alt.binaries.sounds.midi Frequently-Asked Questions "Official" FAQ Maintainer: Jason Thibeault (roguewar@nbnet.nb.ca) "Previous" FAQ Maintainers: Vikram Pant (vikram@midifest.com), Viet-Tam Luu, Dave Williss (dwilliss@microimages.com) To submit additions or changes, see Appendix B. Preface A few points/thoughts I would like to mention: * The abbreviation "a.b.s.m" has been changed to "a.b.s.midi" to avoid confusion with the group alt.binaries.sounds.mods (and possibly others). * All sections marked with a pound sign (#) are being written, and all sections marked with an exclamation mark (!) have not been completed. * Remember, this is the alt.binaries.sounds.midi FAQ; in other words, if you're reading this then this is your FAQ... CONTRIBUTE! See Appendix B for further information on how to do this. * I, Jason, recently took over as maintainer of this FAQ and will be updating it and changing it slightly. In fact, I plan on giving the bloody thing a root canal, weeding out all the bad links. But there may still be some out-of-date links or facts, please let me know right away if any such occurences spring up. Index * Introduction * Changes * 1. MIDI o 1.1 What is MIDI all about? o 1.2 What are "MIDI's"? + 1.2.1 What is GS and how is it different from GM? o 1.3 What are differences between sound files, .MOD-type files, and MIDI files? o 1.4 Why can't I convert a .WAV file to a MIDI file? o 1.5 What is a .KAR file? o 1.6 Can I convert a MIDI to a .WAV file? * 2. alt.binaries.sounds.midi o 2.1 What is alt.binaries.sounds.midi all about? o 2.2 How do I decode MIDI files from a.b.s.midi? + 2.2.1 Can I get viruses from MIDI files? o 2.3 How do I send MIDI files to a.b.s.midi? o 2.4 Some guidelines for posting MIDI files new o 2.5 How to place a request on a.b.s.midi new + 2.5.1 How NOT to place a request on a.b.s.midi * 3. MIDI playback devices o 3.2 MIDI devices + 3.2.1 Gravis UltraSound / Ultrasound Max + 3.2.2 Turtle Beach Multisound Classic + 3.2.3 Turtle Beach Multisound Monterey + 3.2.4 Sound Blaster 2.0 + 3.2.5 Sound Blaster 16 + 3.2.6 Sound Blaster AWE32 + (Add your favorite synth here) - Contribute * 4. MIDI software o 4.1 MIDI sequencers + 4.1.1 Cakewalk / Cakewalk Pro + 4.1.2 Midisoft Recording Session / Studio + 4.1.3 WinJammer + (Add your favorite sequencer here) - Contribute o 4.2 MIDISCAN - MIDI OCR Software o 4.3 MIDI-file players + LOOKING FOR REVIEWS! - To submit review follow sample + (Add your favorite here) - Contribute o Links to MIDI player Collections * 5. MIDI on the Internet o 5.1 What are some FTP sites where I can get MIDI and MIDI-related files? o 5.2 Where and how can I get the official MIDI spec's? o 5.3 Other MIDI-related addresses (WWW, etc.) * Appendix o Appendix A - Off-line sources of information o Appendix B - Contributing to the a.b.s.midi FAQ o Appendix C - List of contributors new o Appendix D - Commercially owned files not to upload Introduction This is the FAQ (Frequently-Asked Questions) document for the USENET newsgroup alt.binaries.sounds.midi (a.b.s.midi). Its objective is to answer questions that would be asked repeatedly (and answered) by users of a.b.s.midi (both new and not-so-new), and thereby free up the newsgroup to new discussion and MIDI files. This FAQ document should be posted on a bi-weekly basis on alt.binaries.sounds.midi, but probably won't, however I shall make a valiant attempt to do so. In addition to this, the most recent revision may be found at any times at these addresses: * FTP: (N/A) * Gopher: (N/A) * WWW: o United States - http://sunsite.unc.edu/pant/midi/absmfaq/ (HTML Version) o England - http://www.iprom.com/midi/absmfaq/ (HTML Version) o Russia - http://sunsite.cs.msu.su/Music/absmfaq/ (HTML Version) o Japan - http://SunSITE.sut.ac.jp/multimed/vmidi/absmfaq/ (HTML Version) o Australia - http://grissom.powerup.com.au/webmirror/midi/absmfaq/ (HTML Version) o Poland - http://sunsite.icm.edu.pl/midi/absmfaq/ (HTML Version) Contributions to this document are welcome (and needed!). No credentials whatsoever are needed, though it might be nice if you know what you are talking about :-) . For more information on how to contribute to this FAQ document, see Appendix B. Any opinions expressed in this document are solely those of their respective authors, and do not reflect in any way the views of the FAQ maintainer (unless if he is the author), or anybody else affiliated in any way with alt.binaries.sounds.midi. Changes * 1 Jul 97 -- Jason Thibeault took over as maintainer. * 17 Aug 96 -- Vikram Pant took over as maintainer. * 17 Jun 96 -- Changed extension from .shtml to .html and took out all Server Side Include things (counter and automatically upaded last mod date). * 5 Jun 96 -- Broke down into smaller files. * 13 May 96 -- Dave Williss took over as maintainer. * 6 May 96 -- Moved the FAQ to a new home. 1. MIDI --------------------------------------------------------------------------- 1.1 What is MIDI all about? MIDI stands for Musical Instrument Digital Interface. It is a standard, first published in 1983 by the International MIDI Association (IMA), that allows different musical instruments (commonly keyboards, but also guitars, violins, etc.) and devices (synthesizer modules, computers, sound cards, etc.) to be connected together. The last "I" in MIDI stands for "interface" and neatly describes what MIDI is exactly. It is a common interface, largely device-independent, that allows different devices made by different manufacturers to communicate with each other. Nothing more, nothing less. MIDI data consists of signals, in the form of a series of codes or "events" that tell an instrument or synthesizer, "start playing this note at this volume," "stop playing this note," "play this channel using this instrument sound," and so on. Of course it is actually more complex than this, but essentially the important notion to grasp is that MIDI data does not describe the sound of the instruments used in a song (in most cases, anyways), but rather how those instruments are used (i.e. played) to form the entire song. For musicians, MIDI offers many possibilities, discussion of which would be better left to someone more qualified. One example of these "possibilities" is MIDI recording, which allows a single musician to compose songs that would otherwise require several people to play the instruments. This makes the art of composing more accessible, so anybody with a decent MIDI synthesizer or MIDI-equipped computer system has the tools to put out some fairly impressive work. For us non-musicians, MIDI is another way to experience and enjoy music. Whereas the 1000 or so MIDI files in my MIDI file collection take up about 20 megabytes of storage, a single digitally-recorded (CD quality) song may take well over four times this space. (Vikram Pant's collection is well over 1500, my personal one is only 1000. I should start downloading more midis! ;) There are two solutions to this storage problem: use a medium that can store this massive amount of data in a compact, manageable form (i.e., DAT's and CD's), or make the songs in such a way that they take up much less space (i.e., MIDI files). Of course, CDDA (CD Digital Audio) and MIDI are two quite different media, and IMHO should not be compared. In other words, those who complain that a song in MIDI format doesn't sound as good as the CD version should go out and buy the CD. (Which leads to another important point: MIDI files are easily, and in most cases freely, distributed, while CD's tend to be comparatively more pricey, and basically impractical (not to mention illegal) to duplicate.) MIDI music should be viewed in its own light, not as cheap approximations "real" music. Innovations such as computers, the