Replies: 4 comments 9 replies
-
|
Hey Johan, many thanks for those insights! Glad you like it. @klerg brought up this topic one year ago, and I wanted to give it a try, as those "original" Soundfont-songs were quite catchy. Altitude is among those that made it into our testsuite - so thanks a lot! :) Currently still doing a bit bug fixing (#1516) so the next version of fluidsynth should behave a bit more accurate. (Creative's documentation about the AWE32 implementation is a mess). Regarding SBK to SF2 conversion: @mrbumpy409 is the expert for this topic and he once wrote up a few steps how to do that. I copied them to our wiki - AFAIK there is no fully automate way with pleasing accuracy. But IIRC, Christian wanted to convert all the SBK on this FTP server to SF2. Maybe he's got some news to share? |
Beta Was this translation helpful? Give feedback.
-
|
Hello That is a nice story to hear from you I wonder if your friend made any midis with the Sound Blaster AWE32 from back in that day. Well, the AWE32 was not that cheap at that time. Yes, my first sound AWE was the AWE64 and it was the same for me. I did not know the "AWE Hotspot" was first called "Sound Blaster AWE32 homepage " just one more good thing to find out. That basically is the same thing I was doing but I must came up empty handed too many times so at some point I said let's do something on it. The dev here told me he needed info on how the filters work on the AWE32 so I had to look this up on google, I was in for a surprise too. Yes, after much digging on the internets. I came about a document Creative Labs calls "AWE32 Developer's Information Pack" and also the "AWE32 FAQ" then the Fluidsynth dev took over from that as it is way over my head. Yes, Altitude was one of the main midis we used to support AWE32 NRPN filters. I knew at least some of the people who made the AWE32 midis were still around today, yes you are right about that. Sure that is fine and back in the 90s was not so easy to make MP3's like today. Yes, the original format is a bit dated as it used the SBK file that predates SF2 file, and I was not able to load it into any banks. Yes, the ROMs need to be replaced. I'm looking into that as well. Yes, the AWE32 FTP archive of midis has been saved probably forever now and that is very good too Just as the Fluidsynth dev says the Creative Labs docs mentioned above have errors and mistakes so it has taken longer than expected to add AWE32 NRPN filter support, but it is going quite well so far. I'm sure one day soon that expert here will be able to convert all the SBK to SF2 so it be possible to load the sound banks that come with all of the AWE32 midis like Altitude, frenetic, Marathon, MYSTRAVE, trancpar, TRUTH, and Uplift. |
Beta Was this translation helpful? Give feedback.
-
|
Hello again! I really like the dedication to this old device. When it came out it was the first "real" synthesizer that was somewhat affordable here in Sweden even for a student. I had to eat macaroni and ketchup for a few weeks ;) My friend and I did a few songs, both separate and together. Altitude, Frenetic, Marathon, Mystic Rave, Trance Parent, Truth are all songs from us :) And Echo and Argue that can also be found on some of the ftp mirrors. The original ftp archive was at Lysator and is still there but has lost some of the files, Lysator. Looks like the mirror Freebsd.org has more files. I will definitely follow this project. Will see if I can find a way to to convert SBK to SF2. I saw this tool, Awave. I haven't tried it because I mainly use Mac at home. I had no problem running Fluidsynth on my Mac :) |
Beta Was this translation helpful? Give feedback.
-
|
Hello Yes it was supported up to the Audigy 2 ZS. Same for me in the USA this was a soundcard that had an actual synthesizer and was possible to buy if you saved up for it. Wow that must have been a chore to do. Very good to know you both made songs, either way it is all good. Yes, these six songs make up the bulk of the AWE32 midis that use NRPN filter messages. Argue and Echo are great, but do not have any NRPN control of the filter. I looked at the two links to the AWE32 midi file collection, and could not find any more midis in the Freebsd.org than at Lysator at all. Sure, that sounds good to hear then. Yes, a few technical points make it harder to go from SBK to SF2, but Awave is a great choice. I do not know how to do this on a Mac, and it may not be possible or likely. |
Beta Was this translation helpful? Give feedback.
