Log inRegister an accountBrowse CSDbHelp & documentationFacts & StatisticsThe forumsAvailable RSS-feeds on CSDbSupport CSDb Commodore 64 Scene Database
  You are not logged in 
FRODIGI   [2014]

FRODIGI Released by :
Onslaught [web]

Release Date :
27 May 2014

Type :
C64 One-File Demo

AKA :
Free Running Oscillator Digi

User rating:*********_  8.7/10 (18 votes)   See votestatistics
*********_  9.2/10 (6 votes) - Public votes only.

Credits :
Code .... Algorithm of Algotech, Onslaught, svenonacid
Graphics .... Algorithm of Algotech, Onslaught, svenonacid
Design .... Algorithm of Algotech, Onslaught, svenonacid
Idea .... Algorithm of Algotech, Onslaught, svenonacid
Concept .... Algorithm of Algotech, Onslaught, svenonacid
Sampling .... Algorithm of Algotech, Onslaught, svenonacid


SIDs used in this release :
FRODIGI(/MUSICIANS/K/Khugiani_Naveed/FRODIGI.sid)

Download :
http://csdb.dk/getinternalfile.php/130122/frodigi.prg (downloads: 815)
http://csdb.dk/getinternalfile.php/130123/frodigi.d64 (downloads: 357)

Look for downloads on external sites:
 Pokefinder.org


User Comment
Submitted by algorithm on 1 June 2014
Proof of concept demonstrating compression, not necessarily amazing quality. It seems that 50% of people cannot recognise the audio at all, while the other 50% can fully.. Compare the speech sections, in particular the taylor dayne and n-trance parts to the original acapella. Formants intact, xpletives clear...

Higher bitrate to arrive (and yes its sounding far better) for speech that is, other types of audio still have issues

No "sideborder" removal. just two d020 splits :-)
User Comment
Submitted by mankeli on 1 June 2014
Did you compress the sample manually with hex editor, or what's your excuse for not including a higher bitrate version?

"Yeah yeah I always have the sideborders open in my parts, but there just wasn't enough time to put anything in there!! honest!!"
User Comment
Submitted by algorithm on 29 May 2014
In comparison to the compression rate (150 bytes per second) the "warbling" speech sounds very recognisable (I hope)
I estimate using this method that i can get crystal clear speech in probably 2 minutes using this approach.

The reason why i included 4 minutes was just a showcase of the compression (in favour of quantity rather than quality) Furthermore, I did not have the higher quality encoder ready in time, and some flashback demo parts to put together, hence released this quickly.
User Comment
Submitted by Oswald on 29 May 2014
why not did you choose crystal clear speech in 30 seconds, instead of 3 minutes of warbling? btw interesting how some drums gets translated into something like musicians would do for drum instruments :D
User Comment
Submitted by celticdesign on 28 May 2014
Impressive.
User Comment
Submitted by algorithm on 28 May 2014
This is only the start. Even with slightly more bit usage per frame, the speech can be made crystal clear. For music it would need some revamp, although the additional data would improve the quality somewhat
User Comment
Submitted by Oswald on 28 May 2014
cool, didnt think this method could achieve this much. I also wonder what would be possible with more information used per tick.
User Comment
Submitted by algorithm on 28 May 2014
42 is only the start :-)
User Comment
Submitted by Bob on 28 May 2014
Frantic: that would be "42"
User Comment
Submitted by algorithm on 28 May 2014
@sith.

By performing analysis of commonly uses frequencies, i can define a custom notetable which will ensure that there are more frequency data allocated for the important sections. This would reduce the warbling.

In addition, increasing the vq note table to 12bits per sample (4096 presets) would significantly near or enough removal all warbling.

For speech this would be enough to sound great considering that the speech as it is with this current implementation sounds good enough - but not great :-)

Would need to rehaul the encoder to support higher quality (in particular the bass frequencies) for other types of audio material however
User Comment
Submitted by Frantic on 28 May 2014
One day Algorithm will optimize, compress and encode the whole internet, in a format that fits on a single 5.25" floppy disk (although not in a lossless format). Then he will retire to a fancy beach somewhere on the globe and consume porn until the day when it is all over.
User Comment
Submitted by Sith on 28 May 2014
There's a lot of bleeping in there but you can understand the speech. Taylor Dayne, ah the nostalgia! The last song part was used by Algo before in 'Demolicious' but this time it is clear and understandable as the other method was too compressing. I'm still waiting for that killer demo from him...
User Comment
Submitted by Fresh on 28 May 2014
Hot stuff Algo. I agree with ccr, I'd like to hear what you can achieve with higher bitrates and/or updates per frame.
However speech, considering the current bitrate, is amazingly intelligible.
User Comment
Submitted by algorithm on 27 May 2014
DKT. There is around 4 minutes of audio in the demo. If you wait a while you will hear three speech samples (taylor dayne, Toms Diner, and N-Trance)
User Comment
Submitted by DKT on 27 May 2014
Kinda psychedelic. Nevertheless great to hear it on C64.
Could you show some speech only samples? I'm curious.
User Comment
Submitted by algorithm on 27 May 2014
The warbling effect can be removed once I update the encoder to use more 'note' tables as there are 512 shared mainly across 3 waveforms in this implementation (with some used for whitenoise) for the xpletive sounds (to make the speech more intelligible)

The next implementation will use 4096 presets (this should nicely fit in 5 bytes (12bitsx3)+4bits volume

I believe the speech by now should be recognisable (even if the original is not heard) but that ofcourse purely depends on the original frequency/none of the original audio and how well the predefined tables and mixes match up to the original.
User Comment
Submitted by TheRyk on 27 May 2014
somehow I was hoping for sth from TheProdigy, but boing boom tchak and Underwater Taylor Dane (Mermaid Dane, hehehe) etc. were also a great way of starting the day :D

Amazing how much music you squeezed into a one filer. And even though the performance still kinda... erh... hurts, I'm also surprised how well I could understand sung language (which of course might be an illusion since I knew almost all the tunes before and my brain will reconstruct any missing information).
User Comment
Submitted by ccr on 27 May 2014
Interesting proof of concept. I'd be interested in what this method can do with higher bitrates and/or more updates per frame.

What metric(s) are you using for the generic algorithm, PSNR?
User Comment
Submitted by Rudi on 27 May 2014
nice!
Search CSDb
Advanced
Navigate
Prev - Random - Next
Detailed Info
Summaries (1)
User Comments (19)
Production Notes
Fun Stuff
Goofs
Hidden Parts
Trivia
Forum
Discuss this release
Sponsored links
Support CSDb
Help keep CSDb running:



Funding status:




About this site:
CSDb (Commodore 64 Scene Database) is a website which goal is to gather as much information and material about the scene around the commodore 64 computer - the worlds most popular home computer throughout time. Here you can find almost anything which was ever made for the commodore 64, and more is being added every day. As this website is scene related, you can mostly find demos, music and graphics made by the people who made the scene (the sceners), but you can also find a lot of the old classic games here. Try out the search box in the top right corner, or check out the CSDb main page for the latest additions.
Home - Disclaimer
Copyright © No Name 2001-2016
Page generated in: 0.392 sec.