It's loud.


Glued to the PCB, not to the case. My understanding is that you can disassemble it as you want, the "boom cage" just sticks to the PCB and the speaker's membrane sticks to the case ?

that makes more sense, i guess i misunderstood ED
 
The speakers aren't glued, they lie open in the cage and you can replace them just swapping them out without any soldering.

They are not in a vacuum, they just need to be sealed while they're being used. It works similar like a damper: If it's sealed, the membrane can't move as far (as the air is limited in the back volume) and therefore needs space for the same power as other speakers.

So you easily assemble and disassemble the unit.

About the questions about the volumes:
Well, Pandora had 100% volume from the volume wheel, but also only 75% volume with ALSA (so similar to the Pyra).
Based on my experience, driving ALSA over 75% (into the red area) almost certainly overdrives, even though the speakers sound okay. My guess is that ALSA itself starts clipping at some point. I guess that's why ALSA usually has 75% or similar as default.

Also, the reason why the Pyra is limited to 75% is because it starts to overdrive already right now. This has multiple reasons:

1. The amp is not limited yet, it can output more power than the speaker can take.
2. The seal is probably not 100% sealing yet.
3. The back volume is not optimized yet.

I'll work on that.
The reason the GPD XD was at full volume was because it still sounded good in real life (I didn't hear it being overdriven), but it sounded a bit worse with the microphone. The ugly reverb was clearly hearable in real life as well - the reason for that is probably because the GPD XD uses the full case as back volume. Which adds reverb.

I'll do more tests as I continue to work on that and will also lower the volume of the GPD XD.
 
  • Like
Reactions: rSl
I'm very surprised that the GPD sounds so terrible (even if it is loud), the Pandora is quieter but sounds much better.

Happy to hear the pyra is both loud and produces a clean sound.
 
Last edited:
Sounds like good results already, with probably a lot of room for improvement.

We need specs and at the very least some audio of the distortion at full power

Alternative reasons:

1. The driver is loose, rattling on its own.
2. The driver is catching on the grille at low frequency xmax
3. The driver is catching on the PCB at low frequency xmax
4. The driver is front-loaded by the grille, turning it into a compression-driver
5. The front case as a baffle is a problem
6. The front waveguide is a problem
7. The front speaker grille is a problem
8. The back waveguide is a problem
9. The diffraction against the PCB is a problem
10. The space for the back of the driver is a problem, could be standing waves, could be many things
11. The openings for ports, battery compartment seal (etc) in the back of the case is a problem
12. The case dampening is a problematic

Many of these lead to a problem for the amp, and

13. The amp has a problem with normal loads

To troubleshoot

free air
  • try the driver outside the case, radiating into full space
Provided we get this far, and assuming ideal working conditions, try the following first with the driver sealed into place with silicone along its edges, then again like that, only now with the front grille completely removed.

dipole
  • use only the upper lid as a baffle
closed case
  • sealing off the back, try it in infinite baffle, needs an area over VAS, use something like a half a kinder-egg with dry silicone along the edges to push it closed over the back of the driver. Try multiple shapes by crunching it or finding something else
bandpass
  • assembling the case, try it in the case
bandpass 2
  • add dampening material to the sealed case.
bass-reflex
  • try sealing off some but not all off the openings in the case
Could also try the GPD drivers or pandora drivers in known to work or these conditions, or,
move the pyra drivers to the pandora or the GPD.

If you do a frequency sweep you can see what happens to the resonant frequency for all of these things.

Here it is at 600Hz, regular driver
f_seas_exotic_loudspeaker_tweeter__x3-06_exotic_t35.jpg

From what i could see on the pictures it looked like the back volume was almost completely sealed.
The low frequencies taper off, and the high frequencies are lost, more alarmingly;
It moves that impedance peak into where it isn't supposed to be, and sharpens it. Meaning the dampening factor is off there seen by the amp. It also moves to a higher frequency, The difference between that peak is greater as a result relative to the other frequencies. If the driver is lost beyond linear excursion, that isn't good, and its especially not good given you can throw any combination of the frequency-band at it while its there.

