How Loud is 1 Watt?

[PeterSt]

35 minutes ago, Kimo said:

I have my doubts.

 

Me too. 🤪

Watch out for good control. Loud does not mean that per se (the contrary). I did not watch the whole video, but this will be about the damping (I guess).

[PeterSt]

18 hours ago, Confused said:

I have a question, I guess aimed at either measurement specialists, or maybe someone with a very good knowledge of multi-meters.

 

I would never use a multi-meter for that. But *if* I do, I will ensure that the frequency (of mind you, a test signal !) is 50 Hz or lower; then the results are a kind of reliable.

 

That I deemed the video in the OP as a kind of gag right from its start, is something else.

[PeterSt]

A general remark: 

 

No means exist that I know of, to calculate the current requirement for music.

Music (obviously) bears a complex wave form and we can't know in advance what the voltage as such will be of all the frequencies added at one time. We also can't know how the speaker responds to all these frequencies at again the same time (this is impedance stuff).

 

Hence we can't predict how good the amplifier must be "in what".

 

Maybe it is telling that the amplifier I designed for my 118dB sensitive speakers, did 82WRMS into 4 Ohms (which is not sufficient for many speakers, as it can go to 1 Ohm).

What net came from this is 0.00053% THD at full gain into 1.5Ohms.

 

There is one real thing only which may let this fail : The power supply. Actually in more detail : the heat dissipation.

We could say that each amplifier can accomplish such a task. The question, however, is for how long.

 

Current capacity progressively degrades as heat adds. And when the capacity degrades too much, the voltage will drop.

Which voltage ? well, that on the most difficult transients (and you don't hear me say that these are the lower frequencies).

... And thus it implies distortion because not all frequencies (voltages for it) drop equally.

 

Clipping is a similar effect, technically different; the total voltage rails can't cope (say that the 20V required at the volume implied) drops to 18V, then all what required more than 18V, will be truncated, which makes the (always sine frequency) square-topped (truncated sine).

[PeterSt]

13 hours ago, accwai said:

So exactly which part of the power supply would be heating up when current demand on the amp side increases?

 

Easy enough: the voltage regulation(s).

 

More easy enough: the ambient of the power supply (other voltage regulations, transformer, a 1 million sllicon chips (if solid state), name it - in the end also resistors of course).

 

It's like a deadly circle; all of these components create heat which already induces self-collapse when not taken care of. This is why heat-dissipation in PCB design is a crucial part of the design.

[PeterSt]

13 hours ago, accwai said:

There is no such thing as difficult transient as far as the power supply goes, in theory at least. Is this correct?

 

I don't think so, but it obviously depends on the device. For example:

 

You may not realize it, but a bit of CPU which is rated at 100W, is capable of boosting up the current requirements from about 0 to 100W (or 20A if you want) in nanoseconds. I'd call that a pretty severe transient;

Now you know why power supplies for a PC are switching power supplies - this means allows these transients without delivering the 100W continuously (that would unnecessarily heat up the PC internally with the known result 😝).

 

... And now you also know why making a linear power supply for the same CPU's hence PC's, is quite a task.

 

Mind you please ... I am mentioning the 100W as example (and not 1000W which is quite common for gamers) because I still try to talk Audio Terms. Thus indeed these PC's exist, as we well know.

[PeterSt]

An additional phenomenon you most probably won't know about:

 

When the power supply of a PC is insufficient (hence the voltage drop because of current requirement is too high), then the PC will shut down. Ah, well, you knew that because this is just the ATX specification and the ATX power supply will just shut down the system. But what you did not think of, is that a power amplifier (etc.) which demands a higher current because of the music content, really behaves the same in the power supply department. One difference: nothing will shut down.

 

... Instead you should be able to hear this ...

 

So the more "infinite" the power supply can supply (current) the less impact the (music) demand will influence;

If a power supply delivers 100%, music can draw that 100% but the influence the music implies on itself will also be 100%.

If the power supply is so over-specified that the same level of music implies only 1% of its capabilities, the influence on the music itself is also 1% only.

 

Not to forget: it is all about how long a demand can be sustained (for delivery). This can be about Mahler in its wildest moments (loud level), but it can just as well be about the rendering of high frequency cymbals for a longer time (sustaining higher frequency transients).

