[SEEKING] Intended Doom SC-55 Music

[twipley]

twipley said:

Here is Doom music as it was intended to be heard!

https://github.com/chocolate-doom/chocolate-doom/issues/245

kopasite.net/up/1/doom-rc-55-soundtrack.torrent
-- you are welcome sharing this with your friends.

Comparison between two different encodings:
-- Roth (LogicDeLuxe) from http://www.abmischung.de/index.html
-- Blume (MusicallyInspired) from http://sc55.duke4.net/index.php

(I do not let my torrent client opened. If you need a seed, leave me a note and I will do so in the following days.)

It might be interesting to compare both of these.

kkaden (over at github) has noted some differences between the two. The quest is to find or attain the original, intended, authentic Doom audio experience.

[kkaden]

They're both really good recordings and I couldn't choose which is "better". I'd just be happy to listen to either in my port of choice, Chocolate-Doom.

As far as finding the originally intended experience goes, I think we'd need Robert Prince to record the music for us on his original equipment.

Having Logic DeLuxe's recordings has been great, but he only ever released Doom and Doom2's music. Musically Inspired plans to release recordings of Doom2 and TNT's music, as well as continuing on to Heretic and Hexen, and eventually the music from other Doom derived games.

Also MI's recordings feature Loop Point metadata for use in ZDoom, so that the music continues as it does normally playing with midis.

[twipley]

kkaden said:

As far as finding the originally intended experience goes, I think we'd need Robert Prince to record the music for us on his original equipment.

This, or inquire about the method each of Roth and Blume used.

Because, isn't after all the only needed equipment the Roland sound card? It is what he intended people using.

I really wonder what is the cause of differences between both above encodings.

EDIT: I've informed both Roth and Blume about the existence of this thread.

[Dragonsbrethren]

The SC-55 is what he composed on. He expected people to use all different sound hardware. I doubt he intended anything.

I mean, Doom even has the GENMIDI and DMXGUS lumps specifically for making music sound good on other cards.

[twipley]

I guess you're right. I meant he would have liked people to use SC-55s. Because, it permits hearing Doom the way it was designed to sound like.

[LogicDeLuxe]

All details about my recordings should be somewhere in the old thread here: http://www.doomworld.com/vb/doom-general/41149-original-full-quality-doom-12-soundtrack/
I don't remember everything. It's a long time ago.

[twipley]

Here is -- I believe -- most of the relevant material from that thread:

LogicDeLuxe said:

Now, I have to dig up this thread, as I have something for your ears.

I got myself a Roland Sound Canvas SC-155. It is basically a SC-55 in a different housing and more knobs and sliders. Same wavetable, same GS engine. And definitely better noise ratio than a SCC1.
I recorded the MIDI files (that link from this thread) through a M-Audio Delta-44, which has 24 bit ADCs providing a decent quality. They are certainly better than the DACs on those old Roland devices, thus more than adequate.
It is mastered in Adobe Audition for CD quality. Nothing beefed up, of course, as the whole point is giving the pure sound of the original tracks to people who don't own one of those Roland devices.

exp(x) said:

I think it would be implemented best by supporting flacs named <lumpname>.flac (ex. D_INTRO.flac) in ~/.chocolate-doom/music/doom.wad/ similar to how savegames are stored.

LogicDeLuxe answered:

You also need loop data, though. The FLAC has to repeat itself while the fading notes and reverb at its ending is still playing. The exact point where the original MIDI ended has to be stored somewhere.

Never_Again said:

Speaking of mastering, the intro (track 01) clips quite noticeably. Unlike the added silence, it cannot be fixed on the downloader's end. I also wonder if you used peak compression. What exactly the mastering part involved, anyway?

LogicDeLuxe answered:

I played the MIDI's provided in this thread earlier to the end, until the last note and reverb is over.Unlike the MP2 tracks, this don't clip, really. I mastered them a few dB quiter than those MP2. There are some multiband peak compression involved on very few occasion. And the peak limiter softens some peaks slightly. I sure compared every mastered track carefully to the original recording at the same RMS loudness to be sure that there is no significant difference. I am the last who would overcompress or clip a recording just to make it excessive hot. The RMS matches those of the late 80's and early 90's which used to be a good compromise between loudness and resolution usage. I usually master at -14 dBFS RMS sine, and this one is even slightly softer. In my experience, there are only few cases where even more dynamic is important, which are mainly classic, some jazz styles and audio dramas. Most mastering guys probably would have edited this much more aggressively, I guess. Compare this sound to those mainstream CD's released over the past few years, for instance. Btw. I am a TurnMeUp supporter.
Other than this, I cut off the mirror frequencies above 16kHz with a FFT filter and did a phase true subsonic filter below 30 Hz to save some power. (Lowest note in those tunes has a base frequency at about 45 Hz.)
No noise reduction involved, since the SC-155 has already a pretty good noise ratio. Thus no need to mess with it.

