April 15, 2022 at 2:02 pm #24478
AdministratorApril 15, 2022 at 2:05 pm
Here are a few photos of the KABD-4100 and the ICP1 which is needed to program the board. There is some confusion about this, so to clarify the ICP1 is reusable. You can program as many boards as you want to it. Once the KABD board is programmed you disconnect the ICP1 and can program another board if you like. This is very nice for someone that wants to start selling speakers.
MemberApril 15, 2022 at 3:37 pm
That’s good. They seem a lot more capable power and input wise. They seem like a reskin of the wondom/sure boards though, nothing wrong with that though. Annoyingly I had a look and they’re still using the ADAU1701 chips which they were using before, so still pretty incapable for low frequency or high precision correction. Good to see they’re widening their product stack though. Hopefully they start to design some of their own boards, maybe even with some ADAU1452s, now that would be great! It would actually bring some competition to the miniDSP HD!
AdministratorApril 16, 2022 at 10:14 pm
One of my Patreons knows a lot about these ADU chips and said something similar. I do wish they would make a high end version with an Infineon chip and an upgraded DSP chip. But honestly, for this price with Sigma studios is a great deal. By the way, you’re right about them being Sure Amplifier boards. They always have been rebranded. I’m not sure if these are, but I would assume they are. Have you got a chance to use either yet?
MemberApril 17, 2022 at 3:36 pm
As soon as I find a good budget soundcard (for the pi) I’ll have a great solution with the raspberry pi in terms of DSP. I already have something good set up with Moode Audio, but the only inputs for that are web based. Spotify connect, web radio, room, locally hosted files, etc, etc.
MemberApril 17, 2022 at 1:32 pm
@elliottdesigns @123toid – So, I need to flash to see how many instructions this design is over, but this is my most recent design for these boards. I’ll probably be pulling out the FIR filter because not enough instructions to be effective. Also, the low end time delay can likely be removed in entirely. The reason is because the frequency is the longest and that driver is the farthest from the microphone. As such, pushing to later the tweeter and mid-woofer should be enough to get them time aligned. Meanwhile, the same is likely true with the gain modules, as the largest driver is usually the least sensitive. So entering a negative value to volume level will likely allow the removal on one of the drivers to then volume level them.
Couple good links. Basically, the first one explains how to window in REW effectively and how to calculate what frequency is the cutoff for the reading. If you do a windowed and overlay that with a ground measurement, you can then come up with a firm reading of what is going on with the driver.
Then you can use the built in REW algorithms to recommend the IIR PEQ filters to use with them set on using either the ground measurement or the gated measurement frequency crossover point. After you put those corrections in the PEQ filter (after first volume leveling and time aligning the drivers, then creating the crossover), you are pretty well done. After that, you fill in the value for the high pass for the low end, then you use the time alignment and gain at the left of the diagram to volume level the speakers for where you are going to be listening (so you do an average of the seating positions to come up with these values) and time align them to finish it off for in room response.
With all that, there isn’t much left for any FIR. I’ll have to see if a baffle step compensation could fit in or not. But it is a start.
- This reply was modified 2 months, 1 week ago by 123toid. Reason: Edited upload from attachment to photo
MemberApril 17, 2022 at 3:45 pm
Yep, definitely not capable of FIR these aren’t. Also, when programming with the miniDSP (non-HD) that uses the adau1701 as these boards do; I found that there was no significant accuracy below 200hz which isn’t particularly useful. However, can’t complain about the performance for the simplicity and price though.
MemberApril 17, 2022 at 3:47 pm
Btw, good schematic, I like the idea of using the limited FIR capabilities for fixing crossover phase. 👍
MemberApril 17, 2022 at 8:52 pm
BTW, some resources on how to generate the PEQ filters from REW. Figure it is relevant for those using REW instead of omnimic and that may not understand where to go to do it.
AdministratorApril 17, 2022 at 10:08 pm
MemberApril 19, 2022 at 7:51 am
What I plan on doing and handing to you and Erin at Erin’s Audio Corner is to create a step by step PDF of how to tune a speaker using the above WIP schematic, REW, and a Umik or similar calibrated mic. This also will cover what can be removed after you have the speaker mostly tuned (mentioning everything that can be removed to free as many taps as possible to do a phase correction, hopefully). Obviously I will have citations at the bottom to these different sources so people know where it came from.