[PeterSt]

7 hours ago, fas42 said:

I find it interesting that some people seem to create rigs that are noisy, when idling - for whatever reason, the gear I've played with is dead quiet ... on  maximum gain, with ear against tweeter, I don't hear anything

 

But what was the efficiency of the speaker(s) you used ?

You know the story about the 6dB more implies twice the power in theory, right ? So if you would be using a 94dB efficient speaker (mind you, measured at 1m and not "somewhere" like the video in the OP implies), and you would be using a 250W amp for that (never mind RMS mesured or not), then this would be similar to 250 x 2 x 2 x 2 x 2 in my situation (118dB). Now let me see ... that would be a 4000W amp.

Then try it again at full gain with your ear against the tweeter. 🤪

[PeterSt]

7 hours ago, fas42 said:

the rectifying diodes

 

They surely create heat too.

[PeterSt]

I actually can't envision how to avoid regulators in a good SS design. They are all over the place for various reasons. I mean, we won't have separate power supplies for e.g. 16V, 5V, 3V and 1.1V, right ? so that is reason one.

Reason two is the noiseless (as in PSSR) regulation, hence a best clean supply (like for oscillators, but so much more).

 

Maybe it is good to see that I am talking SMD/SMT components only, and that this is a quite different world from the huuuge thru-hole (legged) elements. So where the latter are usually inherently noisy to begin with, the small stuff was created to be more noise free (and maybe to save space). So mind you (also talking to Alex), we can do this all in-house and use 0402 size components where applicable (you can't even see those - it's like sand dust). So when I think of something today, next week I'll have it up and running, all soldered to a PCB (the week is necessary for the production of the PCB itself).

I don't think many people (companies) (can) work like this (which is all about speed of development - iterations).

[PeterSt]

11 hours ago, fas42 said:

The classic, square diode bridge rectifier is normally used for any higher powered unit - which is bolted to some heat sinking surface ... not much chance of overheating the silicon, I would think.

 

But Frank, then you still don't get it ...

 

Ales just mentioned the life time of capacitors. This is directly related to heat. OK ?

Each 10C of additional ambient directly influences the life time of the caps. Btw, each 20C too. Haha.

The point (my point) you seem to miss is that each heat radiation from a component should be able to leave the cabinet. If not, heat piles up fast (and you should know with your XMas temperatures over there *if* not in the middle of a wood fire to begin with).

 

I don't think I gave the example, but if I don't have the heat dissipation under control in our (Mach III) PC, *plus* when I don't have the boot process correctly under control (which is a BIOS thing) which makes the system booting from RAM taking 12 minutes for that, the 105W at first becomes 160W towards the end, and the system won't boot because of a fraction of a second the OS adds 100W to it (in W10 this is when the "window" wants to apear). This is in a fully optimized environment (BIOS + OS).

And the only difference from minute 1 to minute 12 is the heat in even an open cabinet. Read: Heatsinks are too small.

Or especially for you Frank: the metal work needs to be more expensive.

Of course we need to keep in mind that this is all in a 10cm height (thus very small) cabinet.

 

 

For understanding, one thing I left out: in this 105W-160W example, the Xeon involved was one tad too "over done" in combination with the 10A (linear) PSU for 12V; replace this 40 core Xeon for a 32 core and 100W (!) remains 100W forever.

 

The importance of this story is ore than crucial, because you won't really notice the increase in heat, BUT you can easily measure it with the IR meter on about each component. That's the heatsink of the PSU, the heatsink on the rectifier diodes, the transformer, the voltage regulators (those who must deliver 10A). In normal mode each is around 45-50C. But in "over done" mode each will be 60C or so.

So what.

... So that extra heat can leaver the cabinet and things get more and more inefficient. And it ends with your music sounding worse and worse.

 

And who is the one telling that systems may deteriorate after a few minutes - and who is even able to hear it on YouTubes ?

Frank.

But now Frank knows why that is (and if the amp only would shut down it would be more obvious).

 

[PeterSt]

FYI:

 

 

yes, that small. But this serves a purpose.

 

[PeterSt]

11 hours ago, fas42 said:

10's of watts is all that's needed, to deafen one - my current active speakers only have 50W driving the mid/bass; and they have no trouble filling the room with SPLs where one has to shout to be heard, if the other is at any distance away.

 

Yes, deafen one. That would be correct. But the big sport is to have a 100dBSPL or so, and NOT deafen.