Looping is one not so trivial think to do. The problem is, that I would chop of the reverb if I'd just cut the end. You could insert it to the beginning in order to prevent this, but then, it won't sound like a real beginning anymore. Thus, the best think to handle this issue would be an engine handling this, ie. restart the track from the beginning while still playing the reverb of the last note.Do you mean the OPL3 and the MU5 example? If yes, it well could be the case. Some of my cables have its colors wrong on the RCA jacks for some reason, and I just quickly made those examples without checking the channels.
I did for the Sound Canvas recordings, though, I think.

LogicDeLuxe said:

Here is the Shareware music from Vanilla Doom recorded from my Terratec Maestro 32 with a genuine OPL3 in stereo: http://www.rapidspread.com/file.jsp?id=vqxe0waz4x
The files loop correctly with a slight compromise: In some songs, the first play has a different panning to successive loops. I edited the songs in a way so that they loop seamlessly. Because of this, you can hear decaying notes from the end of the song at the beginning in some cases. Is this okay, or would you prefer a perfect beginning in favor for loopability? D_INTER has a slight pause at the end, which is exactly the way, Ultimate Doom plays it.

[MusicallyInspired]

Hey, everyone! I apologize for not posting myself but my account hasn't been activated on Doomworld yet and I'm still waiting. First of all, I didn't realize that somebody else had already recorded a Sound Canvas soundtrack for the Doom games so I never recorded it to compete with LogicDeLuxe. I recorded my soundtrack with an external Roland CM-500 module (which contains the same content inside as the external Roland SC-55 module which the soundtrack was primarily composed for, as well included is a variant of the Roland MT-32, more specifically CM-32L) straight into a Tascam audio interface. I did use some multiband compression to boost the levels as much as possible without taking away the dynamics. I tend to agree with LogicDeLuxe about the industry over-compressing everything and losing all sense of dynamics, which existed over 12 years ago.

As for looping, I used metadata loop points that loop each track seamlessly. Rather than cut off the end of the track (which would chop off reverb and the ends of some instrument notes) and loop back to the very beginning, I instead continued the track after the end to play again and then fade out after a certain point, so you can have a decent listening experience outside of the game without tracks getting cut off. I then created an end loop point a little while after the actual end of the track (but before the fadeout) and another beginning loop point a couple seconds after the beginning of the track. This gives the illusion of a perfectly seamless track which preserves reverb and instrument end note data without chopping it off and sounding obvious and jarring. So it really sounds like you're playing the MIDI tracks with an actual Sound Canvas, which would obviously also preserve reverb data and note end information as it loops. The result are tracks that, when played by a port that supports metadata loop tags, will loop forever completely seamlessly and unnoticeably!

I started recording an SC-55 soundtrack for Duke Nukem 3D a few years back and went through several revisions until I got it right with some invaluable help from a couple other community members. The point was to use it with EDuke32 for people who never owned or had a chance to listen to the soundtrack from a real Sound Canvas. I recently released version 4 of that soundtrack which also contains metadata loop points and is included in the Megaton Edition on Steam and other platforms that it may be ported to. After doing Duke, I also did its expansions. At this point I decided to create a website of SC-55 soundtracks and record them for as many games as I could that I had and that supported the Sound Canvas. A few people even sent me demo and beta MIDI files to record, which I haven't gotten around to yet. So far I've done Duke3D + Expansions, Shadows Warrior (the demo, the full game uses CD audio), Blood (demo again), Descent 1 & 2, and now Doom. I plan to do Doom 2 and Plutonia next.

The reason I have duplicate songs in the Doom soundtrack (like d_e2m3 and d_inter) is because the MIDI files themselves were also duplicated in the WAD. I wasn't sure how the game or source ports utilized the music files and I wanted to insure that no tracks were left out in-game. I referenced heavily the music list at http://doomwiki.org/wiki/Doom_music and paid close attention to which levels referenced which music files. I didn't record these tracks twice, however, I merely copied what I had already recorded, so the duplicate files are identical including the metadata loop tags. I never included files for episode 4 levels because, according to the aforementioned list, these levels actually referenced the already existing music files rather than including more copies.

