John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
Jump to Latest
#63,121
billshurv said:
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.
Click to expand...

Yes. sorry for the bad writting. BenchMark uses noise shaping of some type applied to the dither (noise).

How few bits are needed to get a nice waveform is more to the point. And, that the levels dealt with is much lower with 24b files.

I'm looking for low distortion down to very low signal levels. Music files that stay at 24b (HiRes) can do that better partially because dithering would be done at a lower level.


THx-RNMarsh
 
Last edited:
#63,122
RNMarsh said:
is going to substantially IMO have different audible side effects. It is the CD which sound inferiour to the 24b files.

THx-RNMarsh
Click to expand...

There are nothing but good "side effects" reducing quantization error into random noise totally uncorrelated with the signal. The audibility is your opinion, I can't hear 1 LSB of noise from 16bit audio when the volume is set for normal listening.
 
#63,126
Tom Danley said:
In addition, the very best I ever heard from any I was involved with, was at best on par with an old transistor radio.
Click to expand...


Tom, the audio signal that modulates the ultrasonic carrier needs some serious preconditioning, the modulation index needs attention, the modulation method SSB or DSB is important as it demands different signal preconditioning, ect.
The HSS paper is very informative (all the referenced papers are available on the internet)


jneutron said:
I didn't spot that, where is it?
Click to expand...

Jneutron have a look here:

http://www.hse.gov.uk/research/crr_pdf/2001/crr01343.pdf

My tinitus is very sensitive to ultrasounds. A small fraction of any from the above limits makes my ears ring loudly at 3-4 different pitches for hours.

George
 
#63,127
gpapag said:
Tom, the audio signal that modulates the ultrasonic carrier needs some serious preconditioning, the modulation index needs attention, the modulation method SSB or DSB is important as it demands different signal preconditioning, ect.
The HSS paper is very informative (all the referenced papers are available on the internet)




Jneutron have a look here:

http://www.hse.gov.uk/research/crr_pdf/2001/crr01343.pdf

My tinitus is very sensitive to ultrasounds. A small fraction of any from the above limits makes my ears ring loudly at 3-4 different pitches for hours.

George
Click to expand...

Thank you George, very informative as usual..
doc page 15, pdf 22. I hilite relevant info
When considering hearing damage by noise, the concept of intended protection is quite sophisticated, requiring knowledge of the noise dose (level and duration) required to produce a hearing damage response over the range of susceptible individuals. With knowledge of how all of the various factors interact, one may predict what proportion of an exposed population would suffer a specified degree of hearing loss from a known exposure. For conventional broadband occupational noise as might be experienced in any number of workplaces, the idea of intended protection may be applied, as long as one is prepared to grapple with the troublesome social concept of the “low fence” which defines the boundary between acceptable and unacceptable noise-induced hearing loss. As far as sound of very high frequency or ultrasonic frequency is concerned, the dose-response relation is unknown: the limiting levels have been deliberately set low to avoid any response whatever.
Click to expand...
Doc page 22, pdf page 28
3.2 HEARING LOSS, EITHER TEMPORARY OR PERMANENT
It has long been recognised that high levels of wideband or “ordinary” noise, after sufficient duration, will damage the hearing of exposed individuals. In the occupational sense, the exposure will probably last several hours each workday over a period of years. The principal characteristics of occupational noise-induced hearing loss are:
· It is always sensorineural, affecting the hair cells in the inner ear.
· It is bilateral; the audiometric pattern is similar in each ear.
· The earliest damage is manifest as a threshold shift (a loss of hearing sensitivity) at the audiometric frequencies 3, 4 and/or 6 kHz.22
· During stable noise exposure conditions, the hearing losses at 3, 4 and/or 6 kHz will grow quickly over the first few years, and then develop more slowly to reach a maximum level after about 10-15 years. The losses are not expected to exceed 70 dB HL.
· Hearing loss in the lower frequencies takes much longer to develop.
· Once the noise stops, the noise component of the hearing deficit stops growing. Hearing loss due to natural ageing will, of course, continue.
In respect of hearing damage by very high frequency or ultrasonic noise however, there is no such received wisdom. The following sections will review what information is available, while bearing in mind the nature of the hazardous noise and the nature of the hearing deficit
Click to expand...

