Driver movement

[audio2920]

Ah OK. Understood @SIY.

I do appreciate your point that this stuff does require some effort in learning before you can engage with it. I would like to think that as mixer, I have enough understanding, but clearly, being able to create a good mix does not rely heavily on this, nor vica versa.

To entertain your question, I would say only 30Hz...?

 

[NTK]

Ah OK. Understood @SIY.

I do appreciate your point that this stuff does require some effort in learning before you can engage with it. I would like to think that as mixer, I have enough understanding, but clearly, being able to create a good mix does not rely heavily on this, nor vica versa.

To entertain your question, I would say only 30Hz...?

 

[SIY]

To entertain your question, I would say only 30Hz...?

Nope. And trust me, this is something that's been incredibly evident from a lot of posts in this thread.

To put whipped cream and a cherry on it, there was a series of LONG articles in (I think) Wireless World about "First Cycle Distortion" that got a lot of attention from people who wouldn't get that right either. I was asked to review those articles, got about two paragraphs in, and stopped bothering- "The author doesn't understand basic Fourier transforms, and it goes wordily downhill from there." In audio, ignorance is not a barrier to dissemination.

edit: NTK got there first. I am not a rapid typist!

 

[audio2920]

Nope. And trust me, this is something that's been incredibly evident from a lot of posts in this thread.

To put whipped cream and a cherry on it, there was a series of LONG articles in (I think) Wireless World about "First Cycle Distortion" that got a lot of attention from people who wouldn't get that right either. I was asked to review those articles, got about two paragraphs in, and stopped bothering- "The author doesn't understand basic Fourier transforms, and it goes wordily downhill from there." In audio, ignorance is not a barrier to dissemination.

edit: NTK got there first. I am not a rapid typist!

Sorry. Yes. I answered without thinking!!

But that, I did know. But thank you @NTK it's a good refresher.

 

[NTK]

This is the important post

This is the problem with interpreting CSD without a firm understanding of FFT and how these plots are produced.

Look at the time axis, it goes to 13 ms. 13 ms is the period of a 77 Hz signal, which means the frequency resolution is at best 77 Hz. This plot barely gives enough time for 1 cycle at 100 Hz, and is far from sufficient to figure out the decay time at the lower frequencies.

 

[dshreter]

This is the problem with interpreting CSD without a firm understanding of FFT and how these plots are produced.

Look at the time axis, it goes to 13 ms. 13 ms is the period of a 77 Hz signal, which means the frequency resolution is at best 77 Hz. This plot barely gives enough time for 1 cycle at 100 Hz, and is far from sufficient to figure out the decay time at the lower frequencies.

I think that is why @BrokenEnglishGuy is asking for help with others' interpretation of the CSD for a couple speakers of interest. I don't think he is inferring the conclusions and is interested to know what can be observed in the plots - for example what differences are observed between R3 and F208 based upon the CSD plots.

I don't think it helps that everyone is saying he doesn't understand the plots - that's why he is here asking. I've shared the morsels I know, but don't feel qualified to do that interpretation. But I know many here have the experience to be able to give a quick interpretation and comparison of what jumps out from the two CSD plots.

 

[BrokenEnglishGuy]

This is the problem with interpreting CSD without a firm understanding of FFT and how these plots are produced.

Look at the time axis, it goes to 13 ms. 13 ms is the period of a 77 Hz signal, which means the frequency resolution is at best 77 Hz. This plot barely gives enough time for 1 cycle at 100 Hz, and is far from sufficient to figure out the decay time at the lower frequencies.

The problem is i saw too much interpretation of CSD or impulse response wrongly and at this point i got confused about these woofers ( R series woofers )

I wanna know if the woofers from the R series hace enough dampening, or if the frequencys between 200hz-400hz are ringing too much and can cause problems or something, the speaker seems to play very loud and low in distortion..

The woofer works between 20hz - 400hz.

 

[BrokenEnglishGuy]

I think that is why @BrokenEnglishGuy is asking for help with others' interpretation of the CSD for a couple speakers of interest. I don't think he is inferring the conclusions and is interested to know what can be observed in the plots - for example what differences are observed between R3 and F208 based upon the CSD plots.

I don't think it helps that everyone is saying he doesn't understand the plots - that's why he is here asking. I've shared the morsels I know, but don't feel qualified to do that interpretation. But I know many here have the experience to be able to give a quick interpretation and comparison of what jumps out from the two CSD plots.