Plutonia apparently is comprised completely of copies of Doom and Doom 2 music so I'm not sure if I will record a special pack for it alone or not. I'd appreciate input here. Should I create packs for each WAD? Or will Doom and Doom 2 packs suffice? TNT includes music from Doom 2 as well, but also original music by three other composers, so I'll have to make a special pack for that. After that, Heretic and Hexen are next on my list, along with other Doom engine games that also have custom soundtracks (I don't yet know what these all are, but I plan to make my website as complete as possible over the course of time). I've got other games on my list to do that use the Build engine as well as other first person games. The point is to have an authentic 90s listening experience for use with each game's source port. Not all of the soundtracks I've made or plan to make have ports capable of playing back music packs, but I'll still record them anyway in the event that they may get a port that does. Or just for the listening pleasure of those who love the 90s sound.

A bit more information about my recording process, as I have said I am recording from a CM-500, which is basically an SC-55 without any knobs or controls whatsoever beyond a volume knob. The purpose of the Roland CM series line was for gamers who didn't need the extra functionality to compose with, so I imagine they were cheaper (CM stands for Computer Music). The CM-500 is a two-in-one combination module which includes a CM-300 (SC-55) and a CM-32L (MT-32, third generation I believe). A switch in the back toggles between 4 different modes: CM-300 mode, CM-32L mode, and 2 more modes which feature combinations of both CM-300 and CM-32L modes). Obviously the only mode I use for my recordings is the CM-300 mode. I record directly from the CM-500 module into my Tascam US-1800 audio interface. I collect the MIDI files from each game and use the Cakewalk Sonar X1 Producer DAW to playback and record them. I record all tracks as 44.1 Khz 24-bit masters and then bounce down to 16-bit for the final tracks.

Opening the MIDI files in a DAW and recording that way as opposed to simply playing from each game is more beneficial because it's easier to control the environment; how the song loops, where the loop points should be, etc. It's much easier working in an environment that has the exact tempo information relevant to the music so that you can work with exact measures and beats, it makes things SO much easier. With Doom's soundtrack, however, I found that the tempo of the MIDI files never corresponded to the actual music notation so I had to do more work there to get things to loop properly and precisely. As I said, I did to some multiband compression but the reason is because I find that when playing these tracks in a source port the sound volume is far louder than the music volume. For EDuke32 specifically, there was no dedicated music volume slider at first, only the audio slider which also controlled the music volume. So it was impossible to get the music louder than the sound effects. This has since been resolved, and the Doom ports probably don't have this issue, but I find just a LITTLE bit more punch is worth it. I was careful to not go overboard and damage transients and dynamics too dramatically, though. That said, I still have the unmolested masters so I could easily go back and remove the compression and bounce new tracks. I did so with the Duke3D soundtrack as well and planned to release both, but I was being conscious of server space and nobody really seemed to request the uncompressed tracks after I had finished, so I never bothered. At the very least, LogicDeLuxe's tracks are uncompressed so people can choose which versions they like I suppose. :)

That's about all I can think of to say. I apologize for the length. :)

[Gez]

MusicallyInspired said:

I referenced heavily the music list at http://doomwiki.org/wiki/Doom_music and paid close attention to which levels referenced which music files. I didn't record these tracks twice, however, I merely copied what I had already recorded, so the duplicate files are identical including the metadata loop tags. I never included files for episode 4 levels because, according to the aforementioned list, these levels actually referenced the already existing music files rather than including more copies.

Note that Doom II includes a duplicated song that actually exists in two different versions.

MusicallyInspired said:

Plutonia apparently is comprised completely of copies of Doom and Doom 2 music so I'm not sure if I will record a special pack for it alone or not.

For Plutonia, you might look at this I guess. It's either that or using recordings already made for Doom and Doom II.

MusicallyInspired said:

After that, Heretic and Hexen are next on my list, along with other Doom engine games that also have custom soundtracks (I don't yet know what these all are, but I plan to make my website as complete as possible over the course of time).

The Doom Wiki has an exhaustive list. Only the PC games are relevant for your purposes, the PSX/N64 soundtrack was definitely not designed for SC-55!

That said, if you're not afraid of multiplying your workload by untold scores (haha, pun!), many Doom mods have original soundtracks of rather good quality.

MusicallyInspired said:

I've got other games on my list to do that use the Build engine as well as other first person games. The point is to have an authentic 90s listening experience for use with each game's source port. Not all of the soundtracks I've made or plan to make have ports capable of playing back music packs, but I'll still record them anyway in the event that they may get a port that does. Or just for the listening pleasure of those who love the 90s sound.