 

Anyway, let me try to put you up the task to find an article on how much Watts music requires. Work it out for speakers of a given efficiency, if you get far with it in the first place.

Hint: waste of time.

[PeterSt]

1 hour ago, sandyk said:

dynamic range of RBCD is 96dB, and quiet suburban neighborhoods experience ambient noise levels of approximately 45-50 dB, (USA figures)

 

Hmm ... How is that related ?

You seem to imply that of 96dB of dynamic range, 45-50 dB falls in ambient noise ?

... Nah, it does not work like that. Juts put up the volume ...

(but I probably did not get what you wanted to say)

[PeterSt]

1 hour ago, phusis said:

but down into the lower midrange and bass there's no way around added power requirements and large(r) displacement area, period

 

1 hour ago, phusis said:

Power requirements (and handling ability) may be limited with very efficient speakers

 

No difference in the low-end that I can see / notice / measure / (stomach) feel.

[PeterSt]

1 hour ago, phusis said:

but there's a substantial difference between blasting away with 105-110dB's of your typical compressed music material (which is damn loud, I'd say, though perceived less so when reproduced cleanly), and a dynamically uninhibited recording with 105-110dB peaks (with much lower average SPL's). Now read that again and mark the distinction.

 

Good post, if I may say so. It makes me think of an other phenomenon, not discussed that I can see:

 

First off, what you are saying there relates to the average SPL overly compressed music exhibits much (much !) higher than more normally compressed music (a bit of compression is mandatory in the digital environment). This means that the SPL indeed is higher, since SPL is always an average over time. Thus, this already tells us it is not about the peaks really.

 

From the above follows that any amplifier would have much (much) more trouble to sustain that average SPL. I mean, the average level is much (much) higher, and this now is exactly my little subject again (see my last posts).

But wait, because outside of my posts, the subject changed somewhat to what we (humans) can bear ...

 

And now the thing: For us humans the same would apply; we can stand a peak of say 110dBSPL all right but we would not be able to stand a continuous 110dbSPL. Aha ...

 

From the latter follows that the better the reproduction, the less smear occurs (it is *always* about that when we talk accuracy), which also implies more difference between peaks and valley's (same as more dynamic range).

 

I could also bring to attention my ever (and ever) "good measure" of sound: no standing waves in the room (and not by room treatment by by means of better quality);

When we transfer this to "less room loading" you see the same thing happening. Because there's no sustained (illegal) sound pressure, the room gets the chance to unload and make room for the next (50Hz etc.) wave cycle.

 

Trust me; I work with this for over 10 years now as a sheer measure. But what's new to me today is that our ears and brain too need "unloading". And this should be the reason why I play at a level of 90dBSPL the family in the room reading a book etc., for 3-4 hours in a row each day, the whole year long. And this too is a sheer measure: when people start to close their ears or start to )fake) wear headphones, I know I am doing something wrong opposed to yesterday. Usually I feel that coming myself (unhappy about today's sound).

In this realm don't underestimate what I am going through at testing new configurations of ^2 or ^3 cables or whatever next year's novelty will be.

[PeterSt]

1 hour ago, Confused said:

I would be interested in the consensus view of anyone posting in this thread, what do we collectively think the SPL's in the room were during this clip?

 

All right. I am reporting live at listening to the video from post  #5.

 

After 3 seconds I notice this : A very (much too) hollow sound/space.

 

The fact that this is Dutch, already makes me shiver (sorry guys).

 

Garmt (the one with the glasses) only estimates the peak power the amps can deliver. Estimating does not help anyone here.

 

The question about the power meters is answered with "a. over a period of time" and b. "It shows it instantly".

Yeah, sure. The interviewer digs it.

 

The question on the 7W ampflifier is answered regardless the efficiency of the speaker (with the nasty given that this could be about the Harbeth speakers only ... but this is not (the context) how we observe the video).

 

Luckily he doesn't play Yello. But I am already confident that this (sh*tty) electronic music sounds completely different in my system, as the guy(s) don'r recognize the difference between distortion and what can be rendered. I will report later about *that* as I first have to find the music in the first place (hopefully he will tell what it is).

 

Garmt (et al ?) clearly thinks that if those lows are rendered without distortion, the remainder works "as well" (this can be explained two-folded, both in a negative).