Internet, etc. have made it possible for "ordinary" people to express their views, their feelings and ideas on a worldwide scale; MIDI has done the same thing for amateur musicians. As for MIDI tunes that cover an existing work, it's fine to sequence a MIDI file to sound as much as possible like the original tune, but no matter how good the MIDI artist or the synthesizer, nobody can get a MIDI file to sing, nor should we expect it to do so. --------------------------------------------------------------------------- 1.2 What are "MIDI's"? Viet-Tam Luu "MIDI's" and "a MIDI" are terms that have been coined up most likely by computer users, referring to what are formally called "MIDI files". MIDI files are data files that contain a sequence (hence "a sequence," a term referring loosely to MIDI files, and "to sequence", meaning to create a MIDI file) of MIDI events. MIDI events are codes that tell a MIDI device (such as a sound card with MIDI support, a stand-alone synthesizer module, etc.), for example, to play a certain note, use a certain instrument sound to play it, play at a certain tempo, etc. All these events, put together, make up a single song or piece; some MIDI files contain a medley of several songs, although these are infrequent. 1.2.1 What is GS and how does it differ from GM? Warren Buss (wbuss@primenet.com) This is an excerpt from an original article in Electronic Musician 8/91 by Chris Meyer. Some companies feel that General MIDI doesn't go far enough, so Roland created a superset of General MIDI Level 1, which they call GS Standard. It obeys all the protocols and sound maps of General MIDI and adds many extra controllers and sounds. Some of the controllers use Unregistered Parameter Numbers to give macro control over synth parameters such as envelope attack and decay rates. The new MIDI bank Select message provides access to extra sounds (including variations on the stock sounds and a re-creation of the MT-32 factory patches). The programs in each bank align with the original 128 in General MIDI's Instrument Patch Map, with eight banks housing related families. The GS Standard includes a "fall back" system. This means that a Roland GS Standard sound module will correctly play back any song designed for General MIDI. In addition, if the song's creator wants to create some extra nuance, they can include the GS Standard extensions in their sequence. None of these extensions are so radical as to make the song unplayable on a normal GM sound module. This way, compatibility is maintained. --------------------------------------------------------------------------- 1.3 What are differences between sound files, .MOD-type files, and MIDI files? Viet-Tam Luu Sound files, such as WAV, VOC and AU files, consist of what is called "waveform data." Sound is propagated in sound waves, which are (in simple terms) waves of variation in air pressure caused by physical phenomena such as vibrating vocal cords, the vibrations of the reed as air flows through a flute, aircraft breaking the sound barrier, or anything else that produces sound. Waveform data means a digital representation of those sound waves. CD's and DAT's (Digital Audio Tapes) store digital waveform data. (Normal audio cassettes and vinyl records store analog waveform data.) "Digital" means that the analog sound data are converted into numbers in the recording process. In playback, the reverse occurs: numbers are translated back into analog signals used to activate speakers and thus re-create the original recorded sound. Thus, WAV-type files and the like truly "describe" a sound. It is important to remember that (the greate majority of) MIDI files do not actually contain any sound data. MIDI data does not, generally, describe a sound; it indicates how to play specific sounds (at certain pitches, with certain volumes, for certain amounts of time, etc.), in such a way that we can appreciate (or not appreciate :-) which we call "music." MOD-type files could be described as a hybrid of MIDI files and sound files. By this we mean that they have characteristics of both MIDI files and sound files. MOD-type files incorporate sound data, called "samples," and control codes that cause these samples to be played back as music. The actual structure of these control codes is quite different from that of MIDI events in MIDI files, but in both cases they serve to "describe" the song. Programs exist that will convert MOD-type files to MIDI files, and vice-versa, with varying degrees of success. --------------------------------------------------------------------------- 1.4 Why can't I convert a .WAV file to a MIDI file? 1.4.1 The short, simple answer Viet-Tam Luu No program exists that will analyze a .WAV-type file and "compose" a corresponding MIDI file, for the simple reason that with current technology, IT CAN'T BE DONE. 1.4.2 A few more details... Most "proposed" schemes involve operations such as Fast Fourier Transforms (FFT's) and wavelet transforms, etc. These take a .WAV-type file, generally called "waveform data," and, in the case of the Fourier transform, breaks it down into its component sinusoidal wave frequencies and amplitudes. The problem is that none of these algorithms yield the amount of information nor precision needed for the job. The problem is that when, say, the sounds of a hundred instruments in an orchestra become mixed together and are recorded as a single sound (.WAV), much of the information pertaining to each of those specific instruments is irreversibly lost. Since you can't get something from nothing, it is impossible to simply mechanically convert a .WAV back into its component sounds and then into a MIDI file. Humans can transcribe .WAV to MIDI because we are intelligent (or at least we think we are). This means that our brains, with all the previously-learned infor- mation stored in them, can take this scant information and "fill in the blanks." For computers to do this is a similar but much more difficult-to-implement process. "Difficult" simply because of the sheer complexity of the algorithms that would be required, and because of the barely imaginable computing power (that isn't yet available needed to perform all these operations at a reasonable rate. (For a more detailed analysis of the problem, see section 1.4.3.) There is of course a fairly easy way to literally convert a .WAV file to a MIDI file, with wavetable devices that can take instrument sounds saved on disk or such, for example the Gravis UltraSound and the Sound Blaster AWE32. One simply converts the .WAV file into a very large MIDI instrument sound, and then makes a MIDI file playing back that instrument at the appropriate frequency. But what would be the point, when one might as well keep the .WAV file in the first place? 1.4.3 The definitive answer Randy Lynn Tusch My background is in real-time digital audio & video. I have many designs which address many related issues currently in use around the world. My colleagues and I have written pattern-matching algorithms for the last several years and we have established several hard and fast guidelines to governing this process. (I'm sorry if I sound like I'm lecturing... I just want everyone to understand the scope of this undertaking. I hope it will prove to be helpful.) The following should help everyone to see what is involved in any digital recognition process. 1. Waveform data CAN be processed to detect OBVIOUS tempo and rhythm information by looking at the consistant spikes (quick bursts) of sound where there are quick attack and decay characteristics. (such as a drum beat.) Where a pattern is distinguishable, a tempo MIGHT be discernible. (Any irregularities in the flow could throw off the timing parameters the computer may be establishing.) 