I love old game MIDIs too. Very fond of the Ultima games (especially U7 and UW2), TES: Arena and Daggerfall, Hocus Pocus (music was the best part of that game)... And Doom and Heretic of course.

[MusicallyInspired]

Actually, I had intended on making SC-55 packs of Duke3D mods/tcs as well in the beginning. I think I'll move on to mods after the main games I have in mind are done. Thanks for that exhaustive list, by the way! That will come in handy for sure.

And yes, Arena and Daggerfall had fantastic soundtracks. Hocus Pocus as well! Definitely what drew me to the game to begin with. Without that soundtrack I doubt I would have even played it. Really helps with the atmosphere. Definitely doing that soundtrack as well. I'm thinking of getting FPS games out of the way first before moving on to platformers.

[twipley]

MusicallyInspired said:

That's about all I can think of to say. I apologize for the length. :)

Hey, thanks for such an informative post!

I hadn't realized the extent of the lengths you guys travelled, performing all of those subtle modifications over the original outputs. I am not sure of my opinion on this. On the one hand, you guys sure are going over for quality, and are very meticulous in their technique in doing so. On the other, though -- I hope my judgment is not clouded by the fact that I am, in terms of audio mastering, technically illiterate -- some part of myself is left wondering the pertinence of such modifications.

While I realize these are applied for the common good, such techniques as multiband compression, peak limitation, cutting off mirror frequencies, and increasing the music volume (in comparison to the sound one) leave me with a touch of uneasiness. For example, if the original Doom experience is having more sound than music reaching your ears, then so be it! (Isn't it?) Although, as a purist, I might be biased.

However, I think part of what the community (at least, the Chocolate-Doom one) is interested in, is in "unmolested" music. "Nothing beefed up, [so as to be] giving the pure sound of the original tracks to people who don't own one of those Roland devices."

I have a hard time writing all of this without sounding like bashing both of you. Trust me, this is the farthest from my authentic intention. If I could talk to you guys face to face, you might understand that better. My delicate manners seem somewhat obscured by the coldness of all this text.

[LogicDeLuxe]

twipley said:

For example, if the original Doom experience is having more sound than music reaching your ears, then so be it!

Since you could change the volume for music and sfx independently from the beginning on (at least, you can in Doom), there is no intended relationship, but merely personal taste.

However, I think part of what the community (at least, the Chocolate-Doom one) is interested in, is in "unmolested" music. "Nothing beefed up, [so as to be] giving the pure sound of the original tracks to people who don't own one of those Roland devices."

This was my intention to begin with.

For dynamic processing: I didn't altered the overall dynamic, so the transients and punch remains as they were intended. Only few peaks were actually limited in a way that they match other similar parts of the song. There is no intended peak volume to begin with either. Due to jitter in midi transmissions and also due to the nature of effect processors, you rarely get the same peaks twice anyway.
And if you just plug the Sound Canvas in the average Soundcard's line-in connector of the time, you would probably get far more distortions than some very minor dynamics processing would do.

For the filtering: Roland synthesizers of the time rendered at 32000 Hz internally. There is already a low pass with a low order in the device, which I just enhanced with an FFT filter, which stays true to phase. This would make the instrument samples sound more natural, as mirror frequencies tend to sound very artificial. Though since this effect is very minor on those devices, you won't probably hear the difference most of the time.

[MusicallyInspired]

WARNING: another very long post!

twipley said:

While I realize these are applied for the common good, such techniques as multiband compression, peak limitation, cutting off mirror frequencies, and increasing the music volume (in comparison to the sound one) leave me with a touch of uneasiness. For example, if the original Doom experience is having more sound than music reaching your ears, then so be it! (Isn't it?) Although, as a purist, I might be biased.

Because in the old days, if you had a Roland Sound Canvas, the inputs would have had to have been plugged into either your sound card's line input or another input on your stereo system or mixer that your sound card was also plugged into. In either case, the volume of the music is inherently out of the game's control because of the simple fact that the game is not creating the music audio, the sound canvas is. The game has a music volume slider yes, but this controls the MIDI data volume being sent to the Sound Canvas, not the actual overall analog output volume of the Sound Canvas itself. You would've had to have set your own desirable volume before/while playing the game on the Sound Canvas anyway. I do this now with my CM-500 and MT-32 and old DOS games that support them when I use them and have to find a good balance between game sound volume and synth sound volume. So, if the music files recorded are too quiet it's not REALLY indicative of the original's experience since you could have turned it up back then as well. The idea is to find a balancing compromise when recording digitally because it really is an entirely different playing field.