 

The amount of excursion of the woofers (this is what you hear them talk about at 6:05) shows a very low SPL to me (the driver looks to be a 15"). Mind you, this is not a virtue, this is math (but includes damping capabilities of the system). Thus, no visible excursion tells me maybe 70dBSPL and it indeed includes you hear them talking at a not exaggerated level at all.

Then the guy with the blue shirt feels the diver at its edge (the suspension in my poor English). It does not work like that ...

 

 

OK, caught that one. :-)

 

After playing the track, the interviewer starts to draw various conclusions on things like "it is clear that the more power you are giving the better it sounds". ... As if there has been an other test with a lesser amplifier.

I start to sense a commercial here.

 

The sonata sounds terrible to me. So notice that the solo instrument there will be the most difficult to render, which is related to speed (you can call that damping as well).

 

OK, done.

Tbh, it took me 22 minutes to cover this 12 minute video, so I hope it was with sufficient attention.

 

Main conclusion : I reckon we listened to 70-dBSPL or so. Maybe 2 more.

Another conclusion : The video is awkward to the sense of that they show Watts, but not SPL. Same as with the OP video (which I did not watch in full).

 

 

 

 

[PeterSt]

So, fasten your seat belts.

 

The one at the top (see below) is where the music starts in that video.

The one at the bottom is a direct YouTube. Notice that obviously both are YouTubes (don't ask me about the bitrate). Let's regard the inherent quality of YouTube equal in this case (undoubtedly @sandyk is going to tell us.

 

A few remarks in advance:

- I have the album at hand so I will play it tonight through my main system.

- My beloved music *is* this kind of music, but with more (up/down)beat and more intelligence.

- I think I know when electronically based muslc sounds completely wrong (which I kind of announced in my previous post).

- Now see what happens here, and *know* that without the experience you can be told anything. OK ?

 

So my subjective judgment after listening to no more then when the higher pitched sounds start of the direct YouTube (but knowing what to expect):

 

1. The lows seem to render better BUT watch my remark on the hollow space (!! previous post).

2. Listen to the highs at the very start of the track. The "Garmt" one is one pile of distortion (but did you know that when you listened to that *wihout* listening to the direct YouTube version ?!?

3. The higher pitched sounds a but more into the track are 100% exactly what I expect from them, in the direct YouTube.

 

Ad 3. No (electronically generated) sound sounds "as is"; it always requires a source, so to speak. Listen to the direct YouTube and observe the envelope and frequency within the sound of a synthesizer. Compare the both now, and you will easily determine that the Habeth posted one is distortion only.

(and I know, this is recorded through a microphone, but hey, I wasn't the one posting this as a commercial ...)

 

What were we at ?

Ah, right. How 1W would ... etc. Well, that 1000W whatever isn't going to cut it. Or the speakers won't. Or the amplifier itself will not. Or the source. Or ...

Frank's response we can predict:
   See ?

But that's fine.

 

 

 

 

 

[PeterSt]

2 hours ago, Confused said:

7dB0 to 88dB.

 70-88 🙂

(somehow I made the same typo earlier on)

[PeterSt]

6 hours ago, Confused said:

If we assume 2 speakers with 84.5 dB / 1W/1m efficiency, measured from 4 meters or so / 14 feet (SPL decreases by -6dB with a doubling of distance), we get the following:

 

For a 95dB SPL

 

I have the feeling that you twisted things upside down there;

Nobody is stating that we should perceive an amount of SPL at some point in the room; at least I never saw something like that. Thus, your mentioned 84.5dBSPL is sufficient for me to know what will happen in the room if

- I know the distance (and to a lesser agree room size);

- I know the directivity of the speaker (my directive horns only "degrade" 6dBSPL at ~10 meters).

 

I you ask me how loud I play, I would indeed say "90dBSPL" (peaks). I measure this at 1m distance.

If you ask me how loud a piano grand plays, I again will say 90dBSPL (which is measured at 1m distance !).

If I'd like to play real levels, I could tune it on the piano grand (and cymbals at 110dBSPL will be killing).

 

Never ever I would say "and this is with listening at x meters distance".

What I would say though is: and don't listen to this at 1m distance, which distance  would be impossible to listen to a drumkit just the same.

 

... So this is also not how the Harbeth video is to be judged.

What could be done though, is looking at faces, looking at how far they are from the speaker (which is 50cm) and how they not shout at all in order to talk to each other.