QMS is way above spec, leading to
QES being all over the place, meaning intermodular distortion
QTS is down for everything above resonant frequency

Is it a ribbon driver, or am i missing something?
 
Last edited:
  • Like
Reactions: rSl
Sounds great for me and even if the sound quality can't be improved I would be perfectly happy with it as it is.
For me the best news here is, that the speakers can be changed by the enduser without soldering when they are broken. This should also be stated somewhere in the official presskit. I'm thinking here about how much the Fairphone producers talk about how easy it is to change parts. In my opinion it's even better on the Pyra. The speakers are not soldered on an extra Platine, so you could even buy them somewhere else and not only in the official shop as you'd have to do with the Fairphone.

BTW: I'm in the same time zone like ED, so the post was published at 3:45 a.m. ;-) ED, go to bed!
 
Since the volume rolls off for each time the distance doubles, you would get a more accurate recording if you measure at head-distance. That is also more correct given we will use it that way. If you add some seperation similar to that of a head between the microphones, its closer to what we listen to on headphones (binaural recording)

The microphones are overdriven at 3.33, so it will help with that too.

The screen will horn-load the drivers on the pyra, in a way that further removes the sampling position from listening-position, so if you could work in a test with just different positions of that relative to the others, it would be great.

Where can we listen to the original tracks, and are they high-quality?
 
Last edited:
Thank you for the video!
This is what i was was waiting for!! I am thrilled by the sound so far. In my opinion the Pyra is the clear winner. It is more voluminous than the GPD,(which sounds a bit overmodulated) especially with more demanding music, like the Chris Huelsbeck track! Would be interesting to hear all the opportunitys of the speaker.

So far i am very pleased. Thanks ED!
 
I was about to complain that every time I come here I just see updates for PGS and GPD... but I will forgive you this time xD because video is awesome !! :cool:

Really interesting and clear comparison, thanks
 
1. The driver is loose, rattling on its own.

Nope, sitting tightly in there.

2. The driver is catching on the grille at low frequency xmax

Shouldn't, as it fully padded.

3. The driver is catching on the PCB at low frequency xmax

The backside of the speaker is sealed by itself, so that shouldn't be an issue.

4. The driver is front-loaded by the grille, turning it into a compression-driver
5. The front case as a baffle is a problem
6. The front waveguide is a problem
7. The front speaker grille is a problem
8. The back waveguide is a problem
9. The diffraction against the PCB is a problem
10. The space for the back of the driver is a problem, could be standing waves, could be many things

This shouldn't be the case, as we followed the design guide and the speaker is exactly made for the conditions we have here.

11. The openings for ports, battery compartment seal (etc) in the back of the case is a problem

Defnititely not, as that's totally independent of the speaker.

To troubleshoot
free air
  • try the driver outside the case, radiating into full space
Provided we get this far, and assuming ideal working conditions, try the following first with the driver sealed into place with silicone along its edges, then again like that, only now with the front grille completely removed.

This will instantly kill the speaker. The speaker must NOT be run outside a sealed cage, as the membrane will be damaged pretty quickly in that case.


bass-reflex
  • try sealing off some but not all off the openings in the case
Once again: This sounds worse. We ran the speakers without sealing first and they were overdriven with 20% or higher volume.
These speaker MUST have a sealed backvolume to work properly. Everything else will fail.
That's why my guess is that the seal is not 100% yet.

Could also try the GPD drivers or pandora drivers in known to work or these conditions, or,
move the pyra drivers to the pandora or the GPD.

Both these speakers don't need a sealed back volume, they are totally different types of speakers.
And the Pyra speakers can't be used in the Pandora or GPD XD as there is absolutely no back volume (the manufacturer suggests 1cm³ for best results).

If you do a frequency sweep you can see what happens to the resonant frequency for all of these things.

Yes, that's something I have planned.