However, I think part of what the community (at least, the Chocolate-Doom one) is interested in, is in "unmolested" music. "Nothing beefed up, [so as to be] giving the pure sound of the original tracks to people who don't own one of those Roland devices."

I feel I should explain, though it is a bit of a challenge to do so to the uneducated or inexperienced, that the purpose of mastering and all the work we've done to our respective soundtracks, was to preserve the original quality as much as possible. Not to add to it or "beef it up". There are many things I could do to add to it like remove the SC-55's onboard reverb completely and add a newer higher quality reverb digitally, add real distortion effects on the guitar instruments, delay echo, and so on. There are other concerns with recording from a synth module like this that exist simply due to the nature of being recorded and not generated from the real thing. It's impossible to exactly emulate the circumstances of playing with a real existing synth module unless you actually have one. It's the same audio, yes, but it's now being played back as a digital waveform inside the game port, not as an analog signal straight to your speakers. Again the game has no control over the Sound Canvas volume, just the volume of the MIDI data that is SENT to the Sound Canvas. And so, this is a perfect example of the age-old analog vs digital scenario.

If you want THE EXACT same experience unmolested, then you simply have to get a Roland Sound Canvas. :) However, the differences are minor when you look at the broader picture and aren't worth considering unless you're a hardcore audiophile (like me! :D). At any rate, it's not as simple as recording the tracks straight from the Sound Canvas and being done with it. For one thing, what should the analog volume on the Sound Canvas be set to when you're recording music from it? What should the input volume on your sound device that you're recording on be set to? There is no standard or environment to copy from the initial experience because that was analog and not digital, so you have to create a standard to work around. And usually the one that preserves the most quality is also the quietest. Much quieter than any other software on your computer would ever generate. This is why a lot of people still prefer analog recording over digital recording. However, it's impossible to record and playback analog sound from inside a piece of computer software, unless you actually have something like a Sound Canvas. ;)

This is going to sound fairly technical, but I don't know how to explain it any other way. You can skip below if you start to feel your head spinning!

Spoiler

Without trying to be too technical, as you can guess the difference between analog and digital is not a minor one. The sound that comes out of your speakers from the Sound Canvas, while digitally generated, is analog in that it is sent through analog signals and produces vibrations in the air from your speakers to your ears that we call sound. Sound doesn't really "clip" when you turn up your speakers until it hits the speakers' maximum capacity to produce vibrations in the air. When that happens it starts distorting because it's receiving too much signal to reproduce accurately. You can turn it up as much as you want almost alongside the game volume without any adverse effects. This is not the case for a digital recording. The volume sliders on your computer screen do not function the same way as turning the volume knob with your hand. Your computer is sending the digital recording waveform as an analog signal to your speakers, but the recording itself is a digital representation of that initial analog signal from the Sound Canvas. There's a conversion that takes place there. It sounds the same to your ears, that is, it produces similar vibrations from your speaker, but it really isn't. The digital audio spectrum inherently has a volume ceiling that analog does not have (outside of hitting your speaker's maximum capacity) and when you hit this digital ceiling the audio clips and distorts before the signal even gets to your speakers.

So, when recording digitally we have to compensate for this conversion because all of the sudden the music is being generated in the same field as the sound effects, and the sound effects are much louder than the music. There are different formats of digital sound on a computer with varying sampling and bit rates. Higher bit rates means a much more broader spectrum of volume that can be turned up or down. Higher sampling rates means the sound can be clearer and crisper and less muffled because it has a broader frequency range to encompass that sound in. The sound effects in Doom are of very low bit and sampling rates (probably 8-bit, 22Khz). Yet, modern source ports can set the game's sound settings to be as high as 24-bit and 48KHz. This doesn't really change the quality of the sound because those sounds are still 8-bit and 22khz, but they are being played back in a higher quality setting, except that they might sound a bit more grainy due to the extra crispness of the higher sampling rates.

When the music tracks are recorded they then exist in a digital realm of audio, whereas before they didn't. So now in the digital realm we have a volume ceiling which the audio CANNOT go above without clipping and distorting, no matter what your speakers' volume knob is set to. So a standard must be set as to which volume level to use when recording. Initially in digital recording, everything is recorded quietly, but not TOO quiet or you'll hit the "noise floor" which exists at the bottom of the digital volume spectrum, then you'll hear background noise that distorts the audio. Recording quietly preserves audio quality while creating "headroom space" for large spikes of volume to exist without hitting the digital ceiling. However, if you were to use these tracks without altering them in any way you would find that they are simply far too quiet in-game. You could turn down your sound effects to a much lower level and turn your speakers up, but you'd find that after you're done playing every other sound on your computer would be extremely loud.