You are right, i'm asking for help for an accurate interpretation to the CSD, because i want to know the woofer quality from the R series from 2018.
But the woofer distortion looks very good even this loud

 

[sarumbear]

Anyway if for example, the driver doesn't have enough self damping, the driver should stop slowly and keep '' ringing '', but if the driver need to do a faster wavelenght while it's ringing, there will be a raise of distortion in the next wavelenght?

I mean, in the first escenario is a driver playing 40hz wave and then 350hz wavelenght, the second one was play while the driver keep ringing in their first miliseconds
The second scenario is the same driver but only playing the 350hz wavelenght, in this scenario the driver gonna produce the 350hz wavelenght better?

Te second scenario is closer to the reality, but a frequency swap plays the entire 20hz->400hz so.. i don't know, but is not like that test set a permanent wavelenght of 40hz and then start to play randomly upper frequency above 40hz like
40hz -> 300hz
40hz -> 350hz
40hz -> 333hz
And so on, you cannot measure how these transitions are done? or it lacks of an objective purpose? and the swap 20 hz -> 400 hz is enough?

Nothing like you explain above happens. There is no “self damping,” no “faster wavelength,” and “ringing” is a completely different behaviour to what you think.

There’s no need to create “scenarios”. Everything you are concerned with is well understood. All you need to do is either study them or stop worrying and accept people who did are saying.

 

[BrokenEnglishGuy]

You are right, i'm asking for help for an accurate interpretation to the CSD, because i want to know the woofer quality from the R series from 2018.
But the woofer distortion looks very good even this loud

My personal take about this is:
- The cabinet doesn't suffer resonances
- Fr looks linear
- Distortion is very low

Having these facts, the woofer will reproduce the sound accurate and i don't see any quality problem. (?)
*My ears thinks the woofer play very clean without any problem, only the speaker sound weird when is not well integrate in the room.*

 

[sarumbear]

not slow enough is not a thing. every driver can produce 30Hz or even 3Hz. the question is if it is audible. you can create a 5Hz pressure wave with a folding fan in your hand, but you will never hear it because you are not displacing enough air

You need a fan

Eminent Technology: home

 

[audio2920]

Not that this is helpful in any way to the OP, sorry.... but for my own sanity check to clear up why I said "30Hz only" to the simple problem @SIY posed:

I skim read the question and assumed the FT window was one cycle. Stupid that I didn't read more carefully, i was just rushing about...

Anyway, if the window had been as I thought, i.e. the exact duration of the full cycle (t=0 to t=1/30) and didn't include the "DC" zeros pre and post, then I think there'd be no change in rate of change within the window, no asymmetry etc. Hence the FT output would be the 30Hz itself only, right? Or is that also wrong?

 

[ebslo]

Anyway, if the window had been as I thought, i.e. the exact duration of the full cycle (t=0 to t=1/30) and didn't include the "DC" zeros pre and post, then I think there'd be no change in rate of change within the window, no asymmetry etc. Hence the FT output would be the 30Hz itself only, right? Or is that also wrong?

That's correct (ignoring the negative frequency component anyway), but then your FT will only have 30Hz resolution.

 

[audio2920]

Thanks @ebslo! I appreciate it. (Just checking I'd not completely lost the plot.)

 

[ebslo]

Thanks @ebslo! I appreciate it. (Just checking I'd not completely lost the plot.)

Sure. That said, what you describe is really just an anomaly of the DTFT. The actual FT is defined as an integral from -inf < t < inf, so to take the actual FT we must specify the signal in both directions beyond our single cycle. The highly contrived case of taking the DTFT of exactly a single cycle of a sinusoid has very little applicability to anything in the real wold. In addition to our inability to "start" and "end" time, it's only possible to even contrive this case for periods that are an exact multiple of the sample period.

 

[SIY]

Sure. That said, what you describe is really just an anomaly of the DTFT. The actual FT is defined as an integral from -inf < t < inf, so to take the actual FT we must specify the signal in both directions beyond our single cycle.

Which was exactly how the problem was stated. The take-away is that starting and stopping inevitably are functions of bandwidth.

 

[audio2920]

The actual FT is defined as an integral from -inf < t < inf, so to take the actual FT we must specify the signal in both directions beyond our single cycle.

Understood @ebslo & @SIY thank you for taking the time.

I was only asking because I misread the original question, which then got me thinking... if that *had* been what SIY asked...

 

[AdamG]

Thread closed as requested by OP!