Just testet the latest algorithm with different audio tracks.

Just testet the latest algorithm with different audio tracks.
I didn’t adjust any animation parameter while clicking randomly through the music.

What do you think?

are you still using the MSGEQ7?

Yep. 100 fps because of the slow OLED and slow leds.

it looks snappy on the video, i guessed a Teensy. Cant wait to see it finished Viel Erfolg!

Danke. Not sure, if it will ever be finished. But anyway another puzzle piece that will be used for the “Funky Light Synthesizer™”. :wink:

Looks pretty fun. Nice and reactive.

awesome, I saw it live.

This is the point where I seriously need Faster leds! Then every last little detail of the sound will be visible. 100 fps are nothing to show the potential this animation has.

Looks great! Using WS2812? Going to try it with APA102?

Thanks. Yep. Yep. :slight_smile:

Regarding https://en.m.wikipedia.org/wiki/Frame_rate the human eye can see upto 150 fps. In cinemas we have 24 fps. TVs do around 50-60 fps. How useful is it to drive your animation at 400 fps? :slight_smile:

Why do good TVs upscale the video to 200 (or even more) fps? Because it makes movements incredible smooth and soft. One could even do temporal dithering for better color resolution. The more fps the better. 400 is anything but senseless. If I only read audio and write leds I get 120 fps which looks (and feels) way different then 60 fps (because of oled writing).

Seen from an other side: why makes 192k audio sampling a difference while we hear only up to 15kHz? Because so we get a lot more “overtones” which cause interferences down to the frequencies we hear. This helps to increase the immersion. The faster the sampling the better, really, no matter if sound or light.

I don’t buy that audio example either. There were quite a few tests with real audiophile people that couldn’t hear the difference between 128 kBit and 192 kBit mp3’s. :wink:

But well, the Wikipedia page says, that there are conditions, where the human eye can detect shorter images. It might be worth spending a lot of time optimizing from 50 fps to 100 fps, but it’s probably not worth spending a lot of time optimizing from 100 fps to 200.

But hey, it’s your project! You make quite cool stuff and I appreciate that. Keep on going! :slight_smile:

“tests with real audiophile people that couldn’t hear the difference between 128 kBit and 192 kBit mp3’s.”
Well, what can I say? It’s weird to read the words audiophile and mp3 in one sentence. But you know what: in case you’re located in Europe feel invitetd to come arround and we do that tests. With music and with high fps animations. 128/192 kBit mp3, 96/192 kHz FLAC, 100 fps vs. 400 fps. You WILL hear, see and feel a difference, I promise. :slight_smile:

I’m from Cologne, we could do that on a free weekend. And I’d go for it, but for other reasons. :slight_smile:

I don’t see me as audiophile. Far from it! Suffering of tinnitus, shortsightedness and longsightedness (they both do not repeal each other!) I’m a bad judge in this case. :wink:

Just send me an email.

@Stefan_Petrick I’m interested to hear what delays you have between reading the channel values from your MSQEQ7… Any idea what the minimum delay is?

In the video I gave the output 30µs to settle and used “classic” analogread. The datasheet recommends 36µs. 25 or even less is also possible depending on your hardware layout but the then the data contains more noise which requires more agressive datasmoothing which takes time again (and causes latency). At the moment I get around 3000 full 7 band stereo measurements/s when doing nothing else. But I see potential to speed up the analogread itself by manipulating the ADC and reducing the resolution there to 8 bit. Basically I ordered a logic analyzer today in order to have the right tool to push the speed to the physical limit while keeping the results precise. I will report the outcome. @Fin_Belcher

Sure - I showed the code here within the comments: https://plus.google.com/115124694226931502095/posts/iMQiXg3tTH3 @Andreas_Pleschutznig