But right now I am working on something else. Need to get one more MDF 3/4″ board (bringing it to 16 sheets of 3/4″ mdf and 2 sheets of 1/2″ mdf).
MemberApril 19, 2022 at 8:11 am
That looks great! And you lucky sod with all that MDF, I’m so jealous!
I’m currently in the middle of making 4 videos that go through doing REW EQ for rePhase and rePhase itself for FIR files, so hopefully that stuff should help too when I release it. I’ve been using rePhase for quite a long time now so hopefully I can explain it’s use in a way that’s easy to understand 🤞
Any chance you could upload the PDF to the forum, I have a feeling there will be a lot of people new to this from Nick’s new video looking at for a good tutorial on how to do the schematics, and you certainly seem to know your stuff with them!
MemberApril 19, 2022 at 11:48 am
Once I have it made, I gladly will. I just haven’t had time yet. I plan on taking screenshots to insert into the PDF so that people can see clear menu by menu how to get to the settings (One issue I have with many guides, even with pictures, they do not always make clear where the setting is or how to get to it, which then takes up a large portion of time trying to figure out how they got there). I don’t have the manual made yet, but once I do, I’ll tag you on the upload.
And yeah, that is enough MDF to build 6 tower speakers and 1 sub. I’m also going to see after I’m done if I have enough extra scrap pieces to make a prototype bookshelf. That bookshelf will use an SB Acoustic SB26STAC tweeter paired with a Scanspeak 18W/4434g-00.
But, since I missed moving some of the pieces I needed over to the spreadsheet and cutlist, I’m currently redrawing the models in Inventor and making all last minute changes to the design to finalize it. That way I make sure that everything is there for me to start. That is about $1K in MDF already…
MemberApril 19, 2022 at 12:07 pm
Yeah, I’ve yet to see a good sigmastudio guide for speaker DSP. Looking forward to seeing yours when made, also no rush, these things take time! Blimey that’s a lot of money in MDF, almost half as much as what I spent on my car 😂
MemberApril 19, 2022 at 12:32 pm
I’m going to make it for the JAB5 to start. Then modify it as needed for other sigmastudio devices as time goes on. The design portion is the same, but the hardware settings varies and shouldn’t be fussed with for the most part (due to risk of making something give up the smoke).
But I’ll explain going through to do impulse response in REW and how to step that in sigmastudio to get it aligned. Once aligned, how to use the mute function to mute one drive or the other to get a clean reading on volume (also because those will matter for figuring out the first point to start doing a crossover). After you have volume leveled, I recommend to start over and check the impulse response again and then check the volume again if impulse response was adjusted. Then it is picking the crossover point and doing iteration on selecting the right point and trying to prevent any aberrations. From there, you can do the setting of the high pass on the low end of the woofer or sub. That is a balancing of trying to get a lower frequency, but also to prevent low end excursion and distortion (if you prevent it getting close to x-max, then the driver will act more linear, which can help reduce distortion, which in turn can help with compression issues when playing louder, etc.). If doing a 2-channel bookshelf or if you have an external sub, you can just set the low end to where you are crossing over and so long as that is not at x-max, you are done (like an 80-120Hz high pass). After that, I move to setting up the PEQ filters, which should in theory be straightforward.
After you have done all of that, things that can be removed are any mute buttons (you cannot use them outside sigmastudio anyways, so when finishing it to close everything into the enclosure, you do not need that anymore), any invert buttons unused (mainly there for convenience in the event you wired it backwards so you don’t have to fix that), probably the gain for the lowest dB driver, the time alignment of the largest driver that is furthest away from the tweeter (unless a special case where you couldn’t get it aligned, so you adjust all looking for a common subdivision to get all drivers maximally aligned). Also explain the left and right and to mark the speakers so that you don’t confuse which speaker is getting which channel.
From there, it is flashing it with slowly increasing the number of taps on the FIR filter (I start with 50s, then 20s, then 10s). Once you know how many you can add, you put that number in RePhase, then tell it to try to create a flat phase. Whatever it kicks out, you move that over to sigmastudio (there is a text file thing on formatting, but I forgot the details at the moment).