Bottom line, they don't know.

It is interesting that the USAF had 85db spl as the limit early on, page 5 doc, page 12 pdf. Clearly, just set without information.

jn
 
#63,128
scott wurcer said:
There are nothing but good "side effects" reducing quantization error into random noise totally uncorrelated with the signal. The audibility is your opinion, I can't hear 1 LSB of noise from 16bit audio when the volume is set for normal listening.
Click to expand...

Thank you for your patience. I might not be able to either.... if that is all there is to the noise. I just get better over-all results with 24b than 16b. Thats fortunate as CD's have been and are on the way out.

In your opinion, where are the differences which would be audible between playing CD and 24b downloads? Everyone keeps saying 16b is just wonderful and all anyone would/will ever need and no need to change.... But it is changing away from CD and measures and sounds better in many ways on just the playback side. Is it in the elimination of the producing/playback of CD's? That part is missing entirely with downloads.

AES 3269, AES 3394 and AES 8522 are about methods for estimating detectablity and threshold. 16b doesnt always cut it.


THx-RNMarsh
 
Last edited:
#63,129
Anyone know for sure... who makes the most linear and lowest measured distortion 24b DAC 'black box' product down to and below -100dB?

DAC non-lin.jpg IC only.


THx-RNMarsh
 
Last edited:
#63,130
RNMarsh said:
AES 3269, AES 3394 and AES 8522 are about methods for estimating detectablity and threshold. 16b doesnt always cut it.

THx-RNMarsh
Click to expand...

We should just end this here, I guess the studies where no one could tell the difference are just flawed or fraudulent. I thought the one posted last week, but where the files were no longer available was particularly well conducted, as you say a null result across a large sample of listeners including musicians, well healed audiophiles, etc.

I still listen to LP's too.
 
#63,135
scott wurcer said:
We should just end this here, I guess the studies where no one could tell the difference are just flawed or fraudulent. I thought the one posted last week, but where the files were no longer available was particularly well conducted, as you say a null result across a large sample of listeners including musicians, well healed audiophiles, etc.

I still listen to LP's too.
Click to expand...

Alright. That is fine. I think it is an accumulation of many small things all together..... no one thing.... my CD player might not be as perfect as tests and studies or designs werent implemented all that well and did not come close enough to theory or best practices. CD player was just a MSRP of $300.

Thanks anyway. I still listen to CD's too ....

RNMarsh
 
Last edited:
#63,137
Could be that too. But getting only theory for answeres --- which dont always work so perfectly in practice.... .

There needs to be some acknowledgment to that as well.

Listening tests.... I have done them in many locals and I dont always fall in with the majority... including the one I did awhile ago at DIYAudio (5-6 months ago?).... so listening tests results of others is, in this case, not too helpful. But, it could be used as Plan B fall-back 🙂.

I saw no need for dither on 24b as I showed at -90. This morning, in an email, I asked Bob Speer of CD Mastering Services "--- dither is used only when reducing the word length. .... for 24/96 HD dither is not used" for clients such as HDTracks and others for hiRes downloads.

I'm going to go do some real world comparison tests to get to the bottom of my question(s).


THx-RNMarsh
 
Last edited:
#63,140
I don't know why anyone would want to use a physical CD and player except for transferring the data over to a more convenient medium. So it would be nice to see the term "16/44.1" used instead of "CD" because no one's talking about physical CDs, but rather 16/44.1 PCM.

The constant and intentional use of the term "CD" seems to be nothing more than a propaganda device.

se
 
Status
Not open for further replies.
Top Bottom