2. Any single instrument during a solo could possibly be detected as a trackable pattern for recognizing individual notes and chords provided there are no irregularities about the instrument (such as a variance in attack or differing harmonics produced during variations in volume of the individual notes played, etc.). The pattern can be filtered out of a simple mix provided it does not get "lost" in the mix and can be used to identify individual notes played by THAT INSTRUMENT (as opposed to all the other instruments in the mix [The whole point of this process!]) which is the basis for generating a MIDI file. NOW... If this sounds complicated enough, we are about to get into the REAL issue of the WAV to MIDI conversion process. Let's look into what reality brings into the picture... 1. Most music is an amalgam of several instruments (instrument sounds) mixed together to produce the final effect. 2. Instrument sounds are rarely ever "dry" in the mix. (In other words, there is almost inevitably a series of special effects and signal processing techniques applied to each instrument or instrument group such as reverb, flanging, distortion, delay, etc.) 3. Several types of sounds have sharp attack and decay characteristics which can be interperated as "percussive" in nature. 4. The overall subtleties that create the mood and "human factor" (non-mechanical feeling) of the production are the key factors used to set the velocity, sustain, and decay (note-on, sustain controller, and note off) and any related information in the resulting MIDI file. Now let's look into the ENORMOUS task the recognition process would have to undergo... Since the instrument sounds are most often mixed with other sounds, there is no "automatic" way to tell the difference between, say, rich strings, a French horn, and a guitar mixed together. True, we can detect the attack characteristic of the French horn as differing from that of the strings, but at the precise moment the attack is detected, how many other sounds are playing? Since waveform data does not identify that the sound we are currently hearing consists of, say, 3 or 10 or 50 different instruments, the recognition process records THE AMALGAM or MIX of ALL INSTRUMENT SOUNDS present at that moment AS THE INSTRUMENT THAT GENERATED THE ATTACK. After which, the computer will look for a close match to the amalgam recorded previously when another attack is detected. Since that is unlikely to occur, our French horn (with several other sounds) amalgam (or complex instrument) will likely not be perceived by the computer again during the duration of the song. Instead, the attack will inevitabley create a recording of a new "complex instrument" to be identified as yet one more instrument in the MIDI file. The overall effect is one or two notes in the MIDI file for each of hundreds of "complex instruments" over the duration of the song. (Hardly what we're looking for.) THE SOLUTION? We, as humans, can identify individual instruments in a mix of many as a result of a recorded memory of all the different instrument sounds we are exposed to over our lifetimes. When we hear a new sound, we can usually distinguish it from the other known instruments by "extrapolation" of the unknown from the known. Therefore, we can expand our memory (database) of instruments dynamically and (hopefully) perpetually. 1. For the computer to accomplish the same task, it has to be given the same database (memory) of instrument sounds we have acquired (ideally, all instrument sounds known to mankind). Additionally, we have to give it the correct artificial intelligence programming to allow its database to grow as it discerns new sounds. 2. As we may (or may not) be aware, there exists an exhaustive list of special effects (which grows as quickly as the list of known instrument sounds) to further complicate things. Even given a database of known instrument sounds, the computer hears a "dry" guitar (one with no effects- such as distortion- at all) as being a different instrument from the same guitar with distortion and reverb effects applied. Meaning, we now have to expand the computer's database to include the several billion known effect styles. Additionally, we have to program it with the ability to distinguish between new instrument sound and new effect. (I love that one!) 3. Given that we are successful up to this point, we can reward ourselves with the knowledge that we now have the basis for isolating the sharp attack characteristic of a drum as being different from the same type of attack a piano (also a percussive instrument) will generate. This does mean, however, that we must be expecting our percussive instruments to be the traditional snare, kick (bass) drum, hand clap, and any other instrument normally used to establish the overall rhythm or tempo. If a non traditional approach is used, we now have more programming to implement. Fortunately, if it is as simple as a unique sound being used to establish the meter (such as the finger snap used in "I'm Gonna Get You Baby" by Bizarre Inc. to maintain the constant rhythm.) we can simply identify to the computer that it is to track that sound and use it to generate the tempo settings for the MIDI file. But what if the song is a capella (i.e. a song made up of purely vocal melody and harmony having no instrument sounds at all)? We, then, have no instrument sounds to lock onto for the purpose of establishing tempo. Here, again, we have to do our programming to give the computer the ability to use the human equivalent of the intuitive "feel" of the song's tempo. 4. To be fair, we can use more than strictly attack and pattern characteristics to detect an instrument. We could likely use the frequency content and characteristics of an instrument sound to further increase the odds of successful recognition. (This is, in fact, one of the many additional factors used in any recognition process to narrow in on and isolate individual producers of similar patterns, such as two different people saying the same word.) 5. Of the most difficult issues to address, subtleties in the recording are probably the most difficult qualities to translate into MIDI data. Capturing controller data proposes the greatest challange to even a human musician. Of these, sustain and, possibly, pitch-bend would be the easiest to handle. Velocity information (volume of each note as it is played) is a highly subjective concept to the listener. The reason stems from the differences in each person's hearing awareness and sensitivity, coupled with the psychological biasing of the individual's social, educational, and artisic makeup as well as any preferences the listener may have. What a computer perceives may match the programmer's critera while being totally foreign to other musicians. I will leave the philosophy of this portion of our discussion for you to ponder at your own convenience. To conclude this adventure, I must add the following points and summary: ** We have only discussed four MAJOR aspects of the recognition process affecting the possibility of WAV to MIDI conversion. There are MANY more factors to consider. It should be noted that this endeavor makes even the most extravagant voice recognition project look like "baby food" by comparison. (I would like to acknowledge the success of all those who have produced successful voice recognition systems... They have beaten many odds and proved that the human mind has the potential to discover ways to overcome nearly any obstacle if the "user's" own determination and creativity flows abundantly.) In summary: 1. The database necessary to store a decent set of samples of each of the several billion known instrument sounds and samples of the billions of known special effects would take up several MILLION TERABYTES of storage. (Not to mention the allocation of storage space necessary to allow the database to grow!) 2. The number of bytes needing to be evaluated, compared, and filtered through in any one second of recorded time (not necessarily real time) again reaches a value 176400 times greater than the total size of the database mentioned in (1) above. (176400 represents the number of bytes per second in a stereo CD. You have 44100 discrete samples of sound per second times 2 bytes per 16-bit sample times 2 channels for stereo equaling 176400 bytes per second.) If the database was already several MILLION terra-bytes in size, this process would have to work with several BILLION TERABYTES for every recorded second of material. 