Then, it is putting it in your room, using masking tape to mark out placement as you may have to move the speaker to access sigmastudio again. You do the impulse response again for left and right channels from the listening area. Once you determine which driver is off on the delay, add that delay to get the impulse response aligned. And, finally, volume level both speakers with the last gain setting in the schematic to get the speakers playing at the same sensitivity.
Viola! That should give you roughly the best experience you could get with an ADAU1701.
MemberApril 19, 2022 at 4:11 pm
Sounds great. For active crossovers I go off of the same principles as used for passive (based on enclosure design as well as C-C driver spacing). The only difference is being able to use a steeper and therefore lower in frequency crossover which helps with driver integration, that and being able to use linear phase crossovers, that’s what I use active crossovers for anyways. For calibration of active speakers, I apply the crossover and then measure each driver in rew separately, then use bass limited and subwoofer settings to get the EQ in REW done right. We definitely have different methods, but yours is more suited towards lower capability (short FIRs) like these boards have with their ADAU1701 chips, whereas mine is more suited to not having to worry about FIR lengths at all, like the the Pi stuff I’ve been doing. I definitely think this will be an amazing resource for those getting into these ADAU1701 boards!
MemberApril 19, 2022 at 4:44 pm
Just wait for my ADSP-21489 design and guide! That will be FIR driven for a three way (like my tower speakers I am building)…
MemberApril 20, 2022 at 7:28 am
How many taps?
MemberApril 20, 2022 at 7:36 am
Hold on a minute… That’s the chip used in miniDSP 2×4 HD isn’t it? Haha, you’re effectively open sourcing miniDSP stuff. They’re gonna be mad if this stuff gets big, all they’d be offering at that point is ease of use really, assuming it’s cheaper to DIY it that is. Oh no, just thought, would you need a USB-i for that? Those are really quite expensive.
MemberApril 20, 2022 at 8:37 am
So, the Chinese board I bought was about $150 with the USB-i, but you had to solder your own headers on for the USB-i. Meanwhile, if people care about ease of use and don’t want to do it that way, they could actually buy the miniDSP 2×4 HD and just remove the shell and wire it up inside of a speaker. Then it is just mounting connectors and extensions cables to the back.
I would have to search for the specific ones I bought at lower price, as I got that board with the USBi for $154 (I have 3 USBi and 4 boards with the pcm4202 and pcm1798).
Now, I checked out the software that came with them, but it doesn’t allow for FIR filtering, but otherwise can do mostly what you would want.
But once I get that made, I’ll also show that.
MemberApril 20, 2022 at 9:16 am
Weird that they’re selling 3 USBi to go with 4 boards. I would’ve just sold 1 with 4 since that’s all that’s needed. You’ll only ever need 1 USBi after all.
MemberApril 20, 2022 at 10:13 am
The first two boards I bought, I paid $184 for 1 to have the USBi and $145 for no USBi. For the second two, bought at a later date, I was able to find for $154 the board, PSU, and USBi thrown in. So getting extra USBi, while also not paying much more than the one I previously bought without it, it was a deal.
So that is how I got 3. Meanwhile, I am putting 1 of those boards per speaker for the 3-way towers.
MemberMay 23, 2022 at 12:54 pm
So, I tested flashing the 3-way schematic. Haven’t tried the two one yet. Wrote this guide yesterday. Feel free to look through it. Also, check the registers are correct for your device. But, this should give good performance with 15 PEQ filters and 480 taps for the FIR filter (very limited, but those PEQ filters). This is also 24dB/Oct for the infrasonic filter and 48dB/Oct for the crossover. So very powerful double precision filters.
