Grainy sound in the highs drivers on voice only

[Lucifer1to10]

i tried moving the driver on another horn i had around the house, bigger in size same result ,its a grainy sound in the vocals like right behind the vocals like a shadow or a a countour around vocals

 

[voodooless]

What crossover do you have? Do you specifically compensate the CD/horn curve?

 

[Lucifer1to10]

winford wf-2a80-k2 crossover no i dont compensate the horn curve, everything was flat for the tests also if i try high cut or any other adjustments the eq is working ok but the grainy sound remains , just boilled and oppened one driver , i don t think it should look like this inside

 

[Lucifer1to10]

i got a wiim pro ,conected the amp to it and no change ,so it is impossible for me to find the actual source of the problem , also tried another amp with other speakers, same sound '"grainy" " sandy'" only on voice also tested on headphone same thing ,even on Bluetooth i really can't find the problem ,from my testing it results it is not the amplifieres or the speakers being that on headphones i get the same sound , tested line out ,optical out , Bluetooth out and all the in options ethernet,bluetooth line in ,optical in ,played with all the settings in the wiim pro nothing removes that grainy-sandy sound , called a few friends over same reaction they all hear what i hear ,so i guess my hearing at least is fine , i cannot imagine what is going on if u guys have more ideeas let me know please !

 

[Earth]

Hi Lucifer1to10.​

You are describing the sonic characteristics of the Ultracurve well... early digital can be improved, though it should be asked is it worth it?
Everything connected after the Ultracurve will have the characteristics you describe.
These characteristics will be further exacerbated by using metal dome tweeters.

The external options are:

Only use engineered quality coax to connect the digital signals, it's direct unlike spdif/ opticals.

Get a professional filter on the electric mains power to supply all digital devices, ones with coil inductors, not the capacitor only type, which are next to useless.
You will find it difficult to get a low pass filter below 10Khz at the rated current required, 100Khz is fairly standard. This is mains electric, mistakes are not optional, take all precautions.


Internals options are complex, for the experienced and will require the service manual, generally:

Cleaning up the noise into and outputting from all regulators to all digital power consumer chips using MLCC caps to the chips data sheets specs will make a difference.
Doing the same to the DAC's with C0G/NP0 MLCC's to the chips spec will make a big difference, or X7R type if the specified values are larger.

The SMPS power supply will be able to be cleaned up with filters on the outputs.
First Check/ replace the chopper chips as these degrade with time, these are a huge source of noise.

Playing about with everything else after the Ultracurve is a waste of time, it's not the problem, digital square wave noise is what you are trying to reduce.

 

[RandomEar]

Hi Lucifer1to10.​

You are describing the sonic characteristics of the Ultracurve well... early digital can be improved, though it should be asked is it worth it?
Everything connected after the Ultracurve will have the characteristics you describe.
These characteristics will be further exacerbated by using metal dome tweeters.

The external options are:

Only use engineered quality coax to connect the digital signals, it's direct unlike spdif/ opticals.

Get a professional filter on the electric mains power to supply all digital devices, ones with coil inductors, not the capacitor only type, which are next to useless.
You will find it difficult to get a low pass filter below 10Khz at the rated current required, 100Khz is fairly standard. This is mains electric, mistakes are not optional, take all precautions.


Internals options are complex, for the experienced and will require the service manual, generally:

Cleaning up the noise into and outputting from all regulators to all digital power consumer chips using MLCC caps to the chips data sheets specs will make a difference.
Doing the same to the DAC's with C0G/NP0 MLCC's to the chips spec will make a big difference, or X7R type if the specified values are larger.

The SMPS power supply will be able to be cleaned up with filters on the outputs.
First Check/ replace the chopper chips as these degrade with time, these are a huge source of noise.

Playing about with everything else after the Ultracurve is a waste of time, it's not the problem, digital square wave noise is what you are trying to reduce.

OP said he has the same problem with a different source (phone), different amp, speakers and even headphones. So clearly, it cannot be the Ultra Curve if the problem pesists with that device not even in the playback chain. In essence, with the current tests OP listed and knowing that they are subjective (sighted listening) it's not possible to give any solid advice other than "get a UMIK-1 and measure it". Randomly installing mains filters or soldering in SMD caps is the worst advice to give at the moment. "Cleaning up the SMPS" is just nonsense, especially if it's not even remotely clear what the problem is.

 

[PirateQuinn]

It's not the ultra curve unless something is overflowing due to gain issues.

 

[Earth]

So clearly, it cannot be the Ultra Curve if the problem pesists with that device not even in the playback chain.

Of course. Process of elimination clearly written will allow accurate diagnosis. However it is anything but clearly written.

Having owned and worked with the Ultracurve, it was the first piece of digital gear that we had to move on immediately after test, due to a hideous resonant treble ringing inherent in the architecture that we could not get rid of; extremely 'grainy in the top end'. Had I know what the causes of these were are back then, the above post #45 would have been completed in order of priority. But why would you want to do that to a brand new machine? Experience has shown me, when the harmonics of the square wave are ruthlessly reduced and brought back to square, most of that digital mince disappears. Particularly when reducing noise to processing chips that handle the audio.

