From the original SynthMaker forum (2004–2014), with its members’ names and dates.
monx ·
Hey all. I wanna report a problem I've observed, and haven't been able to put a name to.
So I'm building my little synth in SM and all is well until I add my volume module, and happen to set my sustain volume really low. I could hear some intermittent buzzing an octave below my oscillator.
When I increase the volume gradually from 0 to 1 as I'm holding down a middle C, I can hear this buzzing come in and out at a few specific volume levels.
So I ignored this since it wasn't very loud and moved on with the project. That is until I was working in Reason last night, using the NN19 to generate a sampled saw wave, and noticed the same artifact. I don't know why I never noticed it before, but then again, Reason's heavy aliasing also never bothered me until I started building with Synthmaker...
Is this all just "quantization distortion"? Would dithering fix it?
I am attaching sonogram and sound sample of the distortion produced in Reason. (Forgive the noise on the audio. I had to boost the sample since the effect is only prominent at a low oscillator volume.)
MyCo ·
is this recorded or rendered? Don't think that this comes from digital processing, sounds more like a feedback in the analog part of the soundcard.
monx ·
It's recorded. But good point, I should render.
infuzion ·
What resolutions are the Wavetables? Or how are the sounds being created specifically?
MegaHurtz ·
I think it sounds a little like your processor rattling away.
infuzion ·
MegaHurtz wrote: I think it sounds a little like your processor rattling away.
Maybe your CPU gears need more oil.
😈
Actually, I think SM's wavetables act kind of weird in low octaves, but I always assumed it was CPU-cheap the algorithms. Or if it is your soundcard, perhaps you have EMI/RFI issues? Or perhaps your sound card is so good you notice it, but our cheaper soundcards/speakers/wires don't.
MegaHurtz ·
need a little butter 😀
MegaHurtz ·
Actually you can always hear this.
I used to do hightech for a living.. There every processor output got secured with a filter to enshure peak voltage
is kept to a minimum. Else all the nearby signal paths get resonated. And you get a jittery signal.
MegaHurtz ·
What soundcard are you using?
I really want to know becouse most of the others are highpass filtered, not that that`s a bad thing.
But this one lends itself to eeg. It`s a great measurement tool 😀
monx ·
Oil and butter, all very good suggestions, but I think I'll try the dithering stuff first.
I tried exporting the audio from Reason and the output is not distorted at all. That surprised me. It makes this a little more confusing.
I am skeptical that the distortion has anything to do with electrical interference, analog output of the soundcard, etc. It's definitely a "digital" distortion.
monx ·
Update: Ok, I figured it out. It was really stupid.
The "10-Band Software Equalizer" on my Realtek HD Audio driver was enabled. Who knows what other garbage artifacts I was hearing because of this. 😕
I wish I could bypass the entire effects capability of this sound driver and get a little lower latency, because the best I can do is about 30ms... It's the ALC885 codec.
http://www.realtek.com.tw/products/prod ... ProdID=138
trogluddite ·
monx wrote: little lower latency, because the best I can do is about 30ms
Sorry if this is a dumb question, but have you tried the free ASIO4ALL driver? Just search for it on the net - it's totally free, and can really help lower latency with on board sound chips. It's not guaranteed to work, but I know a lot of people round here use it for soundcards without their own ASIO capability.
The Luddite
MegaHurtz ·
The rendering is becouse your processor doesnt get output in the proces.
This is a confirmation that it`s definately your harware.
this has nothin to do with you dithering 😀 unless yr not exact about what you are doing it`s useless.
MegaHurtz ·
What unrounded bit you wanted to lift in a random fashion?
Ps and be shure to design a filter shaped like a gaussian bell.. rbody uses powR these days...
monx ·
trogluddite wrote: have you tried the free ASIO4ALL driver?
I do use ASIO4ALL, and it makes a big improvement. The 30ms I mention is what I get with ASIO4ALL.
On my other machines I get better results with ASIO4ALL. I just wonder if the latency could be improved by a better driver.
I used kX Project with my SB Live cards and that was an amazing driver, I tell you what. I wish something similar was available for all these Realtek chips soldered on motherboards. They are everywhere.
And MegaHurtz... problem solved, but thanks; you were right, it wasn't dithering.
MegaHurtz ·
asio is asio, the latency is due to the buffersize and speed.
infuzion ·
monx wrote: Update: Ok, I figured it out. It was really stupid.
The "10-Band Software Equalizer" on my Realtek HD Audio driver was enabled. Who knows what other garbage artifacts I was hearing because of this.
Makes me wonder how much the "Speaker Settings" change the sound also.
Something worth while to put on the WiKi.
trogluddite ·
monx wrote: I do use ASIO4ALL
MegaHurtz wrote: asio is asio, the latency is due to the buffersize and speed
So sadly, 🙁 your latency is caused by buffers inside the hardware of the sound chip itself. A new soundcard will be the only way to make your latency lower - one with native ASIO drivers and no built in FX/EQ. The good part is that a well designed PCI/PCIe or USB soundcard will give you better audio quality too (lower noise, more accurate sample clock) - in most cases a PCI or PCIe one will perform better than USB (USB controller chips also use a little bit of hardware buffering)
The Luddite
sunsynth ·
Yup, and 'No multi-client with ASIO4ALL'
Have a look that the ASIO driver that comes with the soundcard supports multi-client!
MegaHurtz ·
the internal soundcards are more sensetive to interferance on the other hand.
monx ·
Ah. You guys are telling me all this bad news I do not want to hear. This is an integrated laptop sound chip for Macbook (I'm running Windows), not something I can upgrade. Also I don't like USB dongles.
Here is my Macbook's sound card. http://realtek.info/pdf/ALC885_1-1.pdf
I would be thrilled if someone can explain this datasheet, and point to the hardware buffers that limit latency.
I have a few questions about this stuff. What exactly does 'codec' mean in this context? Is it synonymous with 'sound chip'? Because I have always known 'codec' as synonymous with compression/decompression algorithm.
trogluddite ·
Codec just means COder/DECoder. Although we're talking hardware here, the concept is just the same as the software codec used by Windows Media Player (mp3,wav,...). It's really just a definition of how a sound is encoded as binary digits for communication between your CPU, the PCI bus and the chip itself.
Without studying the timing diagrams in a lot of detail, I can't easily pinpont the buffers - but the sheer comlplexity of the schematic is a big clue to the high latency. Just as with SynthMaker, the more complex your schematic, the more processing power is required. To keep the cost down, the chip will not have a lot of CPU power, so higher latency is needed to stop the audio from screwing up. Each of the different signal paths on that diagram will need a small buffer of some kind, and they all add up.
Sad thing is - all that stuff is there just so you can play war games on your PC with 5.1 surround sound/environment settings/EQ etc. etc. As a musician that stuff is all useless - but even with it all turned off, the complexity and latency remains.
High quality soundcards are little more than a digital to analogue convertor connected straight to the PCI bus - the cost is in the quality, not the quantity of electronics.
As a live musician, I share your pain 🙁 - an entire music studio in a little bag is such a cool idea - but the manufacturers of laptops think we all use PCs for the office, shooting monsters and dowloading porno, so they won't give us the simple things musicians need.
The Luddite
monx ·
Thanks trog. I'm not giving up yet, though. I'm still hanging on to the idea that Realtek could write a low-latency driver for the chip if they were inclined. Notice the part in the data sheet that lists all the software effects.
Wrote something here? To have a post removed or your name hidden, write to [email protected].
SynthMaker