MemberMay 23, 2022 at 1:14 pm
With this small amount of taps, I recommend if you do phase correction. Don’t do linear phase initially, go with your typical minimum phase filters, then correct just the phase using the FIRs as much as possible, that way you can get your taps stretched as far as possible in their usability. I’ll have a look at the guide in a bit. 👍
MemberMay 23, 2022 at 1:23 pm
Great stuff, I had a flick through. I wouldn’t recommend going above 18dB/Oct, 24dB/oct max (especially in higher frequencies where we are more noticeably sensitive to phase change). Or at least not without crossover phase compensation to make it linear. The phase change otherwise is just too high to be worth the extra steepness. If you do use the steeper slopes with minimal taps compensation we have available here, I’d recommend that this should be alright (but not ideal) if the phase was just linearised to 90° or less anywhere above -40dB from the fundamental.
- This reply was modified 1 month, 1 week ago by elliottdesigns.
MemberMay 23, 2022 at 3:30 pm
Thanks for the advice. That is why I had hoped to write it as I was exploring it with my projects, that way I could check how well it worked together, but slower progress than I wanted on building the speakers and a snow day resulted with me inside writing this manual instead. I will incorporate the points made about how steep and how that effects the phase, along with the other elements I talked about in the other thread.
I’ll also need to fill out more on how to use rePhase. But I was lazy and just wanted to get the draft done.
Thank you so much on the notes!
MemberMay 23, 2022 at 4:16 pm
No worries, those notes on rePhase were just directed for this chip though. Not general notes. Such few taps even with small amount of correction will most likely cause pre-ringing, and may unfortunately be audible. I suppose whoever uses it really just needs to make sure they don’t use it for low frequencies
MemberMay 24, 2022 at 9:51 am
So are you saying it’s only 480 taps when running at 48Khz, at 44.1Khz, you can’t run as many?
MemberMay 24, 2022 at 11:58 am
I’m saying that there should be an insignificant amount more going to 44.1K over 48K
MemberMay 24, 2022 at 2:09 pm
Ah, I again see the confusion. I was on about rePhase again. Having the sample rate in rePhase (and the chip too, since they need to be the same), means that the given amount of taps reach much further into the lower frequencies. In my video I demonstrate this, the difference in low frequency filtering performance is actually rather staggering.
MemberJune 5, 2022 at 12:35 pm
Well, I just revised the two designs for the three way and two way crossovers. I added a noise gate and slightly reduced the already very limited FIR filter. But, making sure noise is removed will help more than the phase filtering in this use case. I chose -80dBFS, but considering the signal detection threshold is set to -60dBFS threshold, you can play with both of those to figure out the best balance and agreement. After all, if the speaker trigger is at -60, you might want to filter all signal less than -60dBFS, which is signal at less than 0.1%, which is around the noise floor for the amp unit anyways.
I chose peak for the noise gate instead of RMS to fully limit noise at that amount rather than an average. If you are going to use the compressor to also limit the volume on the high end, then that module needs replaced with the RMS filter as the RMS will allow transient amounts on the top end to exceed the max value set, which can help with dynamics of the audio playback. For the ADAU1701, there isn’t really enough to do RMS and Peak, unless you want to further reduce or eliminate the FIR filter, which you could also do. On a two way, that is easier.
If you are using 1 board to do 4 outputs to 2 bookshelf speakers, then eliminate the FIR filter and copy the rest of the signal chain for the other two channels.
Edit: Made a four-way and dropping it here. made the extra changes so that one volume know would control both speakers and updated the output channels, as that would have been a source of confusion for people.
The four way is made for speakers where the person sets them up to be able to bi-amp the speakers. By setting it up to bi-amp, you can do the control of both speakers externally and you can mount the KABD inside an enclosure by itself, separate from the speakers. And that allows for the speakers to be usable with other external components sometime in the future.
- This reply was modified 3 weeks, 3 days ago by ajc9988.
MemberJune 5, 2022 at 1:31 pm
Great summary! At this point I think we should have a separate topic for these finished schematics/guides. Then maybe you could just edit it when replacing with a newer version. I think it might make it a bit easier for outside sources to find maybe? Anyways, just an idea.
MemberJune 5, 2022 at 2:04 pm
I’ll do that probably later today or tomorrow. But great idea.
MemberJune 10, 2022 at 6:45 am
What is the best way to add HDMI or optical input capability?
MemberJune 10, 2022 at 8:46 am
Probably some kind of soundcard for optical to RCA. There’s plenty out there. The only way I can think of extracting hdmi audio right now is with a HDMI audio extractor.