3. Even with supercomputer technology where it is now, it would take 1000000 of the fastest supercomputers known to man (like a few Cray Y-MP's and DEC Alpha 500's w/4 processors side by side) SEVERAL WEEKS to process ONE SECOND of recorded material. Think of it this way: If you don't mind spending more than the US national debt on computer equipment and waiting a few years for the job to complete, you can have a system that MIGHT accurately convert the digital waveform data of a 5 minute song into a small, compact MIDI file. Otherwise, you can blow a couple of thousand dollars hiring a professional band of studio musicians and engineers who can probably give you what you want in about one day. --------------------------------------------------------------------------- 1.5 What is a .KAR file? Han de Bruijn A .KAR file or (MIDI-) karaoke file is just a special case of a MIDI file. Hence you can play any karaoke file with your favorite midi player, without having to change anything. (Some players require a name ending with .MID though; in that case, just rename the file.) You can see the details if you convert a .KAR file to readable text with help of the "mf2t" programs by Piet van Oostrum. (These are downloadable from ' ftp.cs.ruu.nl:/pub/MIDI/PROGRAMS/MSDOS/mf2t.zip'). .KAR files make extensive use of the Meta Text command, which is defined for standard MIDi files. A karaoke player is a MIDi player which displays the "Meta Text"s while music is playing. A decent one is "MPLAY". It is free software; the author can be contacted at henryso@panix.com --------------------------------------------------------------------------- 1.6 Can I convert a MIDI to a .WAV file? Ruediger Borrmann Try http://www.snafu.de/~rubo/songlab/midi2cs. I haven't actually tried it, but it's a shareware program. Available for PCs, Linux, SunOS 4.1.2, and NeXT 3.3. Jason Thibeault Also available, not only to produce emulated wavetable sound but to transform a midi into a high-quality (even CD-quality) .wav which can sometimes top 4 megs in length, is the great program WinGroove. It's also shareware; look up wg09e.zip on www.filez.com. Available for Windows only. --------------------------------------------------------------------------- 2. alt.binaries.sounds.midi --------------------------------------------------------------------------- 2.1 What is alt.binaries.sounds.midi all about? Viet-Tam Luu Alt.binaries.sounds.midi, or a.b.s.midi is a Usenet newsgroup that exists for several purposes, not all of which everyone agrees upon. A.b.s.midi has no official mandate, but here are a few of the things that are commonly done on it: * Distribution of new MIDI files, including both sequences of existing music and original compositions. * Requests for MIDI files, including specific requests and general requests (by certain artists, etc.). * Discussion of MIDI-related issues, including MIDI software, sound cards and synthesizers, and, believe it or not, MIDI files! :-) A.b.s.m is a non-moderated newsgroup, so it is up its users to decide what is and what is not appropriate for posting, and to keep the newsgroup free of useless, irrelevant clutter. Another newsgroup dedicated to MIDI is alt.music.midi. The biggest difference between a.b.s.midi and a.m.m is that no binaries (files) are posted on the latter; in other words, alt.music.midi is reserved for discussion only. Also, whereas a.b.s.midi tends more towards MIDI with computers, a.m.m is more about MIDI in general, with or without computers. --------------------------------------------------------------------------- 2.2 How do I decode MIDI files from a.b.s.midi? Viet-Tam Luu MIDI files are posted to a.b.s.midi through one of two different methods: UU-encoding, and MIME. UU-encoding ("UU" means "Unix-to-Unix") is simply a way of taking binary data (typically broken down into 8-bit "bytes" each having a value from 0 to 255) and transforming it into ASCII data. This ASCII data consists of printable characters: letters, numbers, and some symbols. In UU-encoding, it takes 8 ASCII characters to represent 6 bytes of binary data. This is why files take up more space when they are UU- encoded. To decode UU-encoded files, a UU-decoder utility is required. Some newsreader programs have built-in support of UU-decoding. Otherwise, the article containing the UU-encoded file should be saved to disk and then processed with a UU-decoder. It is easy to recognize UU-encoded data. It always begins with a line resembling: begin 755 myfile.zip UU-encoded data resembles lines of jumbled characters; these lines are always 61 characters long (except the line at the end of the data), and always begin with the character 'M': M0TJJ)E*J`KH0$`U%V4QI)#3)K.>F=[UI9 ... (shortened) Some UU-encoded files are split up into several posts. For example, the three sections of a three-part UU- encoded post of a file called MYFILE.ZIP might have subject lines like the following: Subject: FILE: "My Title" - MYFILE.ZIP (0/3) Subject: FILE: "My Title" - MYFILE.ZIP (1/3) Subject: FILE: "My Title" - MYFILE.ZIP (2/3) Subject: FILE: "My Title" - MYFILE.ZIP (3/3) Part "0/n" usually contains brief description of the UU-encoded file to follow. Part "x/n" means the xth part of n total parts. "Multi-part" posting is done, by those posting the files, because some newsservers and newsreaders can't handle UU-encoded data longer than about a thousand lines, or 64 kB. Some UU-decoders support multiple- file UU-encoded data; others do not. In the latter case, one must manually cut and paste the parts together. MIME is a facility that allows, among other things, a binary file to be attached to E-mail and Usenet articles. When MIME support is installed, a newsreader program encountering a MIME text will run certain scripts and/or programs to process the MIME article. In this case, a newsreader would automatically execute the appropriate programs to decode an encoded file in a MIME article. Even if MIME support is not installed, or should the MIME support fail, it may still be possible to decode the binary file attached. This is because, in many cases, the standard used by MIME to attach binary files in a MIME text is UU-encoding. Thus, the MIME text could still be saved to disk and decoded with a UU-decoder. --------------------------------------------------------------------------- 2.2.1 Can I get viruses from MIDI files? Viet-Tam Luu No. Every so often there is a panicked person who posts an article to a.b.s.midi saying that such-and-such a MIDI file posted here contains a virus, much to the chagrin and/or amusement of others. Quite simply, MIDI files contain only data as opposed to code and therefore it is impossible to get a virus from a MIDI file. (Unless one tries, for whatever twisted, perverted reason, to execute a MIDI file. :-) Though A Good Point about MIDI and viruses is brought up by Terry Carroll (carroll@tjc.com): However, you can get a Trojan Horse