[AndreYew]

Quote:

Originally posted by Nights_85 would'nt u need quite a bit of energy at that frequency to actually heat up a tweeter? I mean most of the energy in an audio signal (trust me, I doubt anyone would want their tweeter running at say 60watts)goes to the lower frequencies.

As Joe Blogg's quote of the Stereophile graph shows, SACD noise is independent of recorded material: SACD always has high-frequency noise even if you record stuff that has nothing above 100 Hz.

--Andre

 

[Wodgy]

Shouldn't SACD players have an analog low-pass output filter to clean up some of that noise? It's pretty extreme.

 

[Nights_85]

maybe they considered the fact that humans cant here it so they just leave out the extra capacitor. anyway, the less components you have in the circuit, the less distortion

 

[Wodgy]

Yes, but if that noise is causing THD of upstream components to increase (or is likely to do so), there is no net win.

I don't understand why conventional theory tells us that we should be filtering the output of a PCM DAC, but people aren't doing it for DSD,

 

[DesBen]

Psychoacoustic studies have shown that although we cannot directly detect sine waves above 16-20kHz, we can certainly detect their presence in music. One study went further and showed that people noticed an improvements when the system extended to 40kHz instead of 30kHz (ranges might have been different, but you get the idea).

Unfortunately, I cannot recall the authors of the study. I just remember reading the papers as part of my Multimedia Systems course. Maybe someone could provide more details?

 

[DesBen]

High-Frequency Sound Above the Audible Range Affects Brain Activity, AES Preprint No. 3207.

A researcher named Oohashi tested subjects to determine how brain wave activity is affected by ultrasonic frequencies. (High-Frequency Sound Above the Audible Range Affects Brain Activity, AES Preprint No. 3207.) He determined that under blind conditions his subjects were responding to ultrasonic frequencies up to 60 kHz. Oohashi determined that the mechanism for transmission was bone conductivity to a small organ in the inner ear called the saccule, which is wired to the cochlea, the organ responsible for the majority of hearing perception in humans.

see also
http://jn.physiology.org/cgi/content/full/83/6/3548

The experiment "Psychological evaluation of sound quality" is of particular experience to us. And I quote:

A significant difference was evident between FRS and HCS in some elements of sound quality. Subjects felt that FRS was softer, more reverberant, with a better balance of instruments, more comfortable to the ears, and richer in nuance than HCS.

Here, FRS stands for Full-Range Sound and HCS for High-Cut Sound (nothing above 20kHz). Wouldn't you agree that the same has been said about SACD Vs. redbook ?

 

[markl]

IIRC, many SACD players have a filter that elimiates all sound above 50KHz, despite SACDs claimed frequency extension much beyond this. This is actually an irritant to many afficiandas of SACD as they'd rather just have it all, potential risk or no.

 

[Wodgy]

Thanks Markl, that sounds more reasonable.

DesBen, I wasn't wondering about a strict 20 kHz filter (that would just be throwing away information). However, a LPF cut somewhere around 50-80 kHz would be appropriate. Clearly the 100 kHz peak in the graph Joe posted is garbage unrelated to the music.

 

[Alex Altorfer]

Andre,

Quote:

Originally posted by AndreYew ...it's well-known for several years now that certain Bitstream or delta-sigma based CD players caused issues with their ultrasonic noise. The Rega Planet was notorious for this.

Which Rega Planet was notorious for this? The older version or the Planet 2000?

Cheers,
Alex

 

[Joe Bloggs]

Wodgy,

If you take a look at the numbers, the peak of the noise is less than -40dB down, and that is not very significant compared to the energy at lower frequencies when usual music content is being played back (up to 0dB)

Anyway, that chart is for the already filtered output... this player at least doesn't appear to filter until over 100kHz... oh and as theaudiohobby noted in another thread, the spectrum is measured in 1/3 octaves, which measures 'increasing' amounts of noise even in a white noise spectrum, so that needs to be taken into account (apparently it's why the trace for LPCM is also sloping upwards)

 

[AndreYew]

Why doesn't SACD have a filter for higher frequencies? They do (for the reconstruction filter, as well as the extra filter on many players as markl mentions), but placing it too low would kind of go against the SACD mantra (true or otherwise) that low, sharp cutoff filters sound bad. Also, these filters are relatively shallow, so they still allow substantial amounts of energy in.