ModeratorJune 10, 2022 at 9:35 am
I know there are HDMI to RCA / audio out extractors, and I believe I’ve seen optical too, just not sure of whether it was TOSLINK or SPDIF. When in doubt, check AliExpress and you can usually find what you are looking for. That way, you can at least get a name for the item.
MemberJune 11, 2022 at 12:04 pm
How many channels are you trying to extract? What level of HDMI is needed? Are you needing HDMI pass through with audio extraction, or do you already have the HDMI split? Or are you trying to use ARC to then go to an HDMI extractor to then feed to speakers?
I apologize, but I would like to know more information on what your planned setup is in order to even be able to search for an audio extractor that meets your needs. Additionally, is this going into a speaker like a soundbar? If so, then you can shell the audio extractors or might look at a finished product like that at parts express. For example, this soundbar kit. https://www.parts-express.com/Dayton-Audio-SBA302-BT-Sound-Bar-Amplifier-Kit-with-BT-and-IR-Remote-300-880
So I agree with the others, but would need more info to give a solid recommendation for devices with specific features.
MemberJune 12, 2022 at 3:38 am
Thanks everyone for your input,
Sorry for not including enough info.
<div>I am trying to build 2 way bookshelves. The idea is to use all 4 outputs for each driver.</div><div>
I am trying to conceptualise it all in my head, I haven’t started the build.
Need to research how my samsung tv works, I have a soundbar connected to it at the moment but not sure if the splitting is done at the tv or soundbar level.
I had a look at the Dayton soundbar kit plate. That one has 30W per channel, I am hoping to build something with a more whizzbang. Also I would love to play with the KABD’s DSP to see what kind of a result I can get.
MemberJune 12, 2022 at 4:42 am
If it’s stereo as you say, and you have no RCA outs you can use, the easiest is probably optical, then HDMI. Any old audio extractor that outputs to RCA should do the trick, then each of those RCAs into its own amplifier. HDMI option will only work if using an ARC port of your TV (some HDMI ports should say ARC, audio-return-channel). If you don’t care about using your TV for broadcasted content etc, and are just using it as a display, e.g. you have a device outputting hdmi into your tv, there are splitters that intercept the signal rather than you having to use the arc port on your TV.
MemberJune 12, 2022 at 6:47 am
To build on that, if you are planning to buy the KABD-4100, that board has the amplifier built in. There are a couple different ways to accomplish this. If your TV doesn’t have RCA outs, sometimes it has a headphone plug. If you buy a cable that splits out from the headphone to two channels, you can then buy the RCA termination and wire it to go from one to the other. Double check what input plugs come with the KABD-4100 first. I’ll dig out the accessory pack later today. But you can wire up the board to whatever termination you would like, including RCA. You then would just wire the output from each channel of the KABD-4100 to each driver.
Now, if you do not have headphone out or RCA, then I agree with @elliottdesigns that toslink/spdif would be next. since you are doing stereo, practically any hdmi audio extractor should do it. And he is correct, if you do not have a receiver or similar, then you have either the ARC (audio return channel) on the TV (check the manual) or you will need a splitter and then one goes to the TV, one to the audio extractor, then to the KABD-4100. Also, if you want to use a remote for the volume, then you would need to get a device to put in-between the extracted signal and the KABD-4100. You can only control the volume over bluetooth when you are streaming bluetooth music to the device that I know of (playing with the JAB5 from Sure/Wondom). That will act like a pre-amp. An Arylic Up2Stream with a remote and RF extender module should work wonderfully for that purpose. Also, you can connect that to the internet and stream to the speakers that way, allowing for more use than just for TV. But that uses the plug termination like for headphones, so you could do a headphone out from a tv to the arylic, then that headphone out over to the KABD-4100, then out to the speakers.
Since you haven’t started yet, take your time. Do not rush this. Pick out everything that you would need or want in the audio chain. Research each type of termination that is used (later today, I’ll tell you what the accessory pack includes for input for the KABD-4100, but I can guarantee it is not RCA in, so you would have to wire it for that). After figuring out every device, look up wires and terminations on amazon or whatever other retailer where you can find the terminations to do that with. Decide if you will buy something to house the electronics or 3d print or how you want to do that. Because that will help narrow down thinking how you will have to mount the terminations to wire it all up.
