has to be of a higher power than the demodulated acoustic signal, to which I agree.
And yes, the demodulation doesn’t take place at the ultrasonic transducer/air interface but at a surface where the beam impedes. We had another round over this topic (~Jun 2013)
http://www.atcsd.com/pdf/HSSWHTPAPERRevE.pdf
http://ew3.ee.uec.ac.jp/e/wespac9.pdf
George
Last edited:
You have missed the point. Dither on 16 bits and dither (if used.... software gives the choise) on 24 bits is going to substantially IMO have different audible side effects. It is the CD which sound inferiour to the 24b files. The difference is when do you have to start/apply dithering... at what bit level in terms of dB's below FS? If the dithering starts at a much lower level then that is a good thing. [all on the record side.... I misplaced dither awhile back with wwhat should have been noise shaping. Sorry for the confusion]
THx-RNMarsh
Last edited:
In the mid 90’s I built a number of ultrasonic transducers for Gary Kendall at Northwestern U who was researching that heterodyne system and kibitzed a bit for an article on this the editor of Audio magazine wrote.
To work, one needs high intensity sound, high enough to reach significant non-linearity. Since our levitation transducers could reach that easily >165dB, at first it seemed like a natural except one also needed frequency or amplitude modulation and our levitators were a single frequency device.
For Gary’s testing, I made hydrostatically formed focusing electrostatic transducers which were able to get the level needed at 50Khz and had the bandwidth he needed.
Subjectively, these things were annoying for any period of time and after 15 min, most people reported a mild headache or some kind of head symptom.
In addition, he very best I ever heard from any I was involved with, was at best on par with an old transistor radio. I remember Gary’s student (Joe P?) had plans to carry it on.
edit - I just noticed a mention of Joe P on the top of page 12 who apparently did pursue this further.
How it works; Air follows the gas law and while we think about sound as pressure changes and it is, what it actually is pressure changes are above and below ambient pressure (like a DC bias?).
On the positive half of each wave, the pressure is increased and on the bottom half, decreased. What is also happening is the temperature of the air is also going up and down within each wave AND since the velocity of sound is temperature dependant, the positive half of a high intensity wave gradually advances and the “cold half” is retarded so hat what was a sine wave is now a saw tooth shape.
That saw tooth shape has a pumping effect on the air column, an acoustic rectification that causes flow. This was a problem with acoustic levitation as the induced air flow was blowing on the sample you were trying to melt.
Anyway, to make “sound”, you produce that kind of intensity well above audible frequency and then modulate the level (amplitude modulation) or the carrier frequency (frequency modulation seemed more difficult).
Your input audio modulates the carrier, the air non-linearity rectifies it into an audible sound.
Best,
Tom Danley
Last edited:
- Status
- Not open for further replies.