Hi Bob,
You recommended in your book to make the servo time constant 1s. Is there any reason for not making the time constant longer? A longer time constant would improve low frequency margins.
David.
Hi Franco,
I never heard an engineer say this or imply this; all engineers I know are very well aware that this is NOT so. But I may have missed something.
Otoh, I would assume that the inner ear etc works the same whether you are listening to live sound or to reproduced sound. Therefore, to a first approximation, the workings of the inner ear are irrelevant if you want to design audio equipment that faithfully reproduces live sound.
Nevertheless, if you know how the inner ear works and know about the psychology of hearing, you may be able to cut corners in your equipment design and still get a credible reproduction of sound. An example is high-sample rate (256k or higher) MP3 which has been shown to sound very believable in a controlled test.
jan didden
Jan,
This is one I have to disagree with. I think the environment leading up to the live event cause great shifts not only in perception, but actual physical changes such as tightening of the eardrum, increased adrenaline levels etc.
ES
Out of my depth here with electronics, I'll bow out with this comment (but will enjoy following the thread). Assumptions (compromises?) are always part of the design process and I think Jan's position regarding live sound IS a lofty place to start. Of course the overshadowing consideration regarding the Organ of Corti per se is matching volume levels because as volume increases frequency acuity decreases. To that, I would add my opinion that realism in reproduced sound is as much a result of the sound quality of barely audible signal components as the quality of much louder 'primary voices'. The 'details' have to be good because we sense them more acutely. Hearing with the natural MIX of acuities is what seems realistic for me... I think there are many secondary considerations concerning high frequency processing by the ear sensors but who doesn't like bandwidth?
I see Jacco beat me to the punchline but ... The anticipation of a live event versus a recording is a difficult thing to reproduce - partying with friends (as often done before a live concert) is not always practical!
best regards,
Frank
This is true, but larger is the event itself. When your brain is experiencing
music it is taking in all sorts of sensory input, which is why watching a live
performance or even a video is so different than two-channel in the dark.
Of course, you can supplement the latter with tubes that glow and nicely
illuminated meters on your amplifier.
This is true, but larger is the event itself. When your brain is experiencing
music it is taking in all sorts of sensory input, which is why watching a live
performance or even a video is so different than two-channel in the dark.
Of course, you can supplement the latter with tubes that glow and nicely
illuminated meters on your amplifier.
Of course but I was trying to distinguish the sensory issues from the actual performance of one's ears. They really do behave differently from the preconditioning.
Crack, bang ouch...grmph weak soapbox!
Far too simple, and obvious and [/i]complete[/i] an answer to the 'make it sound better' basis of DiyAudio...
Anyone fancy a hug?
Nice try Ed. Tightening, yes, eardrum, no
jan
OK the Tensor Tympani muscle contracts increasing the tension on the tympanic membrane and reduces transmission.
ES
The muscle actually pulls on the maleus, which in turn tenses the tympanic membrane. The reflexes that run that muscle are merely protective (AFAIK) of extreme volumes, damping tympanic membrane amplitude. It's easy to hear when the reflex is working...
listening to commercially produced reproduced music is more about appreciating a impressionist style artistic rendering - it is never going to "be like the real thing" in many sensory and even audio dimensions
since there are many styles of recorded music rendering it seems that the amplifier's role is to be transparent and convey the all of the recording's artist's, those playing instruments, choosing, placing the mics, running the mixing board "intent" as expressed in a few channels of electrical signal
once the goal is restricted to accurate amplification of an electrical signal the question becomes "how accurate" and in what electrical signal dimensions
psychoacoustics, physical limits of recording chain components does enter in evaluating the required accuracy of amplification
amplifiers should convey the signal components that we interpret as image depth or even "emotion" but without "editorializing"
we don't know the mapping from electrical signal to mental/emotional state but we can still fall back on Signal Theory and build amplifiers that convey as much Shannon-Hartley “information” content as practical
since there are many styles of recorded music rendering it seems that the amplifier's role is to be transparent and convey the all of the recording's artist's, those playing instruments, choosing, placing the mics, running the mixing board "intent" as expressed in a few channels of electrical signal
once the goal is restricted to accurate amplification of an electrical signal the question becomes "how accurate" and in what electrical signal dimensions
psychoacoustics, physical limits of recording chain components does enter in evaluating the required accuracy of amplification
amplifiers should convey the signal components that we interpret as image depth or even "emotion" but without "editorializing"
we don't know the mapping from electrical signal to mental/emotional state but we can still fall back on Signal Theory and build amplifiers that convey as much Shannon-Hartley “information” content as practical
Yes I forgot the malleus acts as a lever to increase the travel of the activation. Thanks.