program. Consider the following post: Newsgroups: alt.binaries.sounds.midi Subject: Great bach fugue - Fuguegm.mid (1/1) Date: 18 Aug 1996 22:42:25 -0700 begin 644 evilprog.bat 696-H;R!497-T:6YG#0IP875S92`-"@`` ` end The subject header says it's a nice Bach fugue. However, look at what's actually encided: it's an evil program "evilprog.bat." If you use a newsreader that downloads and loads attachments (e.g., Agent), it will decode evilprog.bat, and execute it. It doesn't care what the subject line is -- it knows it's a batch file that needs executing from the "begin" line. Now, a good newsreader (again, e.g., Agent), will have an option to ask the user before executing (as opposed to loading) a file. I have that option checked, just for this reason. But a real clever guy could, for example, include a Word file that included a macro that did all sorts of nasty things. That would, quite frankly, get right by me. So a word to the wise -- although MIDI files can't encode viruses, a file disguised as a MIDI file may actually be a Trojan Horse. --------------------------------------------------------------------------- 2.3 How do I send MIDI files to a.b.s.midi? Viet-Tam Luu MIDI files posted on alt.binaries.sounds.midi are always encoded in such a way that allows binary files to be sent as ASCII text. This is done usually in one of two ways: UU-encoding and MIME attachments. UU (Unix-to-Unix)-encoding is done using a UU-encoder; different UU-encoders exist for DOS, OS/2, Windows, etc. and can be found at your favorite FTP sites. Once a MIDI file has been UU-encoded into an ASCII file, that ASCII file is then read, cut-and-pasted, etc. into the article posted to a.b.s.midi. MIME is a document format that allows attachments to accompany an E-mail message or Usenet posting. Certain scripts are associated with each type of attachment. Often, MIME binary-file attachments use UU-encode to encode binary files. The main advantage of MIME is that it's easy to use; with a MIME-supporting news- reader, one can just specify what (MIDI) files to attach and this is taken care of immediately. However, not all newsreader programs support MIME so many users don't like it. --------------------------------------------------------------------------- 2.4 Some guidelines for posting MIDI files Viet-Tam Luu Though there is no officially agreed-upon way to post MIDI files on a.b.s.midi, the following guidelines are often and generally observed by those who post there: "Subject" line of the article: Posted files usually have a subject line that begins with "FILE:", followed by the title of the actual piece (if any or if known), the file name of the MIDI posted, and the expression "(X/N)" where N is the total number of articles this posting takes and X is the sequence of this article out of those N articles. If there is a comment or introduction in a separate article accompanying the file data, this is denoted as article "(0/N)". For example, take the subject lines of several articles constituting a posting of a file called MYFILE.ZIP: Subject: FILE: "My Title" - MYFILE.ZIP (0/3) Subject: FILE: "My Title" - MYFILE.ZIP (1/3) Subject: FILE: "My Title" - MYFILE.ZIP (2/3) Subject: FILE: "My Title" - MYFILE.ZIP (3/3) Generally it's a good idea to split an article bigger than 64 kilobytes into several postings, because some newsreaders can't handle such "big" files. Compression: Compression works well with most MIDI files, because of the way those files tend to use repeating patterns, etc. Most MIDI files will shrink by 60-80% when compressed. Compressing posted MIDI files is optional, though it is quite useful when posting several MIDI songs in the same article. The compression utilities most often used, by far, are PK-Zip and Info-zip which produce compatible files. Title: It's a good idea to include the full, known title of the MIDI file being posted. It annoys many users to have to download a file with no title, only to discover they already have it, or didn't want it at all to begin with. --------------------------------------------------------------------------- 3. MIDI playback devices --------------------------------------------------------------------------- 3.2 MIDI devices --------------------------------------------------------------------------- 3.2.1 Gravis Ultrasound / Ultrasound Max Viet-Tam Luu ----------------------------------------------------- | Gravis UltraSound (GUS) | | (Advanced Gravis Computer Technologies Ltd.) | |-----------------------------------------------------| | Chipset: Gravis GF1 | | Output channels: 32 @19kHz/ch down to 14 @44kHz/ch | | Output rate/resolution: 44.1 kHz, 16-bit | | Digital input: Stereo, 44.1 kHz 8-bit | | MIDI synth.: wavetable, patches on disk (6 MB), | | 32-note polyphony, GM | | Custom sample upload: Yes. | | On-board memory: 256 kB DRAM, expandable to 1 MB | | Connections: MIDI adaptor, high-speed joystick, | | mic. in, line in, line out, amp. out, | | CD-ROM audio in | | Expansion: Proprietary daughtercard support | |-----------------------------------------------------| | Driver support: MS-DOS/Windows, OS/2, Linux | | Compatibility: Sound Blaster and MT-32 emulation | | Suggested retail price (USD): $150 | ----------------------------------------------------- ----------------------------------------------------- | Gravis UltraSound Max (GUS Max) | | (Advanced Gravis Computer Technologies Ltd.) | |-----------------------------------------------------| | Chipset: Gravis GF1, Crystal CODEC | | Output channels: 32 @19kHz/ch down to 14 @44kHz/ch | | 2 @48kHz/ch | | Output rate/resolution: 48 kHz max., 16-bit | | Digital input: Stereo, 48 kHz 16-bit, compression | | supported in hardware | | MIDI synth.: wavetable, patches on disk (6 MB), | | 32-note polyphony, GM | | Custom sample upload: Yes. | | On-board memory: 512 kB DRAM, expandable to 1 MB | | Connections: MIDI adaptor, high-speed joystick, | | mic. in, line in, line out, amp. out, | | CD-ROM audio in, CD-ROM controller | | (Mitsumi, Panasonic, Sony) | | Expansion: Proprietary daughtercard support | |-----------------------------------------------------| | Driver support: MS-DOS/Windows, OS/2, Linux | | Compatibility: Sound Blaster and MT-32 emulation | | Suggested retail price (USD): $200 | ----------------------------------------------------- Viet-Tam Luu When the GUS was first introduced into the market in 1992, it brought in a term that was virtually unknown in the PC sound card scene: wavetable synthesis. The quality of MIDI playback that it offered, at the time topped virtually every other card. Whether it is the "best" MIDI card today (short of high-end professional cards such as those by Roland or Turtle Beach) now is a matter of much debate. Still, it is the card of choice for many PC users for several reasons: its low price, compared to "rival" cards such as the Sound Blaster AWE-32 or Wave Blaster; its digital output capabilities, making it ideal for games and playing MOD-type files; software support (which has improved greatly, because of the GUS's popularity despite its biggest "drawback", the lack of hardware- level compatibility with the Sound Blaster) compared to other cheap wavetable cards. The GUS's major weakness is in digital sampling. Firstly, it is only 8-bit. Secondly, GUS's have a cut-off filter that filters out most input above a certain frequency which is quite low, as low as 6 kHz in the case of early GUS's. The addition of the 16-bit recording daughtercard does solve the latter problem. This problem has been addressed in the GUS Max, which records with 16-bit resolution, and has a much better frequency response. The GUS Max also features an "digital sound processor," actually the Crystal CODEC which adds hardware audio