From what i could see on the pictures it looked like the back volume was almost completely sealed.
The low frequencies taper off, and the high frequencies are lost, more alarmingly;

The frequencies itself are pretty good.
See this picture:

The top one is the original file, the bottom one the recording through the speaker.
The frequency curve actually looks pretty good.

Don't forget that the microphone also influences the recording, which can be the reason the waveform looks like it has less volume.

waveform.jpg

The high frequencies aren't lost, why do you think they are?
It's pretty crystal clear, just seems to have some resonance frequencies left.
 
I'm very surprised that the GPD sounds so terrible

After owning a GPD G5A, I'm not. Their build quality is absolutely terrible. Wouldn't be surprised if it has two different speakers, whatever they pulled out of the box that day. Ah man, I shouldn't be so negative. I just have a sour taste in my mouth after that fiasco.

Man that Pyra looks so hot. I want one so bad.
 
This is perfekt, because i want my Pyra to akt also as an Mp3 Player for Sport, whit a big SD Card..
The Question is, how good dos a Bluetooth Headphone work out of the box..
My Headphones are able to controll the Musik Playback from my Ipod Touch 5, so if this also works whit Pyra, this would be great..
 
After owning a GPD G5A, I'm not. Their build quality is absolutely terrible. Wouldn't be surprised if it has two different speakers, whatever they pulled out of the box that day. Ah man, I shouldn't be so negative. I just have a sour taste in my mouth after that fiasco.

Well, I know I found at least two different speaker manufacturers in the GPD XDs, as they clearly look different from console to console.
 
My Headphones are able to controll the Musik Playback from my Ipod Touch 5, so if this also works whit Pyra, this would be great..

Mainly the Bluetooth things that works with Apple things, does not work with everything else... but maybe you are lucky and the Pyra driver could be hacked someway :)
 
Mainly the Bluetooth things that works with Apple things, does not work with everything else... but maybe you are lucky and the Pyra driver could be hacked someway :)

Huh? Care to elaborate. My bluetooth stuff works with android, apple, windows... even my Pandy. What do you mean?
 
This is perfekt, because i want my Pyra to akt also as an Mp3 Player for Sport, whit a big SD Card..
The Question is, how good dos a Bluetooth Headphone work out of the box..
My Headphones are able to controll the Musik Playback from my Ipod Touch 5, so if this also works whit Pyra, this would be great..

Bluetooth headsets simply take what is more or less a line level digitized output from the source plus a volume code then amplify that using the headset's internal amplifier. It has nothing to do with the device's internal amplifier and speakers.

The Noise Canceling wired ones also use an internal amplifier - thus the battery they need.

On device speakers and normal wired headphones, on the other hand, rely on the amplifier within the device.

So - if you're using BT or NC headphones (anything with it's own battery in the headphone), then the quality of the amplifier and DAC inside the headphones has a lot more to do with your sound quality than those in the device you're connected to. Generally I've found wireless and NC headphones have much poorer sound quality (regardless of source device) than a set of quality wired headphones plugged into the Pandora.

@EvilDragon - glad to see you're making the speakers work as well as possible. It is a proof of the quality and attention to detail we're counting on.

The larger question to me, though... Have you connected a Pyra up to a set of high end (wired) headphones? I know it's much harder to get the quantitative results like you have above, but I'm willing to take a qualitative opinion if you have one.

Great to see things coming together!
 
For mee, a Bluetooth Headphone is still the usefullest in Sport and Gymn,
In Gymn because a Wire is pretty in danger whit all of these Machines..
And at running because then the "Musik Device" can stay in a Waterproof Bag..
Also at sport i dotnt need that great Musik Qualitie..
For better quality, and everything beside sport, i have wirred Headphones..

Im happy, that i dotnt need a short cable extension to use my Headphones at Pyra like on my Pandora..
 
Very nice for a first test! The Pandora sounds kind of awkward after listening to the other two devices :confused: I only use the speakers for games though (that I don't play on any other devices), so I never bothered to compare the quality. The GPD sounds a bit muffled.

@EvilDragon Thank you for your choice of music. You helped me identify a song stuck in my head for a week :D
 
Back
Top