Enginerding with Weeb Labs

[sq225917]

The very first to camera piece seems out of sync, the rest is fine. The voice is weird though, it sounds booth mic'd with no reflection, it's totally diffuse, like it's been done with ANC.

 

[Weeb Labs]

The very first to camera piece seems out of sync, the rest is fine. The voice is weird though, it sounds booth mic'd with no reflection, it's totally diffuse, like it's been done with ANC.

A couple of people have mentioned seeing a sync issue but it is really just an illusion generated by the combination of my tendency to lead mouth positions for optimal pronunciation and the contrast between my voice and appearance. If you watch for the "P" in "output" during that first shot, you will see that it aligns correctly.

The office/HT room is very well treated and the microphone (hypercardioid) boomed from just above, so the recording is very dry and "intimate", with a low noise floor. It was my intention to emulate a large diaphragm condenser without actually having one in frame.

 

[Weeb Labs]

I stumbled upon this little gem earlier today and since then, so many fun implementation ideas have sprung to mind. More detailed information can be found in the family datasheet.

If this IC's performance even approaches what is claimed in the datasheets, then I am utterly perplexed as to its apparent lack of appearance in a single consumer product.

TL;DR: This is a lossless wireless audio transceiver with up to four PCM channels (two I2S), native USB Audio, bidirectional audio, a bidirectional data sub-channel, USB HID emulation of a mouse and keyboard, various GPIO, a simple pairing system with up to four receivers in each network and 10ms of end to end latency (depending upon configuration).

I have already ordered a batch of PCBs and look forward to experimenting with these next week. If anybody has experience with this IC which they would be interested in sharing, then please do!

 

[Weeb Labs]

This evening, I completed one of my lengthiest vintage radio restorations to date. It is a customer's set and so much has been done to it over the course of the last six months. Here is a photo taken when I first collected it from the customer.

And here is what it looks like today.

This is a list of work which I have carried out.

• Complete capacitor and resistor replacement
• Volume/tone potentiometer replaced
• Dial artwork recreated from scratch and printed to a new pane of glass
• New LED dial lighting (emulates original appearance, with greater uniformity)
• Speaker cloth soaked in petroleum and washed to remove discoloration and warping
• Cabinet stripped, treated for woodworm, filled and then re-veneered
• IF alignment
• All front metal parts sanded and polished to a mirror finish (knob metal included)
• Band select buttons whitened
• Magic eye tube replaced with my own replica (almost indistinguishable from the original, with a much longer lifespan)
• Chassis cleaned, sandblasted and polished

I am quite pleased with the outcome. Time to get this back to its owner!

 

[sq225917]

More details about the glass printing process please, that looks mega.

 

[tomtoo]

Could you tell more about
"....Magic eye tube replaced with my own replica (almost indistinguishable from the original, with a much longer lifespan)...."

I love this tubes.
Maybe thats why i always read weed labs?

 

[jtwrace]

@Weeb Labs No out of sync here at all and I can't believe that's your natural voice. Very fortunate!

 

[Weeb Labs]

Could you tell more about
"....Magic eye tube replaced with my own replica (almost indistinguishable from the original, with a much longer lifespan)...."

I love this tubes.
Maybe thats why i always read weed labs?

The replica consists of a 1.5" display and an ATTiny85 on a PCB with a bridge rectifier and LDO. The AGC grid feeds one of the ATTiny85's ADCs through a variable divider and the board is powered by the 6.3V filament winding. The panel below is an RGB OLED but I used an IPS LCD for this particular radio.

The code is fairly straightforward and just maps the ADC value to the shape rotation. Simple but emulates the original tube very accurately. There is also a 1/16 diffusion gel in front of the panel, which softens the image slightly and hides any otherwise visible pixels.

 

[Tks]

@Weeb Labs

Are those true RGB matrix? Or are they silly nonsense that LG uses ala WRGB?

(btw nice work on the BFI a while back, when I saw the result I was stunned). I'm here waiting for 480Hz monitors to see if they're going to get close enough to CRT or will we need to wait for the predicted 960Hz to finally put CRT motion handling behind us as something we've surpassed..

 

[Weeb Labs]

@Weeb Labs
Are those true RGB matrix? Or are they silly nonsense that LG uses ala WRGB?

Yes, it's a full RGB matrix. The datasheet is here. Not that the subpixel layout is particularly important in this particular application.

(btw nice work on the BFI a while back, when I saw the result I was stunned). I'm here waiting for 480Hz monitors to see if they're going to get close enough to CRT or will we need to wait for the predicted 960Hz to finally put CRT motion handling behind us as something we've surpassed..

Thank you! With the current batch of 270Hz ULMB IPS panels, I think we are more or less there. Scanning backlights might be an interesting development and is already technically achievable with the current FALD backlights.

 

[Tks]

Yes, it's a full RGB matrix. The datasheet is here. Not that the subpixel layout is particularly important in this particular application.