-
I'm not so sure about that. My experience with Awave's SoundFont conversion has always been poor and plagued with attributes failing to convert. Perhaps it is better now? As far as I know, it is best to use E-MU SoundFont Librarian for SBK to SF2 conversion, but this won't run on modern Windows. I had to use Windows 98 in a virtual machine to make it work correctly. Johan, if you have any other old SBK files that you would like converted, share them with me and I can convert them for you. I'll just need to set up another Windows 98 virtual machine since I am on a new computer now. |
Beta Was this translation helpful? Give feedback.
-
|
Well, maybe Awave is not a great choice, but it has to be worth a shot to try. You must have used a very old version of Awave as I see in the changelog for 11.6 here it says "Improved conversion of SoundFont v1 (.SBK) and added support for reading ROM waves from the SB Live-driver file "1mgm.sf2" (when reading .SBK or .SF2 files referencing ROM waveforms)" What I used in the past was Vienna SoundFont Studio, not sure how well that works compared to E-MU SoundFont Librarian. |
Beta Was this translation helpful? Give feedback.
-
|
Hello! I am mostly interested in listening to my old songs for nostalgic reasons :) Nice to relive the good old days. I was impressed with what I heard when listening to Altitude and Marathon via Fluidsynth. Even if it is not 100% I think so far it was good enough. I did try SpessaSynth but it does not support the filters. The ftp archive that we used to have is not maintained by either me or Jesper. It just exists as "mirrors". It would be nice to have an area somewhere with AWE32 songs. Maybe also make some nice AWE GUI frontend for Fluidsynth ;) |
Beta Was this translation helpful? Give feedback.
-
|
Hello Yes, I know what you mean and it is the same way for me as well for the most part. Sure, but it is not as old for me. Nice, I was going to ask you how Altitude, Marathon, and Truth sound in Fluidsynth. Well, it may never be 100%. Yes, same here I'll take close to 90%. I hope SpessaSynth can work with AWE32 NRPN soon too. Sure, I'm not sure who maintains the ftp archive but that is something anyone can look up. Yes, and that is good thing. If you search for it more AWE32 midis can be found on some sites online, I'm not sure how is an AWE GUI frontend going to be for Fluidsynth ? |
Beta Was this translation helpful? Give feedback.
-
|
Hi Johan, What a delight it was to read your story! This whole dive into old AWE32 tunes has been quite nostalgic for me as well. I was a teenager when the AWE32 came out, and I knew I needed to have one. It was the first sound card I purchased with my own money, and I immediately became hooked on composing and creating SoundFonts. I grew up in a small town in Wisconsin, USA. My family didn't have internet at the time (mid-nineties), so I would bring boxes of floppy disks to my piano lessons. My teacher was a doctoral student at the University of Wisconsin, and he would let me into the computer lab so I could fill up my floppies with downloaded SoundFonts and tracker music. I also learned how to do my own sampling and eventually created GeneralUser GS, which is still only 30 MB in size as it originally had to fit inside the maximum expanded AWE32 RAM (28 MB). During this time, I also created a few AWE32 MIDIs, though none were really circulated except for March No.1 in D Major, which won sixth placing in the 1997 Creative Inspire International MIDI Contest hosted by Creative Labs. The first place winner was the fantastic The Bloodwar AWE32 MIDI by Jess D Skov-Nielsen.
That page definitely rings a bell.