When you're playing with a real external synth module, none of the above technical jargon matters because you can set the volume independently by hand with a real world knob without worrying about distortion. Two completely different settings of sound (the Sound Canvas and the game audio from your sound card) are being output to the same speakers (sometimes different speakers, depending on your setup). In-game in a source port, the digital volume sliders act differently. Now the digital music recordings ARE being controlled directly by the game and the music slider affects the digital audio output of each song rather than the MIDI data because there is no MIDI data! So now the volume of the digital recording has to be of an acceptable listening level AND stay within a margin of digital decibel units before distorting. So, to compromise as a way of "turning up" the music without the volume spikes hitting the digital ceiling and clipping/distorting, you compress the audio waveform data. This "squeezes" the soft and loud volume levels closer together so that the softer parts are louder but the volume spikes aren't hitting the digital ceiling. Actually, what's happening is the loud volume spikes are being compressed down closer to the same level as the softer ones and then everything is turned up, which is basically the same thing. I use a multiband compressor which only compresses certain frequency ranges (like high pitched or "treble" sounds, and low pitched or "bass" sounds and every range in between) so that the compression is less noticeable. This is the struggle I had when recording my soundtrack and why I chose to go with a bit of compression in the end. But as I said, I do have the masters unaltered and could always release that. But seeing as Logic has already done so, I see no need! :)

The idea is to make it louder. As loud as the other sounds coming from the game, but not too loud as to overpower them. As it ever is even with a real Sound Canvas, the idea is to find a balance. Modern sound designers have to do the same thing when putting sounds and music into a modern AAA title. Something they didn't have to do back in the DOS days (unless you were listening to Adlib music). It's not an easy job either. The more control you have over sound in a game the more work you have to do and compromises you have to make to balance it all out because in the digital realm, unlike the analog realm, you have an extremely specific, limited, and defined workable area that you can work in.

And that goes for any other digital medium like graphics (size and number of pixels, number of colours, number of animation frames), or even disk storage space. It's not infinite. Everything with computers is a collection of different combinations of ones and zeros and as such have inherent limitations. We can keep making higher quality content and hardware and push those limitations higher, but they will always be there. No matter how good a quality digital photo you have, it will never be the same as the actual scene. No matter how big your hard drive gets you will eventually reach the maximum capacity if you keep storing stuff. No matter what bit rate or sampling rate quality level you record an analog sound in it will still be limited to a specific scale of volume units that you cannot go above. And because the industry is the way it is, people will keep making music that pushes the limits of the digital ceiling to try and sound "the loudest" on the radio, in the theatre, and even on your computer in software form. Google "the loudness war". So we will probably never have enough room where we don't need to get close to the digital ceiling.

I have a hard time writing all of this without sounding like bashing both of you. Trust me, this is the farthest from my authentic intention. If I could talk to you guys face to face, you might understand that better. My delicate manners seem somewhat obscured by the coldness of all this text.

No offense taken, Twipley. You're concerns are quite understandable and I detected no offense in your post. :) And I'm sorry, once again, that this is so long. I tend to get very wordy.

EDIT: I see that Logic replied already (and much more succinctly than I have!) while I was typing. Anyway, I hope that I explained my decision well enough (probably overly so).

[twipley]

Both versions sound quite different. Especially, in some particular respects.

I am quite a critic, although my musical tastes (and audition!) are not acquired enough for me to adequately judge. At any rate, though, it is an outstanding advantage that both works (packs) were performed in such a complete independence from one another. That is, nobody copied no one in their technique. Two independent minds have done it their own way.

I am not to be a critic here. I could not; I do not even have the vocabulary. I (as others would) might be interested in reading more development in this regard.

I am just happy that two such independent efforts produced the respective results they have. We are lucky to have something like this to compare.

EDIT: I believe my preference (this is really a layman opinion, about which even I am not sure!) is Roth's pack for listening outside of the game, as more of the music seems to stand out; and Blume's encodings for listening inside of the game, as the music do not seem to thrust itself too much to awareness. More testing is needed before developing this opinion and grounding it into direct experience rather than speculative anticipations, though, as I myself have never, ever, played the game using such packs. I am stating this swoon opinion because I feel this thread would benefit from such inputs.

[twipley]

Gentle bump, so that people be informed of the above edit. :/

[twipley]

Hello world!