A.F.A.I.K. it reacts to stress like all other muscles. I think the effect you refer to is when it is at it's limit of increasing tension. My understanding is that it really works as a volume physical volume control to keep the rest of the hearing mechanism in an optimal range. That is why sudden impulse noise can cause so much damage.
OK, you piqued my curiosity so I looked it up - the threshold for T.T. muscle activity in 'normal' mice (a decent model for you and I) is only 50 dB and the muscle activity increases up to 120 dB. Reflex latency starts at 4msec and decreases a bit to 3.5 msec when loud - hence the poor protection against cochlear damage. So the attenuation reflexes ARE working most of the time we hear reproduced music.
Over and out,
Frank
Nice work, but the issue is does preconditioning such as the crowd noise, alcohol, movement to get to your seat etc. cause a general change in muscle tension particularly the very small Tensor Tympani muscle?
BTY a mouse may not be a good model for me, do you have any data on rats?
Hi David,
There is no fundamental reason not to make the time constant longer, but some practical matters may intervene. First, a longer time constant, if it results in a lower sub-sonic frequency cutoff, will increase the settling time of the amplifier to its final, small, offset. In the meantime, the initial offset might result in some sound that is not masked by a turn-on delay. Just speculating here. The other thing is that I like to keep the size of the integrator capacitor to 1uF or below, so as to allow the use of a good quality one without spending a large amount of space and money.
I will usually go for an inverting integrator fed by a 1 Meg resistor and having a 1 uF feedback capacitor. I'll often use something like a JFET OPA604 for the op amp in the integrator.
Cheers,
Bob
Thanks Bob,
It is a dual comp jfet input so I don't expect there to be a lot offset to correct.
I may be able to get away with a long time constant and still use a large feed resistor.
I will be using a delayed output so settling time is not a issue.
I plan on using a nonstandard very low impedance at the non inverting input. I can do this because this amp will only ever be a one-off sitting in my living room.
Since this is the case is there any technical reason to not feed the non inverting input from the servo? There is a benefit of having the input filter involved by doing this and it simplifies the servo design.
Cheers,
David.
Last edited:
Yes.
Several years ago I was a member of an international discussion panel, that discussed amongst other things, this subject.
(the panel members was sound engineers and doctors specialiced in hearing damage), (I was one of the sound engineers, Bruce Swedien (very nice guy btw) was one of the other sound engineers).
I can write alot about the subject, but for the moment I will only answer your question and the answer is, Yes.
Cheers
S
Short answer: Probably
Longer answer: The primary literature only hints at the answer and it seems probable. Both physical activity and anxiety can affect T.T. tonus. ...and don't forget the smallest muscle in the body, the stapedeus, which also participates in modulating the tympanic membrane... Both receive inputs from serotonin-containing neurons and that stuff is, among other things, an anti-depressant and excitatory neurotransmitter. Nothing proven, but lots of reasons to suspect that sensory reception is different 'live'. However, are we talking only about amplitude? Dunno. Much recent work in the field is aimed at designing implants for patients with impaired or absent receptor function. The neurocircuitry is interesting and complex, with pitch, source localization, and 'startle' components handled (more or less) separately. So, we're not going to get any further along this little side-tour...
I encourage you all to pay attention to your intrinsic auditory background 'noise' and also your sensitivity. What do you hear in a truly silent space? My ears change noticeably week-to-week and if I have a cold, nothing sounds 'right'. The amount of background I experience makes a number like .01% THD seem like it should be nigh unto imperceptible but that is only to a first approximation, since the ears are not microphones.
Frank
The original issue is do our ears perform the same listening to a recording as they do a performance. I think we all agree the answer is no. So the next issue is how can we expect a recording to sound like a live performance?
I think the answer is twofold. It never really can but we can in a Picasso like way enhance certain elements to increase the impact of the recording to come closer to a performance. This can be anything from an alcoholic beverage, having friends over, to deliberately introducing distortions!
Bob, sorry if it seems I hijacked this!
ES