How did designers get such usable guitar/pianos/ep/even strings out of tiny sound modules?

Anything about hardware musical instruments.
RELATED
PRODUCTS

Post

I'm genuinely perplexed by this. I am in love with some of the proteus 2000 sounds, and the entire composers rom it comes with is only 32 MEGABYTES.. yes mega bytes! How?

The JV 2080 had 8MB and i used it non stop and i listen to my mixes from then and i LOVE THEM!

maybe, just maybe, my stuff with software synths has become a bit too sterile and clean.. perhaps the noise from these lower bit converters actually added mojo..?

But still, i found all the sounds very playable.. unlike huge patches of some current romplers drenched in verb, many of the old romplers just use the lightest touch of verb, and layer beautifully into a song.

Perhaps these multi gigabytes of realism demanded today, really just doesn't matter at *all* in electronic music, and only for film music?

I'm having more fun now than i have had in many many years.

Is it also that these guys were just masters at creating seamless loop points for sustained presses? Obviously velocity levels are limited, yet again, i love playing the emu guitars. :shrug:

Post

I don't have a lot of experience with them but I always just assumed it was a combination of well looped waveforms, no velocity layers, and clever use of EQ, dynamics, and effects, all done by expert programmers like Eric Persing adding up to very unrealistic, but sugary sweet instruments that work in mixes. Listen to the Flaming Lips album The Soft Bulletin, or the Mercury Rev albums from that same time period. My understanding is that those were done on the Orchestral expansion board of some Roland ROMpler, and those records/sounds blow me away.

Post

so well looped waveforms is probably the big one as i suspected, and only one or max two velocities..

eq /dynamics.. good point...


and of course expert programmers.

So those programmers are probably the best of all time really, hey.

Looking for some flaming lips! looking forward to it :) Cheers

Post

TheoM wrote:so well looped waveforms is probably the big one as i suspected, and only one or max two velocities..

Looking for some flaming lips! looking forward to it :) Cheers
The Flaming Lips and Mercury Rev were both psychedelic indie-rock bands that were being produced by the great Dave Fridmann. They both took kind of an orchestral turn around the same time and I'm fairly certain that a Roland ROMpler + Oschestral Expansion board was the basis of all the orchestral sounds on these albums.

These aren't always the best songs on those records but here are some good examples of the orchestral sound from the old Roland ROMpler used on these albums...

The Flaming Lips:

https://www.youtube.com/watch?v=mVsyJtCsqeA

And some mercury rev from that period...

https://www.youtube.com/watch?v=Y_2c_E_c-U0

https://www.youtube.com/watch?v=M94tvKB-pL4

Post

Maybe they're not neccesarily run these 32-megabyte samples but rather synthesize sound at run time? Huge sampled libraries are relatively modern trend, early engineers tried to simulate behavior of insytruments with algorithms and circuits, for instance Karplus-Strong filter.
Blog ------------- YouTube channel
Tricky-Loops wrote: (...)someone like Armin van Buuren who claims to make a track in half an hour and all his songs sound somewhat boring(...)

Post

Its down to talented people who worked with what they had at the time. Considering the technological and financial restraints, they did a fine job.

Beats todays 'designers' hands down... no preset morphing or randomization then :uhuhuh:

Post

Honestly I think a great deal of it is the amount of polishing AND the DAC.
Having the proteus 2500 rom card and having various Emu incarnations there is something about that rom card in a command station which is magical before effects are added.
Synapse Audio Dune 3 I'm in love

Post

Six years ago, I had a chance to talk to Timothy Swartz of Digital Sound Factory. Along with some other talented designers, Tim made many sounds for the Ensoniq and E-mu samplers and ROMplers in the late '80s and early '90s.

I complimented him on this very thing, and tried to find out how the E-mu Proteus modules sounded so good when they had less sample memory than NI was using for one piano sample. Tim was gracious with his time and information. He said basically the same thing you've mentioned here: creative looping; judicious use of EQ and compression; and in some cases, using samples from one instrument as the basis for another instrument (which reduces the number of samples you need in the first place).