Back in early January, I had emailed Bobby in order to ask him if he would be interested in giving us some input. I have not much to do with his reply since for me it reads like Chinese, but I am guessing others might find it enlightening plenty.

twipley said:

Hey, Mr. Prince!

First off, let me state that I really enjoyed reading your Doom-music-creation-experience descriptions over at http://www.doomworld.com/classicdoom/info/music.php

Secondly, I was principally contacting you in order to inquire about particular opinions of yours. That is, I would want some input from your part.

See, I think this is an important matter -- indeed not one's of the world's biggest problems, but surely the biggest problem facing by projects such as "Chocolate Doom," aiming to reproduce the Doom experience as it was "intended" to be (especially, audio-wise).

The matter at hand concerns digital reproductions of the original Doom music. Two independent (which is a definite plus) audio engineers have converted MIDI files to digital formats, with the quest in mind of remaining faithful to what you had in mind composing the tunes. (I am hosting a torrent so one could compare both versions.) The thing is, "what you had in mind" is somewhat inaccessible to us! This is why I feel you could provide a much-pertinent input to the community (for the example, commenting those packs, or guiding future pack improvements).

http://www.doomworld.com/vb/doom-general/66795-seeking-intended-doom-sc-55-music/

Yours truly,
twipley

Bobby said:

[...] Somehow I lost this email until today when I was searching for another email. I apologize for the late reply.

If you compare the SC-55 and the FM sound card versions and the SC-55 sounds very similar to the FM versions, that's what I was attempting to do (or had in mind). Of course, if I had been able to do digital versions from the start, I would have used "real" instruments in many of the songs, but not all of them. The FM sound card, while not sounding like "real" instruments, had its strengths for creating certain types of songs. They are the ones that would be difficult to reproduce using "real" instruments. The songs with a "guitar like" leads would sound great with real guitars. The SC-55 comes closer to that goal than the FM sound card. But, the FM sound card allowed me to invent my own instruments, and some times those instruments inspired a song to be written around them.

I hope this helps you in some way. I'm not a purist when it comes to instrument choices. I think if a digital audio file sounds good to listeners, the goal of the music has been met.

I wish you all continued enjoyment working on Chocolate Doom. I'll look forward to what you come up with. [...]

Best regards,
Bobby

[fraggle]

Sounds like he misinterpreted your Chocolate Doom comment. But interesting nonetheless.

There was talk before about doing FLAC recordings of the Doom music. Are they available at all?

[kkaden]

Twipley posted a link to a torrent with two versions of Doom's music recorded and encoded in FLAC in his first post

[fraggle]

Apparently I'm not paying attention. Thanks.

Can someone seed the torrent? I can't find any peers to download from.

[plums]

Also requesting a seed.

How big are those files anyhow? Could they not be hosted on mediafire or something similar, in addition to a torrent? Even if it's in parts.

[fraggle]

The MusicallyInspired ones are available here as linked in the original post at the top of the thread, I downloaded those and am now seeding them. It's on a server of mine in the background which I'll leave running.

Would be nice to seed the LogicDeluxe ones as well.

I skimmed over the thread (which I'm ashamed to say I hadn't read in detail until now), and noticed this:

MusicallyInspired said:

The reason I have duplicate songs in the Doom soundtrack (like d_e2m3 and d_inter) is because the MIDI files themselves were also duplicated in the WAD. I wasn't sure how the game or source ports utilized the music files and I wanted to insure that no tracks were left out in-game. I referenced heavily the music list at http://doomwiki.org/wiki/Doom_music and paid close attention to which levels referenced which music files. I didn't record these tracks twice, however, I merely copied what I had already recorded, so the duplicate files are identical including the metadata loop tags. I never included files for episode 4 levels because, according to the aforementioned list, these levels actually referenced the already existing music files rather than including more copies.

Plutonia apparently is comprised completely of copies of Doom and Doom 2 music so I'm not sure if I will record a special pack for it alone or not. I'd appreciate input here. Should I create packs for each WAD? Or will Doom and Doom 2 packs suffice? TNT includes music from Doom 2 as well, but also original music by three other composers, so I'll have to make a special pack for that.

In the comments on the bug to add support for this to Chocolate Doom, I propose a scheme where the replacement songs get selected based on a hash of the MUS found in the WAD. This would have the advantage that if you have the Doom and Doom 2 music, the Plutonia music just works as well. I really think this is the best way to go with this stuff: it will just mean bundling a small config file with the music files.

After that, Heretic and Hexen are next on my list, along with other Doom engine games that also have custom soundtracks (I don't yet know what these all are, but I plan to make my website as complete as possible over the course of time).

