Re DC, there is no and there may not be any long term DC component in the music signal. Yes the wave may be asymmetrical in a short term view, but it would have zero DC when integrated over longer period. As Scott has already mentioned, multitones are asymmetric and very similar to music signals.
DC or missing DC across coupling capacitors is a different story and should not be mixed or mismatched with a common asymmetric music signal shape. We may not send a permanent DC to our speakers
.
DC or missing DC across coupling capacitors is a different story and should not be mixed or mismatched with a common asymmetric music signal shape. We may not send a permanent DC to our speakers
.
The reality is that if I only wanted to produce an audiophile quality speaker I achieved that quality long ago. What I decided to do instead was to repackage the design into something more modern and not a wooden box with a speaker in it. Most speaker builders are anything but, they are box builders and XO designers at best using commodity drivers. I've taken the road less traveled and developed the speaker from the magnet to the cone and designed a high quality acoustic enclosure that couldn't practically be done with wood while having the acoustic properties like wood.
So here I am in 2017 and I want to use current high quality devices. The larger of two amplifiers happens to be class-D from TI. It's the TPA3251, a high powered chip with loads of output power that will drive one device in BTL mode and only used below the 1% distortion level. If you look at the typical marketed output power it is listed at 10% distortion levels, obviously something to be avoided. Maximum output from the chip should fall about 120 watts@4 ohms and with an efficiency around 86db at 2.83v input into 4 ohms that should be loud enough I think at about 107db full output. I am intentionally limiting the max output to leave plenty of headroom with both the amp and the driver, at that output it should be able to run all day.
So here I am in 2017 and I want to use current high quality devices. The larger of two amplifiers happens to be class-D from TI. It's the TPA3251, a high powered chip with loads of output power that will drive one device in BTL mode and only used below the 1% distortion level. If you look at the typical marketed output power it is listed at 10% distortion levels, obviously something to be avoided. Maximum output from the chip should fall about 120 watts@4 ohms and with an efficiency around 86db at 2.83v input into 4 ohms that should be loud enough I think at about 107db full output. I am intentionally limiting the max output to leave plenty of headroom with both the amp and the driver, at that output it should be able to run all day.
Jan that's a good question and rather than give a quick answer let me think about my answer. I can think of some things that though real probably don't make for a compelling story, facts don't sell. So the question is what is the initial hook that catches your attention. Believe me, Jan, Demian already was asking about financial planning and marketing, none of it is simple but it isn't rocket science either, probably don't need a PhD to figure some of this out! Probably won't hurt all that time taking business classes.
ps. The Linear Audio crowd would probably appreciate the engineering effort and the sound quality from something unlike anything else.
ps. The Linear Audio crowd would probably appreciate the engineering effort and the sound quality from something unlike anything else.
Last edited:
Scott, Pavel,
As I often point out there really is no such thing as DC. We are all mortal. There is just very very low frequency.
Now many audio components are capacitor coupled to be 3 dB down at 20 hertz, the better ones at 2 hertz or even below.
As has been done one can do a simple rectangular wave when positive and negative values occur for different lengths of time and thus different levels. Of course it is AC but what it may show are among other issues are bits of thermal stuff.
Now there are of course many other ways to test for such issues, but I encouage playing around. I do try to discourage silly explanations for what may be observed.
As the three stooges joke "What happens you cut off one of a man's ears? He don't hear so good. What if both ears? He don't see so good. His hat will fall down over his eyes." (I do not reccomend you try this at home!)
As I often point out there really is no such thing as DC. We are all mortal. There is just very very low frequency.
Now many audio components are capacitor coupled to be 3 dB down at 20 hertz, the better ones at 2 hertz or even below.
As has been done one can do a simple rectangular wave when positive and negative values occur for different lengths of time and thus different levels. Of course it is AC but what it may show are among other issues are bits of thermal stuff.
Now there are of course many other ways to test for such issues, but I encouage playing around. I do try to discourage silly explanations for what may be observed.
As the three stooges joke "What happens you cut off one of a man's ears? He don't hear so good. What if both ears? He don't see so good. His hat will fall down over his eyes." (I do not reccomend you try this at home!)
Ed the discussion re: asymmetric waveforms always has and still lacks any understanding of the very basics of signal processing. Remember first cycle distortion, more nonsense?
Your point is the same one some audiophiles make when they claim the math uses integrals from plus to minus infinity so Fourier theory can't apply to real music.
Last edited:
But what happens if you feed weather satellite information into a pipe organ?
Celebration | Dacapo Records
I rather like it
Celebration | Dacapo Records
I rather like it
I'd have to ask, or I could get an apps guy to call you. You certainly would be a bona fide customer. I'm afraid any board houses we use are probably outside your budget.
No, in fact the R/C ladder model of DA is easily extracted with a bridge and ordinary stimuli. The information in paper you linked is easily done with a crafted multi-tone signal.
Using waveforms with abrupt steps adds the possibility of mixing information about the large signal transient behavior with information about the linearity of the transfer function. You actually get less information because the results become hard to unravel. Remember that was always your argument for not using a null test.
I'd be reluctant to get too far in scope creep. I mean you're already trying to break into what essentially is a high end Sonos-like system. Emphasize what makes you special. If there is later demand for such features (which I'd do first in software from a phone/tablet/Alexa before asking users to shout at their speakers.
The DC term for asymmetrical waveform was just a descriptor of a non-symmetrical waveform.... whose average single cycle was not zero DC. JC got it right with his paper shown. That's all. However, is there any bias shift or other operational weirdness behavior with short term?
I am set up now for the multi tone tests. But they don't exercise the amp same way.
THx-RNMarsh
Yes they do if you could figure out the proper phase relationships. The Hirata waveform can be decomposed into sine waves (bandlimiting it would only help make things more clear.)
I guess advancing knowledge and actually teaching folks some electronics/signal processing is not welcome here. Too narrow I guess, OK I'm chucking this thread, have fun.
Wouldn't a differential gain/differential phase test pull out any issues related to DC offset? I guess the only other relevant aspect would be drift from temperature/time/humidity etc. that would exacerbate any DG/DP issues. Content both by definition and numerous blocks in any chain removes DC. One interesting aspect would be the sonic impact of the high pass filters in the chain. A typical recording chain would have a high pass acoustic filter at the microphone (usually higher than we want to think), a high pass filter at the mike preamp, a high pass filter at the mixing console, a digital DC removing filter at the ADC all before you have the original capture. Then assuming the DAW doesn't do anything you have a digital stream into an DAC with a high pass filter to remove residual DC, then preamp and poweramp with high pass filters and a speaker that is also intrinsically limited in low frequency response. This would constitute a 9 pole high pass filter. In a real sense Richard is right to ask how that would affect the acoustic content and how audible it would be. This looks to me to be a real master's thesis project. More than I have the patience for.
- Status
- Not open for further replies.