That said, not all of the sounds in those early ROMplers were necessarily "great". I'm not sure how well some of the sounds in the Korg M1 (for example) would work in a mix today. Some of the piano sounds are a bit dated, and I've found some of the organ sounds to be rather harsh in the higher registers (or "sloppy" in the lower ones) or not quite as expressive as some newer synths. Then again, some are classics. It all depends upon the sound you're trying to get.

Still, I love those early ROMplers and learning about the hoops the designers had to jump through to get as many good sounds as they did out of them.

Steve
Here's some of my stuff: https://soundcloud.com/shadowsoflife. If you hear something you like, I'm looking for collaborators.

Post

I don't know that those old ROMplers sound "great" when compared to modern Kontakt libraries. Certainly the pianos were all crap by comparison. But that isn't really the point. Sometimes you need an enormous, excruciatingly detailed sample set in order to recreate an exact replica of an original instrument. Then again, sometimes you don't. If all you're really looking for is a cool and interesting sound, that can start with just about anything. After all, how many great sounds have been made with only a singe cycle sawtooth waveform? I once sampled a single violin note from old vinyl ("A child's introduction to the orchestra" or something like that), threw it into Kontakt with some filter modulation and a couple of other effects, and I ended up with an amazing string synth type of patch. With that sort of methodology in mind, it isn't surprising that the raw material in a JV-1080 (which I also still own) can be used to make some great sounds. Also, don't forget that the output of those old ROMplers were often run through some great outboard effects. A PCM-70 can work wonders on a mediocre sound.
Incomplete list of my gear: 1/8" audio input jack.

Post

deastman wrote:I don't know that those old ROMplers sound "great" when compared to modern Kontakt libraries. Certainly the pianos were all crap by comparison. But that isn't really the point. Sometimes you need an enormous, excruciatingly detailed sample set in order to recreate an exact replica of an original instrument. Then again, sometimes you don't. If all you're really looking for is a cool and interesting sound, that can start with just about anything. After all, how many great sounds have been made with only a singe cycle sawtooth waveform? I once sampled a single violin note from old vinyl ("A child's introduction to the orchestra" or something like that), threw it into Kontakt with some filter modulation and a couple of other effects, and I ended up with an amazing string synth type of patch. With that sort of methodology in mind, it isn't surprising that the raw material in a JV-1080 (which I also still own) can be used to make some great sounds. Also, don't forget that the output of those old ROMplers were often run through some great outboard effects. A PCM-70 can work wonders on a mediocre sound.
great post. Just to clarify one thing, I never meant they sounded better than modern libraries, i meant, how on earth did they sound so good and playable with what they had to work with.

Post

the attack is strongly weighted, if that's convincing, the ears kind of believe the rest.. first impressions.. when you're using sounds instead of judging their finesse, how much more of the sample do you hear...

..works for mixing as well apparently.. raising the volume just at the beginning of a part adds attention/apparent intelligibility instead of making it feel buried/masked/swamped. "pro mixing tip" :p

could also be a relative level of quality because improving the quality of samples doesn't necessarily do much to improve the quality of using samples so perhaps there is "more animation" in these simple sounds being relatively "sensibly/skillfully" augmented with synth architecture than there is in the starkness of a huge sample library.
you come and go, you come and go. amitabha neither a follower nor a leader be tagore "where roads are made i lose my way" where there is certainty, consideration is absent.

Post

I haven't looked at modern expensive software samplers nor auditioned modern high-price sample sets.

Some years ago was playing with the final version of gigasampler before it went obsolete. Giga had some good features, but IMO capability for giant numerous samples played direct from HD was the main strength it was exploiting. Compared to the nicer earlier hardware samplers, giga didn't have near as many low-level parms with which to fine-tune each sample in a multi-sample set.

Dunno what is "under the hood" in a proteus ROM multisample or roland multisamples. Generally presented as a single block one can choose, though if you pick a certain multisample, would take some black-box reverse engineering to find out how many samples are in each black box of a multisample. Maybe under the hood, EMU, Ensoniq, Roland inherited some of the many fine-tweaking individual sample controls that were in their samplers, with the details written to rom and hidden from the user?

There were just many low-level ways to fix-up a sample set, smooth it out, on the hardware sampler, compared to giga. For instance, on EPS samplers you could fine-tune a sample pitch, and then independently fine-tune the loop pitch, to avoid the sample "going out of tune" when it hits a short loop.