SMPS's have only become quiet in the last few years, but they are still not relaible.
The older ones require attention if clear reliable audio is the goal.

 

[RandomEar]

Again: The problem was also present without the Ultracurve in the signal chain. So either the problem is somewhere else, or there are multiple problems with the exact same effect and the Ultracurve is one source of them (very unlikely).

Also, what does "harmonics of the square wave" refer to? There are no square waves being played here. And what would the SMPS have to do with any of this? Can you show any evidence for your claims?

 

[Redacted]

"harmonics of the square wave" is a chart-topping album by "Dave and the Sig Gens"

I'm sure it's just me, being an old git, but I've lost track of this thread and can't identify if there has been a test setup that has identified a cause.

If I thought it was the recording itself, I'd probably bin it, or ADC it and see if I could process it. In the end, all actual 'classical' recordings are just tribute bands doing a version (you're not allowed to say this as it hurts feelings) and there are often many versions available. Many recordings that are entirely mic'ed up are horror stories though.

 

[Milenial44]

If I had already swapped drivers, crossovers, amps, DACs, sources, and even connected the compression driver with just a capacitor, I'd stop blaming the electronics. I'd start looking at the horn/driver integration or the crossover point. To me it sounds like the HF driver is playing too much of the vocal range and it's giving that harsh, grainy character.


I'd also try lowering the crossover frequency or at least padding the horn down a few dB with an L-pad. If instrumental tracks sound clean and it's only voices that are bad across every source, I'd honestly suspect the speaker setup or tuning rather than a faulty component.

 

[Earth]

Worst square wave signal I've seen on a 'stock' output:

Same machine as above trace. Brought back to 'square':

Another machine:

The stock digital out:

The cleaned up digi out:

Square wave signals attract noise, that's their nature. Sinusoidal wave less so.
Square waves (digital signals) are susceptible to interference that manifests as distorted harmonics of that square wave.
If the 'on' and 'off's of the binary stream are not clearly defined, what ever is down steam from that, will have to resolve it some how, error correction, interpolation and so on.


Here is a really good example of what a square wave should look like in the real world on an digital audio output:

Try an image search for: square wave rise times on an oscilloscope trace

 

[antcollinet]

Worst square wave signal I've seen on a 'stock' output:
View attachment 550135

Same machine as above trace. Brought back to 'square':
View attachment 550136


Another machine:

The stock digital out:


View attachment 550137

The cleaned up digi out:
View attachment 550138


Square wave signals attract noise, that's their nature. Sinusoidal wave less so.
Square waves (digital signals) are susceptible to interference that manifests as distorted harmonics of that square wave.
If the 'on' and 'off's of the binary stream are not clearly defined, what ever is down steam from that, will have to resolve it some how, error correction, interpolation and so on.


Here is a really good example of what a square wave should look like in the real world on an digital audio output:


View attachment 550139



Try an image search for: square wave rise times on an oscilloscope trace

Not sure how digital output scope plots have anything to do with the topic of the thread.

If you are replying to a specific post that puts yours in context, then you should quote the post you are replying to. As I have done yours.

 

[Earth]

Well with the information provided....and from experience:

A first favorite looks like it's the metal dome/horn combo, bright and brash.

A close second is the Ultracurve and their bad layout of noise suppression in the stock circuitry which has at best a grainy top end.

The scope readouts are examples of some stock bad digital audio outputs and what good ones should look like.
How is any DAC is going to resolve bad digital outputs like those shown, without causing problems in the top end, artifacts, more glass, brittle, sharp, grainy effects...

 

[Earth]

Unless of course the disassembly and reassembly of the tweeters has not been done concentrically to manufacturers tolerances, or is it a manufacturers error?
#43 ...the last picture, if that's not oxidization, looks like the internal vertical face of the diaphragm has been vibrating against the stator magnet.
Check for wear on the coils side also, diametrically opposite would be what is shown in the image.
I don't know how anyone could reliably achieve that precision in a home workshop environment unless the diaphragms have well designed locating points to the stator already built in....or, you have specialist equipment.

The central metal cylindrical plug that focuses the magnetic plane for the coil, can get knocked out of true if dropped or disassembled or assembled out of true, that would also have a similar effect. Most designs are glued, but some are there by the grace of magnetism...

 

[antcollinet]

How is any DAC is going to resolve bad digital outputs like those shown, without causing problems in the top end, artifacts, more glass, brittle, sharp, grainy effects...

That is not how digital errors manifest. Digital errors result in a whole frame of multiple samples going missing. The result is drop outs, clicks, pops etc. Not "problems in the top end" nor, any of the other attributes you've mentioned above.

If you are not getting either dropouts or clicks then there are no digital errors - no matter how bad the digital waveform.



And you STILL need to be quoting the posts you are replying to - please - it is somewhat difficult to tell what point you are making without it.

 

[Earth]

By all means make a meaningful contribution to the thread, it's better than criticising general observations of the real world.

Oh, you are not a moderator.

 

[Salt]

Someone on this forum once replied to one of my comments "hold your horses" ...