| |
The Shadow
Registered: Oct 2007
Posts: 304 |
Mp3 to sid
Is there a way to convert mp3s to sids? |
|
... 39 posts hidden. Click here to view all posts....
|
| |
algorithm
Registered: May 2002
Posts: 707 |
Some update. with two channels.. No ring modulation or additional channel yet (although this example does use two channels) Hoping to implement filter usage as well.
https://www.dropbox.com/s/uxfng2g2fw4pixx/output2test.wav
For the previous one channel version, I have been using $d418 to control the volume. With two or more channels however, changing the sustain is one of the options.
(I may update the encoder to use one volume per frame along three channels, but the selection of the waveforms would not be as rich)
The actual c64 output has some audible distortion when using more than one channel. due to the requirement of changing sustain per frame to any value (for each channel)
New to the SID chip, Would be great if sustain values can be changed higher than previous without having to resort to gateoff/testbit. Still testing, and may try other alternates |
| |
algorithm
Registered: May 2002
Posts: 707 |
Groepaz. It is synthesising speech. My encoder (updated one) analyses and synthesises chunks via replacement with waveforms and has a 'basic' engine of analysing xpletives. etc |
| |
chatGPZ
Registered: Dec 2001
Posts: 11510 |
oh, ok. different than the earlier ones then :) |
| |
algorithm
Registered: May 2002
Posts: 707 |
But yes, it does not artificially recreate phonemes, it analyses vocal hallmarks from the source, and analyses and finds individual waveform, pitch and volume. |
| |
Mixer
Registered: Apr 2008
Posts: 460 |
Perhaps you can do the slow moving adsr by using only Attack ,Release and gatebit. If you're updating the sid at 9 rasterline intervals, you could make sure that the A never reaches FF and move to Decay/Sustain phase and cause problems.
Once you have the slow amplitude curve of one voice formant, you need to convert that it into intervals and match those with SID A and R value timetables. |
| |
algorithm
Registered: May 2002
Posts: 707 |
I will see what can be done about this issue. There are some issues with unwanted tones which I am tweaking the encoder to remove or minimize. Once these have been done, I will work on getting rid of the clicking issue. |
| |
TheRyk
Registered: Mar 2009
Posts: 2466 |
hmh the sound result of that youtube by algo reminds me of a release here on CSDb... some kind of demo, probably 1 filer... must be about 2 or 3 years ago. sample of some post-grunge rock music tune Y2K+? ... melody was alright, at some stage instrument sounded like cassete player thrown into the ocean... quite cheerful even though... somebody even linked origano... narf... damn memory... Maybe anyone understands and even identifies my description & can give me a link :D |
| |
algorithm
Registered: May 2002
Posts: 707 |
It was probably "chase that feeling" by Alih, although that was a multipart demo. He uses ring modulation/sync (two channels) |
| |
TheRyk
Registered: Mar 2009
Posts: 2466 |
Exactly! \o/ made my night, not finding it drove me crazy, even leafed through my votes & comments xD THANKs mate! |
| |
algorithm
Registered: May 2002
Posts: 707 |
Another update here..
3 channels used this time. However a quick workaround where I average the volume for each chunk (based on the individual chunk volumes for 3 channels) This produces more artifacts.
Use 8580 only (as I am using $d418 per frame for volume) otherwise click galore.
https://www.dropbox.com/s/t84chpie56afs97/test3chanuniversalvol..
The output of the encode as below. (once the clicking issue is sorted out, it should sound better)
https://www.dropbox.com/s/g0lhmrsfad6rvbt/testoutput3.wav |
| Previous - 1 | 2 | 3 | 4 | 5 - Next |