compression support. The GUS Max also features more RAM included, and an on-board CD-ROM controller. --------------------------------------------------------------------------- 3.2.2 Turtle Beach Multisound Classic Hussam Eassa ----------------------------------------------------- | Multisound Classic | | (Turtle Beach Systems) | |-----------------------------------------------------| | Chipset: Proteus 1/XR, Motorola 56001 DSP | | Output channels: 2 @44.1kHz | | Output rate/resolution: 44.1 kHz (max.), 16-bit | | 24 bit internal data path | | Digital input: Stereo, 44.1kHz 16-bit | | MIDI synth.: Wavetable synth., patches in ROM | | copied to onboard RAM, 32-note poly. | | On-board memory: 4 MB ROM + 4 MB RAM, no expansion | | Custom Sample Upload: No. | | Connections: MIDI adaptor, joystick, line in, aux | | in, line out | | Expansion: None | | System req.: 1 port, 1 IRQ and a 32 KB user- | | defined BIOS Address segment. | |-----------------------------------------------------| | Driver support: Windows 3.x (Win '95 rumored) only | | NO support for any other OS / env. | | Compatibility: None. Windows drivers only. | | Street price (typical) (USD): $350 (see note below) | ----------------------------------------------------- Note: Although the TB Multisound Classic is officially discontinued, It is still available in limited quantity directly from TB and large mail order outlets. The TB Multisound Classic was (and still is in many ways) the premier MIDI/digital sampling PC sound card. The Proteus MIDI engine is well regarded by many MIDI enthusiasts and it's digital record/playback performance is second to none even today. It features two banks of 384 MIDI patch sets that are mappable to the 126 GM patch set. All samples are 16-bit uncompressed. The digital audio is unmatched by anything from other vendors. It features ruler flat response (+/- .5dB) from DC to 19kHz. The distortion (<.02%) and noise levels (>86dB) are very low. The Multisound Classic did not have a wide mass appeal due to the high initial price ($1000) and the lack of any type of DOS support. This eventually resulted in TB superceding it with the Multisound Monterey. --------------------------------------------------------------------------- 3.2.3 Turtle Beach Multisound Monterey Hussam Eassa ----------------------------------------------------- | Multisound Monterey | | (Turtle Beach Systems) | |-----------------------------------------------------| | Chipset: ICS WaveFront 2115, Motorolla 56001 DSP | | Output channels: 2 @44.1kHz | | Output rate/resolution: 44.1 kHz (max.), 16-bit | | 24 bit Internal data path | | Digital input: Stereo, 44.1kHz 16-bit | | MIDI synth.: WaveTable synth, Patches in ROM | | 32 voice polyphony. | | On-board memory: 4 MB ROM, optional 4MB RAM | | Custom sample upload: Yes. | | Connections: MIDI adaptor, joystick, line in, aux | | in, line out. | | Expansion: None. | | System req.: 1 port, 1 IRQ and a 32 KB user- | | defined BIOS Address segment. | |-----------------------------------------------------| | Driver support: Windows 3.x. | | Compatibility: MPU-401. Windows Drivers. | | Street price (typical) (USD): $320 | ----------------------------------------------------- The TB Monterey is the heir to the Multisound Classic. It inherits exactly the same digital audio processing circuits and hence the same legendary sampling performance. (See Multisound Classic) The MIDI section remains somewhat controversial. The Proteus 1/XR of the MS Classic is replaced with the ICS WaveFront synthesizer. This new engine is capable of adding user-defined amounts of chorus and reverb and is quite sophisticated in that regard. It also allows the upload of samples to onboard expansion RAM although it has been observed by many that the upload speed can be frustrating. This is due to the slow serial link that is used (the Sound Blaster compatible Wave Blaster interface to the RAM). Some Multisound Classic diehards still maintain that the quality of the patches of the ICS does not come up to the level of the Proteus 1/XR patches of the Classic. In spite of this, it is still highly regarded with a strong following. Wider appeal probably will not be forthcoming due to the lack of Sound Blaster compatibility (hardware or software) and the lack of a CD-ROM interface. --------------------------------------------------------------------------- 3.2.4 Sound Blaster 2.0 Viet-Tam Luu ----------------------------------------------------- | Sound Blaster 2.0 (SB) | | (Creative Labs, Inc.) | |-----------------------------------------------------| | Chipset: Creative, Yamaha OPL-2 | | Output channels: 1 @44.1kHz | | Output rate/resolution: 44.1 kHz (max.), 8-bit | | Digital input: Mono, 15 kHz (max.), 8-bit | | MIDI synth.: FM, 11-note polyphony, GM instruments | | On-board memory: None | | Connections: MIDI adaptor, joystick, mic. in, line | | in, amp. out | | Expansion: None | |-----------------------------------------------------| | Driver support: MS-DOS/Windows, OS/2, Linux | | Compatibility: Sound Blaster 2.0 | | Street price (typical) (USD): $50 | ----------------------------------------------------- ----------------------------------------------------- | Sound Blaster Pro (SBPro) | | (Creative Labs, Inc.) | |-----------------------------------------------------| | Chipset: Creative, Yamaha OPL-3 | | Output channels: 2 @44.1kHz | | Output rate/resolution: 44.1 kHz (max.), 8-bit | | Digital input: Stereo, 44.1 kHz (max.), 8-bit | | MIDI synth.: FM, 20-note polyphony, GM instruments | | On-board memory: None | | Connections: MIDI adapter, joystick, mic. in, line | | in, amp. out, PC-speaker in, CD-ROM | | interface (Creative) | | Expansion: None | |-----------------------------------------------------| | Driver support: MS-DOS/Windows, OS/2, Linux | | Compatibility: Sound Blaster 2.0 | | Street price (typical) (USD): $75 | ----------------------------------------------------- This was a great card... when it first came out, about ten years ago. It offered what the de facto standard at the time, the AdLib card, didn't: digital input and output. It impressed many people (except Amiga users). The problem is that it's still around and being sold. No-one, but no-one, will argue that this is a good card; still, it's better than the PC speaker, which is why people still buy it. And it's cheap, dirt cheap. And the zillion clones of this card are even cheaper. --------------------------------------------------------------------------- 3.2.5 Sound Blaster 16 Viet-Tam Luu ----------------------------------------------------- | Sound Blaster 16 (SB16) | | (Creative Labs, Inc.) | |-----------------------------------------------------| | Chipset: Creative, Yamaha OPL-3 | | Output channels: 2 @44.1kHz | | Output rate/resolution: 44.1 kHz (max.), 16-bit | | Digital input: Stereo, 44.1 kHz (max.), 16-bit | | MIDI synth.: FM, 20-note polyphony, GM instruments | | On-board memory: None | | Connections: MIDI adapter, joystick, mic. in, line | | in, amp. out, PC-speaker in, CD-ROM | | controller (Creative), CD-audio in | | Expansion: Daughtercard support | |-----------------------------------------------------| | Driver support: MS-DOS/Windows, OS/2, Linux | | Compatibility: Sound Blaster 16, MPU-401 | | Street price (typical) (USD): $100 | ----------------------------------------------------- This card is basically a step up from the Sound Blaster Pro (a stereo version of the Sound Blaster 2.0), with support 16-bit playback and recording, a CD-ROM controller, and support for a daughtercard. It is not, by itself, regarded as a serious card for playing MIDI files, as all it has is a stereo version of the OPL-2 FM synthesizer chip found on the original Sound Blaster