Search options (Page 25 of 137)

I am going to look into this. The problem should be fixable.

I recently touched up the prosound my old GBC that I've used for experiments. I'm tapping the signals off the same point (basically right off the CPU pins). But I've done away with the caps and just placed resistors in series, about 100 ohm or so. This DC couples the signal completely for the flattest possible frequency response. This relies on the receiving equipment having DC blocking caps, which is almost always the case.

Whether or not you put the cap in series, I recommend putting a resistor in series as well as a protection against a short circuit to ground or an output to output connection. With just the cap in series, there is a dead short to the other side for frequencies above DC (0 Hz) which may damage the CPU if an incorrect connection is made.

You can use a polarized capacitor for the signal coupling if you put the + side toward the CPU. The CPU will always produce frequencies between 0 and +5V, so the polarity requirement should not be violated.

I'll have to try the extra cap on the voltage rail to see if it gets rid of the horrible buzz.

Bro, do you even hex?

http://www.wolframalpha.com/input/?i=pi+hexadecimal

3.243f6a88 85a308d3 13198a2e 03707344 a4093822299f31d0082efa98ec4e6..._16

First off, you should probably use the newer version of Pandocs:
http://problemkaputt.de/pandocs.htm

Secondly, you can experiment with these things in LSDj.
1) This setting corresponds to the length setting in an LSDj instrument. When then bit is 0, the channel continues to play forever (or until the envelope runs out or you kill it.) When that bit is 1, the channel stops playing after a period set by NRx0.

2) Yes, all 32 of the 4-bit samples are played sequentially in a loop. The volume, btw, is set by shifting each sample 1 or 2 bits to the right, which means the sample isn't just attenuated, but also "bitcrushed" to use modern terminology.

3) This setting corresponds to the envelope setting in a pulse/noise instrument in an LSDj instrument, or an E command. You can rewrite a new value to restart the envelope, but this will produce an audible artifact in the audio, and it's up to you to time this right by whatever methods your playback routine is using.

4) This is the shape setting in an LSDj noise instrument. The hardware has a Fibonacci linear feedback shift register which is producing pseudorandom data. The bit setting determines where the taps are located. The other settings determine how fast the generator is clocked, and in effect which frequency, or period it has. 7-bit means that period is so short that it sound tonal. 15-bit means the period is long enough to sound like white noise, but nevertheless, the output is periodical This can be heard in LSDj with a standard noise instrument. When played at "C-3" you can hear repetitions a couple of times per second.  At "C-4", it repeats so fast that the repetition starts to sound rhythmical.
The data is output as a single bit tapped from the shift register, so it actually just flips between 0 and 1 really fast. This can be clearly seen in a recorded waveform on some shape settings.

Hope that helped. Also, you should totally hang out in the IRC channel #gbdev on EFNet.

Ever heard of collabs? I once participated in a single song made by 4 people.

Edit: But I see the seller's mistake. The file list starts counting from 0 and ends on 7. Haha!

アクアス = Aquas, so that's just the song title. So basically it detected that the notes and tempo are similar enough to some other version of the song.

A track off of this CD:

http://www.cdbaby.com/cd/ventipetrov

Apparently the same piece played on piano, in the same key and about the same tempo. Here are the two tracks mixed together for comparison:

http://www.gg8.se/temp/cygnus-comparison.mp3

Dang robots...