--Andre

 

[Music Fanatic]

Quote:

Originally posted by DesBen Psychoacoustic studies have shown that although we cannot directly detect sine waves above 16-20kHz, we can certainly detect their presence in music. One study went further and showed that people noticed an improvements when the system extended to 40kHz instead of 30kHz (ranges might have been different, but you get the idea). Unfortunately, I cannot recall the authors of the study. I just remember reading the papers as part of my Multimedia Systems course. Maybe someone could provide more details?

This is a famously disputed point in audio science. There have been studies on both sides.

 

[markl]

I wonder how many recordings (master tapes, not LPs or CDs) actually have any information above 20KHz to be reproduced on SACD anyway? Do most microphones, recording devices, analog tape itself have much ability to record much information above 20KHz? If people are recording digitally, do their digital recording devices contain filters that block all sound above 20KHz?

 

[WmAx]

Quote:

Here, FRS stands for Full-Range Sound and HCS for High-Cut Sound (nothing above 20kHz). Wouldn't you agree that the same has been said about SACD Vs. redbook

I, too, have paid attention to this research. I have a few problems with this:

(1) IMD(intermodular distortion) of the system at it's two settings was not measured/published, and correlated to a known threshold for human hearing tolerances(requring reference too or a new study in itself).(A BIG ERROR). Mention of IMD was made one time in the paper, and no specific data was provided. They also slightly mis-state that it should not be a problem in this test. To be specific, they quoted:

"To overcome this problem, we developed a bi-channel sound presentation system that enabled us to present the audible LFCs and the nonaudible HFCs either separately or simultaneously."

Problem is, that IMD is an artifact that will be caused by mixing the signals at line level OR at the acoustic level. It escapes me why they seem to think this is not the case. If one worries that it may be in the 'original' signal, well, a precision 44.1/16 recording/playback system will record ALL of the audible fundamentals of this distortion IF it was originall present. The form of IMD here, the problem is, is likely to be a result of the playback system itself. From what I can gather, they basicly had a supertweeter they engaged/disengaged. Another noteworthy point is that that could not induce these 'results' with the high frequency content alone. They state :

"Moreover, it is the combined presentation of HFCs and LFCs, not HFCs alone, that specifically induces the enhancement of alpha-EEG and activation in the deep-lying structures"

Further concern and investigation should be given to IMD content of the system, especially after this result.

A pertient fact omitted here when mentioning the 1978 study, is that the test subjects were all audio professionals with trained hearing skills, and proper statistial analysis and testing methodology was used. It's held up to scrutiny for over 20 years now. I only mention this becuase they keep citing this old study as if their new test is demonstrating credible different results.

(2) In the case of the 'subjective listening test', a standard accepted statistical/testing format was not used in this study.

Thus, this study is not accepted as a standard. If the researchers want to establish, as a standard, that a bandwidth exceeding RBCD's capability is important for human auditory system, more work is required. Notice that the 1978 study is still the accepted standard, and is cited as such by the JAES.

Personally, I will go with the more careful of the two studies. If they are able to prove that a higher bandwidth is important, fine. They have not accomplished this yet.

-Chris

 

[Music Fanatic]

Quote:

Originally posted by markl I wonder how many recordings (master tapes, not LPs or CDs) actually have any information above 20KHz to be reproduced on SACD anyway? Do most microphones, recording devices, analog tape itself have much ability to record much information above 20KHz? If people are recording digitally, do their digital recording devices contain filters that block all sound above 20KHz?

Yes, absolutely, to record on standard redbook, a brickwall filter is required. This is an important point that is routinely ignored by those who claim that high frequency audio equipment is required to avoid phase distortion (besides the more obvious fact that almost all speakers have far worse phase distortion than any brickwall filter in regular use.)