card. Software exists that allow wavetable MIDI playback with the digital sound hardware, but this is no replacement for true hardware wavetable support. To add wavetable support to this otherwise lame card, several different daughterboards are on the market, varying widely in prices and capabilities, including Creative Lab's WaveBlaster, the Gravis Ultrasound ACE, and two different Roland Sound Canvas cards. However, considering the total price of the SB16 + daughtercard, it may be more economical to consider one of the many cheap cards out there (by Logitech, Orchid, Gravis, etc.) in the same price range that offer built-in wavetable synthesis. --------------------------------------------------------------------------- 3.2.6 Sound Blaster AWE-32 Hussam Eassa ----------------------------------------------------- | Sound Blaster AWE-32 | | (Creative Labs, Inc.) | |-----------------------------------------------------| | Chipset: EMU8000, Yamaha OPL3 | | Output channels: 2 @44.1kHz | | Output rate/resolution: 44.1 kHz, 16-bit | | Digital input: Stereo, 44.1kHz 16-bit | | MIDI synth.: WaveTable synth, Patches in ROM | | and RAM, 32 voice polyphony. | | On-board memory: 1 MB ROM, 512KB RAM | | Memory Expansion: 28 MB RAM | | Custom sample upload: Yes. | | Connections: MIDI adaptor, joystick, line in, line | | out, microphone in. | | CD-ROM Support: Sony CDU-31A/33A, Mitsumi LU005 & | | FX001 Series and Creative CR-523 & | | 563 | | Expansion: Wave Blaster connector. | | System Requirements: 2 ports, 1 IRQ and 2 DMA | |-----------------------------------------------------| | Driver support: Windows 3.x, Win95. | | Compatibility: Sound Blaster-16, Adlib, MPU-401 | | (software driver) (see note). | | Street price (typical) (USD): $300 | ----------------------------------------------------- Note: The AWE-32 is software MPU-401 compatible in DOS as shipped. It can be upgraded to hardware DOS compatibility by addition of any of the available Wave Blaster compatible daughterboards. The AWE-32 is Creative Lab's entry into the mid-tier MIDI market. It appears to have been designed to provide MIDI performance somewhere between the low end FM cards and the High end cards such as the Roland SCC1, RAP-10 and the high-end Turtle Beach Multisound cards. Exactly where it actually falls in that range has fueled many sound card "wars," especially with Gravis fans. This card is really a jack of all trades and is quite capable. It features very good backward compatibility and upgradability. The EMU8000 provides the capability to add chorus and reverb in DOS and Windows. But as with jacks of all trades, it cannot be viewed as a master of any specific area. The MIDI performance is quite good but the use of either more RAM and custom patch banks or a daughter board is really necessary for exceptional performance. Either of these options adds to the cost significantly. The digital audio section is very good but falls short of that of the Turtle Beach products. All in all, this is a good card for good quality, general purpose use with full Sound Blaster compatibility and good upgradeability. --------------------------------------------------------------------------- 4. MIDI software --------------------------------------------------------------------------- 4.1 MIDI Sequencers --------------------------------------------------------------------------- 4.1.1 Cakewalk/ Cakewalk Pro --------------------------------------------------------------------------- 4.1.2 MIDISoft Recording Studio --------------------------------------------------------------------------- 4.1.3 WinJammer --------------------------------------------------------------------------- 4.2 MIDISCAN - MIDI OCR Software Steve Adams I picked up MIDISCAN at a computer trade show for $99. The fellow had his scanner and computer set up to give demos. After letting me pick a piece on my own, we digitized a Clementi sonata. Sounded pretty good. He must have had a Pentium. I have a Compaq Deskpro 386/25 here at work and an IBM 486SLC2/66 at home. Neither of my computers "converted" a file as fast as his did. In fact my home computer locked up several times on a Telemann fugue. Tried it at work also, and to my surprise it locked up also. Memory is not a problem, my home machine has 16 meg and my work machine has 8. Must be something with that particular piece. Have tried a variety of music the past several days. Mostly classical. Did a Telemann fantasia (3/4 page size), a Telemann fugue, a one page Beethoven piece and the Brahms waltz. The Telemann fantasia converted pretty well. I sequenced this piece before, and MIDISCAN did a decent job. As mentioned, couldn't convert the fugue. The Beethoven was the simplest, and was converted the most acurately. How it works: One needs to scan the sheet music first. I only have access to a Sharp flatbed scanner hooked up to a Mac quadra here at work. Used Photoshop. MIDISCAN only reads TIFF files. Photoshop can output this format, just be sure to save without compression. MIDISCAN cannot handle LZW compression. Also if using a Mac, save in PC TIFF format, not Mac TIFF format. Then dumped the TIFF files to floppy. (Here a procedure tip: Each 8x10 scanned sheet of TIFF format is approx 1 MB in size. LZW will reduce that to about 150 kB. What I did was save on the Mac with LZW. Put a whole bunch on floppy, read on the PC and used my PC version of Photoshop to resave without LZW.) Open MIDISCAN (a Windoze program) and tell it which TIFF files to convert. If a piece of music is more than one page, (or several TIFF files, no problem, just list them all, in order to be converted). MIDISCAN then takes over. On the 486SLC2/66 it takes about 3-5 minutes to convert each TIFF file. On the 386, about 10-15 minutes. On the fellow's Pentium (I assume) it only took about a minute. MIDISCAN then displays in split screen fashion both the original TIFF scan and its interpretation. They are displayed "side-by-side" so that one can see each measure for easy comparison. One then has the option to use the built-in editor to "clean-up" the intrepreted file, (and some needs to be done.) The editor is fairly easy and intuitive to use. One does not need the manual to run this program. Finally, another drop-down has the option to save as a MIDI file. (Again, the two files included are the raw intrepreted version, and a slightly cleaned up version. The next step would be to read in the MIDI file into a sequencer for final clean-up.) My impressions on how MIDISCAN did: not bad. All text is ignored. The manual recommends that one "delete" the text from the TIFF file before the conversion step. They include a simple TIFF editor to do this. It consists of identifying the upper-left and lower-right corner to delete. I ran my test with all text as part of the files. No problem. The program did indeed ignore text, except in a couple of cases where finger numbers were intrepreted as a note. Most key signatures were intrepreted correctly. It had a problem with the time signature, but that was easy to fix. Seems to have a problem some times distinguishing between a quarter note and a half note. Again, pretty easy to fix, and occured only about 2-5% of the time. In this Brahms piece, slurs were intrepreted as ties. Again, easy to edit out. All dynamics are ignored, so these will have to be put in by hand in your sequencer. Accidentals are picked up, but I question its accuracy of intrepretation. One thing I noticed that it did not pick up is the stacatto notes. It pickes up the note, but not the stacatto indication. I tested version 1.1c, which is the latest available. Got a coupon for upgrade to 2.0 The only difference is that 2.0 includes TWAIN support and another minor one that I cannot