Without the low-level sound engine adjustments, lots more individual sample editing needs to be done before loading into the sampler for final tweaking of the set as a whole.

Though if you can use as much hard drive memory as you want on a sample set, maybe its not worth messing at all with tight loops.

Sample quality control was generally good on the romplers. At the moment I don't recall samples in romplers that played out of tune. A reason I've not tried out modern high-end software samplers or libraries, other than frugality and don't have time to fool with it-- A few years ago I checked out some sample libs that to me seemed expensive, though it is relative. Spend more than the price of a dinner and a movie, it is expensive. Heck, nowadays a dinner and a movie is expensive.

Some of the sample libs from a few years ago just had bad quality control IMO. Or maybe better termed inconsistent QC. Pianos sampled from instruments that obviously hadn't even been properly tuned before sampling. If yer gonna spend hours sampling a grand piano, heck at least hire a piano tuner first! Sample sets with mis-matched level and tone between sub-samples. Out-of-tune sample sets. Great care needs to be taken with short-looped samples to get the loop precisely in tune with the rest of the sample. But even with long un-looped samples or long loops, at least tweak the long loop so that it doesn't sound like a washing machine when its playing the long loop. Even with un-looped samples, at least level-match the samples, tone-match the samples, properly tweak the heads and tails, and TUNE the damn thangs. :)

For instance if you sample a guitar, even if the guitar is perfectly in-tune and the player is playing as in-tune as he can do it, for some reason when you slice up those samples and put them in a sample set, they still need pitch fine-tuning, and to my ear sounds awful if this is not done. It is an art to make a sample set playable. One can throw together a sample set where if you play any individual note, it sounds great, but when you try to play music on it, doesn't work at all. Takes expertise to turn the samples into a playable instrument.

Maybe modern software sound libs are impeccable and don't have such problems. Just in the past, looked at NFR sample libs and thought, "They want $200 for stuff that doesn't even play in-tune?"

So anyway, so far as I can tell, it is rare to see gross quality control problems in rompler samples. Maybe it doesn't suit the taste, or maybe the fidelity isn't good enough because of limited memory, but at least what is there, is usable.

Post

Because Eric Persing

Post

i don't think persing had anything to do with Emu? Perhaps I am wrong.. but proteus 2000 series (they were all released as cards and then each card also had it's own module, that could also hold three other cards, for example, any of those modules could hold the proteus 2000 rom, called the composer rom. I have the composer/rob papen techno synth construction/ planet phatt/ xtreme lead (my second favourite after composer's).

Also, Persing did some of the sounds for roland not all.. only he knows exactly how many. The point is, omnisphere is NOT doing regular rompler stuff, so how do we even know if Eric programmed regular acoustic instruments for roland, or if he did the more esoteric pads and ambience rhythms like the classic "flying waltz"?

Emu under the hood, by proteus 2000 stage was z plane filters (50 types) and 32 mb sample rom slots. ALL of emu's modules of that era (if not all???) were sample based entirely. I have no idea where they sourced most of their stuff from. The z plane filters were one of the best for the time, sounding reasonably juicy on electronic stuff, but i suspect the main reason for good playable guitars,pianos etc was just damn good efficient sampling as discussed, with seamless loop points. I never had a problem playing long sustained notes with any rompler really from the old days.. those guys just knew what they were doing (including of course Eric Persing). I think that probably answers a lot of the question.. there were extremely talented people there that were experts in getting the most out of almost nothing (in this case, the nothing being 32 megabytes of rom).
I honestly wonder how some of the sound designers of today would fare when faced with such restrictions.

I can link some sound on sound reviews of emu modules if anyone is interested, i read all of them :)

Post

I thought I remember reading Dave Rossum saying something about using his custom chipsets to compress the data, and since it was a custom chip it was fast enough to do on the fly. I can kinda buy into that, as I had the soundfonts of all the proteus and planet phatt sounds and found them very underwhelming. My brother has an original proteus and I am amazed at how great it sounds for 4mb(?) of data.

Post Reply

Return to “Hardware (Instruments and Effects)”