Once you know every termination, everything that needs wired, how you will mount it, the enclosure, etc., then buy the components. But don’t buy first then change your mind like I did. I have a couple JAB5 I’ll use in future products now, but that is a couple hundred sitting idle. I might put them in some bookshelves for my nieces in the future or something.
Hope that helps a bit.
Also, the splitting happens at the TV, generally. That then gives you PCM or other standards out, or uses the ARC. That way you know, the sound bar isn’t splitting it unless you plug into the soundbar first and then the soundbar to the TV. I don’t think there are many products that do that these days, but could be wrong. Many soundbars with HDMI in use the ARC from the TV. But check your manuals, it should tell you.
- This reply was modified 2 weeks, 3 days ago by ajc9988.
MemberJune 12, 2022 at 7:53 am
As @elliottdesigns said, and going by the attached photo I will put the HDMI in to the too hard basket for now.
Better option would be the optical, I am concerned however about converting the signal from digital -> analog -> digital -> analog. Not sure if that will have an effect of the end signal quality.
I still have the option of using the bluetooth, however the concern there is the connection reliability (although I’ve heard that BT 5.0 is quite reliable), and the need to use time delay when watching TV to sync the picture and sound.
Arylic should work as they have the S/PDIF IN expansion board that does exactly what I need. The problem is they only offer 50W per channel @ 4Ohm, and the DSP looks not-so-advanced compared to the KABD’s sigma studio, I might be wrong though, not sure.
Thanks @ajc9988 for the good advice about finishing the build in my head first before I make a hole in my pocket. I will have fun though no matter what I decide. I am starting to think that maybe buying both the Arlyc and the KABD and see how they sound, play with the DSP’s and assess the amps etc, before I make the decision on what to use and finish the build.
MemberJune 12, 2022 at 1:36 pm
So, I was actually meaning this:
The up2stream pro v3 is a pre-amp board. It is not an amplifier board. You can put this in between the tv and the KABD-4100. You can buy the additional S/PDIF in or out, along with a DAC if you want.
You are right to worry about the conversions, as each component may have a certain noise level. Now, you can noise gate with the KABD so that any noise below that point is chopped off the signal before amplification. Can’t fix the noise from the amplifier itself, though. So keep that in mind.
For example, the first one is an HDMI passthrough with audio extractor for a stereo out using RCA. Just put this in between your set top box or cable box and the tv, then feed the audio into the arylic I linked using RCA to headphone jack, then use the headphone to headphone to link to the KABD-4100, then out to the speakers.
For the second one I listed, it has 5 HDMI in, 1 HDMI out, and headphone stereo out. Just do headphone terminal straight line to the Arylic, then same cord to feed to the KABD-4100, then to the speakers. Done.
Now, if you need 4K@60Hz and HDCP 2.2 support, here is a product that claims that.
And when I say headphone, I mean 3.5mm Aux stereo out. But you get it. Make sure whichever audio extractor you use, you check the features, such as resolution and refresh rate, output terminals, HDCP support (more important than you think, and the avedio are some of the few listing that support; if you don’t get an HDCP handshake, some video will not playback).
Further, if you do go with the Arylic for a setup like this, I’d use the accessory DAC as the KABD-4100 doesn’t have a digital in. That port is used to link together two KABD-4100s to build an 8 channel audio situation. So you would have to have a DAC to switch the digital internet content over into analog to send out the 3.5mm Aux port.
On the expansion boards, I’d grab the DAC and the IR extender, then pickup the remote.
What you are basically doing is building a pre-amp receiver. You have the HDMI audio extractor feeding into the Arylic as a pre-amp, which can also stream audio from your internet digital services, which then converts the digital to analog for output or does passthrough output to the KABD-4100 DSP and Amplifier combo. You then can use that remote to adjust the volume of the entire chain from across the room.
So, you are building your system around your required specs, rather than buying a receiver that does it, but might only have some of the features you want and many you don’t which increase the price to buy a receiver.
Hope this gives you more information.