This would be great! Note that Hexen actually had a high quality version of the music on the install CD, so arguably that's the "intended" version. But it would be good to get Heretic, and also Strife!

[kkaden]

now seeding the whole file

[plums]

fraggle said:

Note that Hexen actually had a high quality version of the music on the install CD, so arguably that's the "intended" version. But it would be good to get Heretic, and also Strife!

However, the Hexen CD only has some of the songs, presumably since the length of an audio CD was insufficient to hold all of them. I believe they're from an SC-55, but I'm not positive. Either way, they might be useful as a reference point for doing the rest of the songs.

http://doomwiki.org/wiki/Hexen_music

Thanks for the seeds/links.

[Devalaous]

One day there'll be a super version of TNT's soundtrack. One day. Until then, its NightFright's arachno soundfont versions.

[MusicallyInspired]

Yeah, Hexen has redbook audio that's right. I also have the Saturn version as well which I think is a seperate recording itself different from the PC CD version. The problem with these tracks is that, as someone pointed out, there are many missing and the other bigger one is that they fade out which makes seamless looping impossible, which is the biggest feature of my packs. Also I believe the redbook tracks were reworked and recorded with the SC-88, which is a slightly better module than the SC-55, but which the games don't actually support. It would sound nicer but it's not really authentic to what the original games were compatible with. Also, I don't have one :). Or it possibly could be some other module(s).

The SC-55 is good, but usually when they recorded redbook audio for MIDI based soundtracks they went with something a little better and then mastered the recordings afterward. They did that with Warcraft II, the bonus track on Duke3D Atomic, and others. I'm all for quality, but if I was to make higher quality recordings I might as well just go all out and use modern realistic VSTs and samples. That's not authentic, though.

[Avoozl]

I thought Hexen 2 had the redbook audio.

[plums]

Hexen 2 also had CD audio, don't know what MIDI synth but it's one that's much higher quality than the SC-55 and family. All songs were available as CD audio, unlike the first Hexen, but some maps shared music tracks regardless of whether you used MIDI or CD music.

Hexen 2's expansion had new music and also had CD audio. The first hub reused songs from the original game, but since these songs weren't on the expansion CD, you only got them if you used MIDI playback. With CD audio, both hubs would use the new music. Source ports that allow the use of ogg files let you play back the correct music tracks at high quality in the first hub.

Heretic 2 had CD audio only, no MIDI playback at all. A shame, at least as far as "resources to pillage for use in Doom" is concerned.

This is probably more info than anyone cares about :p

[fraggle]

I don't remember anything from Quake onwards having MIDI music. Did Hexen 2 really have MIDI, not just CD audio?

[Gez]

fraggle said:

I don't remember anything from Quake onwards having MIDI music. Did Hexen 2 really have MIDI, not just CD audio?

Yes, Hexen II has MIDI. The files are a bit weird as they have a lot of silence at the end, I don't know how they handled loops.

A few of the songs are remixes of Heretic songs. CASB2.MID is derived from Heretic E1M6's song and EGYP1.MID is derived from Heretic E1M1's song, for instance.

They've been used in Doom mods, too. Scythe II uses EGYP1.MID in MAP07 and EGYP3.MID in MAP09; Epic uses CASB3.MID in MAP03, Epic 2 uses CASB3.MID in MAP10 and EGYP2.MID in MAP14.

Edit: if you open Hexen II's pak0.pak and pak1.pak in a pak archive viewer (SLADE 3 works fine), you'll find them in the midi folder. There are casa1.mid, casa2.mid, and egyp3.mid in pak0.pak, and casa3.mid, casa4.mid, casb1.mid, casb2.mid, casb3.mid, egyp1.mid, egyp2.mid, meso1.mid, meso2.mid, meso3.mid, roma1.mid, roma2.mid, and roma3.mid in pak1.pak.

Portals of Praevus adds tulku1.mid to to tulku10.mid; but I don't have that one at the moment (since it's not on Steam for some dumb reason).

[kkaden]

I'm really happy to see more people getting into this conversation, especially Fraggle. I would love to see support for digital soundtracks added to Chocolate-Everything, and hope to see support spread to all of the popular source ports.

Thanks so much to MusicallyInspired for his great work on the soundtrack for Doom (and the future recordings), and to LogicDeLuxe for pioneering (as far as I know, if someone else, props to them =) the recording work using the Sound Canvas.

And of course Thanks to Twipley for bringing this conversation to life in the original Choc-Doom bug/support thread and here in the DoomWorld forums.