remember now. The upgrade will cost me another $50. (Not sure I'll get it.) Finally, I guess the program did okay. As far as a user getting music into MIDI, it falls someplace between someone who knows how to play the piano real well, and can use his/her sequencer to record while they are playing and someone (like me) who has to step-enter the notes one by one. I'll probably continue to use the program for the first cut at getting the music into the computer, but not enough to buy a scanner for myself at home. All in all, quite a programming feat. Remember, this really hasn't been done on a commercial scale before. MIDISCAN is the first to pull it off. They deserve credit for that, even if the program is not 100% accurate (technology development has to start somewhere.) --------------------------------------------------------------------------- 4.3 MIDI-file players --------------------------------------------------------------------------- MIDI Players Looking for very brief info on MIDI players for DOS, Apple, Windows, UNIX, Amiga and more. --------------------------------------------------------------------------- Sample Review Player's Title Author/Creator Platform Version Number Freeware/Shareware Size (if linking to file directly) Link (to website/file location) --------------------------------------------------------------------------- Links to MIDI Player Collections Title: The MIDI Software Shack from MIDI-Fest Notes: MIDI players from most all platforms URL: http://sunsite.unc.edu/pant/midi/software_shack.html Title: MIDI Players & MIDI Plugins on the Net Notes: Good place to download many MIDI players/plug-ins URL: http://www.aitech.ac.jp/~ckelly/midi/help/midi-players.html --------------------------------------------------------------------------- 5. MIDI on the Internet --------------------------------------------------------------------------- 5.1 What are some FTP sites where I can get MIDI and MIDI-related files? These sites may still be out of order, I've not yet had the time or opp- ortunity to check them. If you wish, please check these sites, and e-mail me with the changes or information on each. Thanks! Site/address: ftp.uwp.edu Directory: /pub/music/lists/kurzweil URL: ftp://ftp.uwp.edu/pub/music/lists/kurzweil Notes: Kurzweil K2000 stuff. Site/address: ftp.funet.fi [128.214.6.100] Directory: pub/msdos/sound/cakewalk URL: ftp://ftp.funet.fi/pub/msdos/sound/cakewalk Notes: Cakewalk (.WRK) files. Site/address: mort.isvr.soton.ac.uk Directory: /pub/pc/cakewalk URL: ftp://mort.isvr.soton.ac.uk/pub/pc/cakewalk Notes: Cakewalk stuff. Accessible through WWW http://www.isvr.soton.ac.uk/People/ccb/Cakewalk. Site/address: mitpress.mit.edu Directory: /pub/Computer-Music-Journal URL: ftp://mitpress.mit.edu/pub/Computer-Music-Journal Site/address: ftp.mcc.ac.uk Directory: /pub/cubase URL: ftp://ftp.mcc.ac.uk/pub/cubase Notes: Cubase archive. ite/address: ftp.waldorf-gmbh.de URL: ftp://ftp.waldorf-gmbh.de/ Notes: Information, patches, other MIDI stuff. Site/address: ftp.und.ac.za Directory: /pub/pc/midi URL: ftp://ftp.und.ac.za/pub/pc/midi Notes: MIDI files, programs for Sound Canvas and MT-32. Site/address: musie.phlab.missouri.edu Directory: /pub/korg URL: http://musie.phlab.missouri.edu/pub/korg Notes: Information on Korg products. Accessible by WWW (http://musie.phlab.missouri.edu/pub/korg/) Site/address: ftp.rc.tudelft.nl Directory: /pub/midi URL: ftp://ftp.rc.tudelft.nl/pub/midi Notes: MIDI site for original MIDI compositions only. --------------------------------------------------------------------------- 5.2 Where and how can I get the official MIDI spec's? The most recent specifications for the MIDI standard may be obtained, for a modest fee, from the International MIDI Association, at the following (snail-) mail address: International MIDI Association 23634 Emelita Street Woodland Hills, California 91367 USA --------------------------------------------------------------------------- 5.3 World-Wide Web MIDI Stuff URL: http://www.inetnebr.com/~williss/midi.html Notes: MIDI page, lots of links Contact: David C. Williss (dwilliss@microimages.com) URL: http://www.servtech.com/public/jglatt Notes: MIDI Technical Fanatic's Brainwashing Center Contact: Jeff Glatt (jglatt@servteck.com) URL: http://www.aitech.ac.jp/~ckelly/SMF.html Notes: Links to virtually every MIDI site on the internet Contact: Charles Kelly (ckelly@aitech.ac.jp) URL: http://stud1.tuwien.ac.at/~e8925292/bestmid.htm Notes: The "Very Best of GUS MIDI" Collection Contact: Gerd Reichinger (e8925292@student.tuwien.ac.at) URL: http://www.harmony-central.com/MIDI/ Notes: WWW document devoted to MIDI Contact: Scott Lehman (slehman@MIT.EDU) URL: http://www.prs.net/midi.html Notes: The Classical MIDI Archives Contact: Pierre R. Schwob (prs@prs.com) URL: http://www.MIDIFest.com/ Notes: Vikram's MIDI-Fest - over 100MB of MIDI's Contact: Vikram Pant (vikram@midifest.com) This place is the very best! I can't plug it enough. Thanks, Vikram, for setting up such a great site! --Jason --------------------------------------------------------------------------- --------------------------------------------------------------------------- Appendix A - Off-line sources of information Bibliography Many books have been published on MIDI and MIDI- related issues. An extensive MIDI bibliography is maintained by Piet van Oostrum (piet@cs.ruu.nl). Unfortunately, due to its size it cannot be included here. The latest version of the bibliography can be obtained by FTP from ftp.cs.ruu.nl [131.211.80.17] in the MIDI/DOC/bibliography or by E-mail from mail-server@cs.ruu.nl (send a message with HELP in the body). Official MIDI specifications See section 5.2. --------------------------------------------------------------------------- Appendix B - Contributing to the a.b.s.midi FAQ Vikram Pant (vikram@midifest.com) Send submissions to the FAQ by email to vikram@midifest.com and start the subject with "MIDI FAQ:" Please don't send me email asking where to find something or how to configure some sound card. If I know, it's already in the FAQ. If you have any questions about the a.b.s.midi FAQ, don't hesitate to E-mail me. Questions about MIDI-related topics should be sent to a.b.s.midi or better yet for technical help try alt.music.midi, not me; I don't claim to know anything much about MIDI. Please don't send me the entire FAQ if you have added or revised a section, just that section. Links are NOT accepted. The A.B.S.MIDI FAQ is not going to link to every single MIDI site on the internet but only to a select few. Pages with organized links are acceptable. Unless you are writing about a specific MIDI device/ synth, don't play favorites, i.e. don't "bash" other sound cards. Be objective. The above also goes for people, as for sound cards. Discriminatory statements/remarks, overt or implied, of any form will not be tolerated. Enough said. As the FAQ maintainer I reserve the right to make editorial changes/corrections to submissions, or to refuse them outright. I will not, however, add edited submissions to the FAQ without the permission of their authors. Unless I make drastic changes. --------------------------------------------------------------------------- Appendix C - List of contributors The alt.binaries.sounds.midi FAQ has been made possible through the contributions of the following people (if your name isn't here, but should be, send E-mail to roguewar@nbnet.nb.ca): * Steve Adams * Han de Bruijn (rcpshdb@dutrun2.tudelft.nl) * Hussam Eassa (eassa@earth.execpc.com) * Mark Johnson (ymj@pacbell.net) * Terry Law (law@legend.pl.my) * Viet-Tam Luu * Piet van Oostrum (piet@cs.ruu.nl) * Randy Lynn Tusch (drt@rec1.roc.servtech.com) * David C. Williss (dwilliss@microimages.com)