And so if I say that the lot plays at 72dBSPL, it is that measured at 1m from the speaker.

If you say it is 88dBSPL it is again at 1m distance from the speaker.

... Never ever I tried for whatever reason to play at 90dBSPL at my listening position. It is just n/a.

 

Still room size matters and my 90dBSPL is in a room of 12x8x3m. At 10 meters distance from the speakers, we can quite easily talk (but with shouting), and we now know that in my case this is about 84dBSPL.

Point is and remains that an amplifier doesn't know about this all. This just tries to squeeze out that 90dBSPL and how I perceive that is not related.

 

I am not sure this all makes sense. But with my own reference, I could try to perceive the same loudness in that say 4 meter "deep" room of the Harbeth video, want to listen at my levels which is 84dBSPL at the listening position, and thus should imply 96dBSPL from the speaker at 1 meter (for the normal dipole speaker).

But now think again ...

... Those people were talking 50cm from the speaker and you could hear them talk;

Senseless or not, I just talked to my SPL meter (of course at 1 meter distance) and with a little emphasis in my voice (to overrule a virtual speaker) it showed 84dbSPL.

 

Blahblahblah 😁

[PeterSt]

 

11 minutes ago, yamamoto2002 said:

voltage becomes larger when music signal is more pulsive.

 

Very nicely executed !

 

 

This one will show the most because it is continuous square and it contains the most frequencies (infinite amount at pure square).

I am not sure whether it can ever be more than this, but if not, it would be representative for the most complex music !

... interesting ...

 

[PeterSt]

31 minutes ago, Confused said:

Anyway, lots of ways of looking at this, so let me ask.  If we are trying to calculate how many watts, what numbers would you use for the calculation?  And if we assume 70dB or so, very close to one speaker, the calculated answer will be a fraction of a watt.

I considered it could be more complicated; so just some blahblah again:

 

I played the track yesterday, and it is all but a loud track. Also, to everybody's surprise, there's hardly LF in it, meaning hardly anything under 200Hz. There is, but this is lower level sine stuff (remember this is (I think) 100% electronic music (the remainder of the album is the worst you can imagine). However ...

 

If I counted right, once in the 4 main beats there is sub-low, and this rambles on for a second or 2. I didn't really perceive this at listening to the track, with the notice that I did not play it really loud either. Btw, this sub low seems odd as in "an anomaly" (can't see the reason why to do this once per 4 beats). *That* will definitely ask a bit more from the amplifiers.

Now remember the guys (at 50 cm distance) talking about the woofer not moving ... correct. The low that you hear does not come from the woofer because ... it is not low. And feel the woofer during the 10 or so seconds this "on the 4th beat" is *not* in order, and you thus don't feel a thing.

 

Over here it is an all but powerful track - just like electronic music usually is. It's dumb stuff only and not "audiophile" at all.

But now again think of that hall sound in that room in the first place. All starts to reflect and will imply standing waves of different (emphasized) frequencies.

 

I'm drifting off ...

[PeterSt]

17 hours ago, sandyk said:

Weird things happened when I tried to mix this with the audio, with a redirection back to other music on the website

 

https://www.dropbox.com/s/jc8mufi1lgqb8mf/How loud is 1 Watt (with an 87db Loudspeaker)-0x0002.aac?dl=0

 

Thank you Alex;

The weird thing could be about the track we are talking about not being there at all (the Sonata's track is), while the length of your audio is again 11:x minutes.

But let it be, please.

[PeterSt]

38 minutes ago, Confused said:

I am not sure if this will add any insight, but here is is anyway, this is for the Laptev Sea track:

 

Haha, probably not. But if this is the Habeth YouTube, you may consider doing/posting the direct YouTube as well (or the other way around). We should be able to see differences easily (not sure what for, but it could be fun).

[PeterSt]

2 minutes ago, Confused said:

and a Dutchman talking nonsense

 

Hahaha, you got me !

[PeterSt]

On 12/10/2020 at 3:29 PM, phusis said:

but that you don't find the lowest frequencies to be "musically [..] very interesting" to me is indicating that you haven't experienced what this range, when reproduced effortless and cleanly via lots of displacement area and properly implemented, can do to the musical enjoyment.

It should be noted that Frank explicitly doesn't want any reproduction under a 100Hz.

Beats me why.

IOW, you will be correct there.