While I found this video to be entertaining and awesome, I don't like how they're trying to compare the compressor & 'tube' etc settings to 30k outboard gear and imply you're just as well with this...
This is a really neat demo, but like all demos, has a lot of tricks behind it. Analyzing demos is sort of like reading a mystery -- much of the fun is figuring out how the "twist" ending will be pulled off.
In this case, his first trick is from the demo C64MP3, which just plays the Tom's Diner clip (2 minutes of 8 bit samples in 64 KB).
The trick there is that the female voice works in a narrow range, so he is able to make a very small sample table. The higher frequencies are regenerated with the SID chip using the noise function. That's why it sounds so breathy or scratchy. The amount of noise for each sample was measured on a PC.
The real-time effects are done through one of two methods:
1. Digital domain -- timestretching or compression just involves repeating a sample or discarding it, respectively
2. Analog domain -- the SID has a lot of neat features, including analog filters. So he can post-process the samples by just reconfiguring the hardware. The sub-bass is regenerated by playing a triangle wave with one voice at the proper frequency (again, driven by the tables generated by the PC).
Very neat stuff. Knowing the tricks behind it makes it even more amazing. BTW, the SID has an analog-in port so you could hook up an external audio source and do a lot of the filtering or mixing with SID voices there. I'm not aware of anyone doing that, but I'm not an expert on this scene either.
5 comments
[ 3.0 ms ] story [ 25.7 ms ] threadStill cool though!
In this case, his first trick is from the demo C64MP3, which just plays the Tom's Diner clip (2 minutes of 8 bit samples in 64 KB).
http://noname.c64.org/csdb/release/?id=87985
The trick there is that the female voice works in a narrow range, so he is able to make a very small sample table. The higher frequencies are regenerated with the SID chip using the noise function. That's why it sounds so breathy or scratchy. The amount of noise for each sample was measured on a PC.
The real-time effects are done through one of two methods:
1. Digital domain -- timestretching or compression just involves repeating a sample or discarding it, respectively
2. Analog domain -- the SID has a lot of neat features, including analog filters. So he can post-process the samples by just reconfiguring the hardware. The sub-bass is regenerated by playing a triangle wave with one voice at the proper frequency (again, driven by the tables generated by the PC).
Very neat stuff. Knowing the tricks behind it makes it even more amazing. BTW, the SID has an analog-in port so you could hook up an external audio source and do a lot of the filtering or mixing with SID voices there. I'm not aware of anyone doing that, but I'm not an expert on this scene either.