Thank you! With the current batch of 270Hz ULMB IPS panels, I think we are more or less there. Scanning backlights might be an interesting development and is already technically achievable with the current FALD backlights.

View attachment 152381

So one thing about IPS. That can go die in a ditch (and all the backlight bleeding to death IPS implementation of garbo gaming company monitors on the market). Appreciate the viewing angles (still not better than Plasma tho), but hate everything else (color gamut can be had on VA just fine).

Also to the dumps with FALD until we get to 10,000, hell 50,000+ zones at least. That godforsaken blooming on flagship implementions (regardless if it's a TV or monitor). Just no please.

Now, for a sane reply...

That image you posted, it's like no actual motion blur, but looks exactly like inverse ghosting you get from heavy overdrive. But the kicker being, this isn't inverse ghosting, it seems it's ghosting in the opposite (more appreciated) direction. I've never seen this. Usually that sort of ghosting as you know happens ahead of the thing in motion, not trailing it like yours is here.

What the heck has your mad scientist self done? We need to get you making some of these displays for people.

And lastly about the OLED matrix. It's simply nice to see that the full RGB matrix isn't extinct outside of reference displays from Sony. I wonder where the supply of these small displays is coming from seeing as how LG would rather kill themselves than even attempt to offer a display with full RGB as the matrix. I'd want to say just failed parts off the factory floor while making a reference display that failed? So it gets cut into little pieces like this and sold (since you can't sell a reference display even if it has a flaw as small as one of those little things). Who's makin this stuff lol?

 

[tomtoo]

The replica consists of a 1.5" display and an ATTiny85 on a PCB with a bridge rectifier and LDO. The AGC grid feeds one of the ATTiny85's ADCs through a variable divider and the board is powered by the 6.3V filament winding. The panel below is an RGB OLED but I used an IPS LCD for this particular radio.

The code is fairly straightforward and just maps the ADC value to the shape rotation. Simple but emulates the original tube very accurately. There is also a 1/16 diffusion gel in front of the panel, which softens the image slightly and hides any otherwise visible pixels.

View attachment 152349

Man, thats advanced.

 

[Weeb Labs]

I have recently been playing with the idea of capturing impulse responses from automated room correction solutions such as Dirac Live and ARC, using a REW loopback.

Those impulse responses can then be windowed to a sane number of samples, exported as a WAV or text file and loaded into your preferred convolver or DSP.

This effectively enables one to export their Dirac Live or ARC correction to any desired device and correct multichannel configurations using DRC packages which would ordinarily support only stereo. A common use case might be system-wide correction under Windows using Equalizer Apo, which lacks sufficient VST compatibility to load Dirac Live or ARC.

If you were feeling particularly malicious, you could even use this to take advantage of the Dirac Live trial period and export a permanent copy of your correction.

 

[JayGilb]

This evening, I completed one of my lengthiest vintage radio restorations to date. It is a customer's set and so much has been done to it over the course of the last six months. Here is a photo taken when I first collected it from the customer.

View attachment 152243

And here is what it looks like today.

View attachment 152244

This is a list of work which I have carried out.

• Complete capacitor and resistor replacement
• Volume/tone potentiometer replaced
• Dial artwork recreated from scratch and printed to a new pane of glass
• New LED dial lighting (emulates original appearance, with greater uniformity)
• Speaker cloth soaked in petroleum and washed to remove discoloration and warping
• Cabinet stripped, treated for woodworm, filled and then re-veneered
• IF alignment
• All front metal parts sanded and polished to a mirror finish (knob metal included)
• Band select buttons whitened
• Magic eye tube replaced with my own replica (almost indistinguishable from the original, with a much longer lifespan)
• Chassis cleaned, sandblasted and polished

I am quite pleased with the outcome. Time to get this back to its owner!

Beautiful work on the restoration.

 

[Weeb Labs]

I am currently developing a new universal ASRC board, based upon the SRC4192. It functions somewhat similarly to the Twisted Pear Metronome board but will be sold for a fraction of the price (around $25) and supports TDM configurations.

I don't yet have a webstore of any kind but it will soon be available to those interested. Might have to seek a manufacturer badge.

Actual size.

 

[Weeb Labs]

The first revision Constellation boards are now being manufactured and I hope to receive them on Monday. This is what it will look like.

In the meantime, I have cleaned up the lab. Expect it to remain tidy for all of five minutes.

 

[Weeb Labs]

They're here and they really are tiny! Now waiting for my component order to arrive.

 

[JayGilb]

Maybe I missed it, but is there a end product that will be using this sample rate converter or is this a stand-alone product ?

 

[jtwrace]

Maybe I missed it, but is there a end product that will be using this sample rate converter or is this a stand-alone product ?

Stand alone I believe. It's meant to be like this.

 

[Weeb Labs]

Stand alone I believe. It's meant to be like this.

Exactly. Similar in function to that board but much smaller and with TDM support.