The ROM samples are only part of the problem. Correct playback on newer devices based on the SoundFont 2.01/2.04 spec also poses a significant challenge. Considering no spec fully describing the behavior of SoundFont 1.0 or 2.0 devices exists, I have had to base my conversions on my memory plus the results from tests that others have graciously recorded for me on their legacy hardware. Even with all of this, there are some AWE32 & Live! behaviors that cannot be perfectly recreated on Audigy or any other SoundFont 2.01/2.04-compliant samplers, not to mention how the different synths handle things like out-of-range parameters or loop points, etc. When converting Tekkniko by Niko Boese, I might not have noticed that one instrument was completely failing to sound at all were there not an AWE32 rendering online to compare it to. derselbst wrote:
Yes I do have news! I have converted all of the MIDI+SoundFont files in that archive to SoundFont 2.01/2.04 compliance, including migrating all ROM samples to RAM. I also converted several AWE32 MIDI files that I had that were not a part of that archive. Johan, the three of your songs that I found were converted: Altitude, Marathon, and Truth. Additionally, those under your friend's name (Jesper Nordenberg) are also included: Argue, Echo, Frenetic, Mystic Rave, and Trance Parent, with notes of your collaboration with Jesper noted in the text files accompanying Argue and Echo. So now, what to do with this archive I have created? I would like to figure out the best place to upload the entire converted collection, and would like to be able to update the files over time as I find and fix errors. I would also like to create recordings of these tracks on my Audigy2 and FluidSynth and ideally procure recordings of the original files from someone who can digitally capture audio from an AWE32 so I can best match the converted songs to their originals. And since we're on this topic, I am also sitting on a large archive of tracker music that I painstakingly recorded over many years, trying to preserve the correct sound for each track while also enhancing the audio fidelity. Both the tracker music and AWE32 music recordings will require a fair amount of storage space to host. Does anybody here have a good idea on what would be the best way to go about this? Peace, |
Beta Was this translation helpful? Give feedback.
-
|
That would be awesome! I'd say just put it in a git repo here on GH. (Not git-lfs, as we've learned that the monthly transfer quota for LFS will be quickly exhausted). This way it'd be version controlled and people could contribute e.g. renderings from real hardware. I guess only storing the MIDIs, converted SF2s, and the renderings of the MIDIs using the original SBK on real hardware would be enough. The MIDI and original renderings are not subject to change. Only the SF2 is, which I assume should have a low storage footprint over time. And for the fluidsynth renderings, a script could be included that just renders all of it. You could always put an (autogenerted) Github HTML page on top, if you feel the need for a "nicer" web presentation - think of fluidsynth.org - it's a Github page ;) |
Beta Was this translation helpful? Give feedback.
-
|
This is a great idea! I would really like to get hold of the converted songs :) Btw, Jesper and I are still friends :) Going to have lunch together today. |
Beta Was this translation helpful? Give feedback.
-
I am readying the collection for publication to GitHub (to be published here. It will be soon—I just need to finish writing up the documentation.
That is so awesome! I have decided to place both Jesper's and your compositions into the same folder ( |
Beta Was this translation helpful? Give feedback.
-
|
Great 😀 That will be awesome! You can name the folder whatever you want but go for that if you don’t think the folder name will be too long. |
Beta Was this translation helpful? Give feedback.
-
Hello everyone!
I would like to tell a little story.
About 30 years ago a friend and I bought the Sound Blaster AWE32 as soon as it was released and while we were studying at the university. This started our interest in synthesizers. We also decided to start a webpage about it called "The Sound Blaser AWE32 Homepage" and later renamed to "The AWE Hotspot" due to the AWE64 etc.
Every now and then I google to try to see if there is a way to play old AWE32 songs from "back in the day". I did that again about 2 weeks ago and found a discussion here about adding support for Sound Blaster AWE32 to FluidSynth. And I was up for a suprise :)
I was very pleased to find out that there is a way to do it without the real hardware. I read the discussions and saw that you used some old AWE songs as examples like Altitude that is one of my old songs. Here I am 29 years later and I can read about something I did ages ago :) This proves that if you have publiced something on the Internet it will never go away ;)
I did not create mp3s of my old songs in the 90s so I only have them in original format but do not own an AWE so I have not been able to listen to the songs for thea last 25+ years. Listening to Marathon (another old AWE song by me already in SF2) and the converted Altitude to SF2 gave me one of the most nostalgic trips ever and put a very big smile on my face!
Thank you all developers for adding support for Sound Blaster AWE32! This is very much appreciated! Now it would be nice to find a way to convert SBK files to SF2 format. I guess the ROM samples (if used) are the main problem. Maybe there is a way? I haven't googled that yet.
It is also nice that our old ftp archive of AWE32 songs that users around the world contributed to is still online.
Regards,
Johan Nilsson
Beta Was this translation helpful? Give feedback.
All reactions