[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference napalm::commusic_v1

Title:* * Computer Music, MIDI, and Related Topics * *
Notice:Conference has been write-locked. Use new version.
Moderator:DYPSS1::SCHAFER
Created:Thu Feb 20 1986
Last Modified:Mon Aug 29 1994
Last Successful Update:Fri Jun 06 1997
Number of topics:2852
Total number of notes:33157

938.0. "Sampler Features Discussion - .6 Contains Summary" by DREGS::BLICKSTEIN (Dave) Wed Sep 09 1987 12:29

    Having heard some of the COMMUSIC III submissions, I'm really
    interested in getting a sampler.
    
    I've spent only a short time thinking about it and have already
    determined that there are a few gross of issues I know about in
    choosing a machine, and probably 5 times as many that I don't
    YET know about.
    
    A brief list of some of the more important issues I have thought
    of are:
    
    1) Availability of software (current AND future)
    
    2) 8 bit vs. 16 bit
    
    3) Polytimbral vs. monotimbral
    
    4) Support software
    
    5) Rack mountable (I'd like to avoid having another keyboard)
    
    6) What is on the horizon (new machines expected?)
    
    My question to the experts out there are:
    
    1) What other issues should I be aware of
    
    2) What's out there?  What does it do?  How does it sound?
    
    3) Of course, recommendations (accompanied by supporting data)
    
    So far as I know, my minimal requirements are polytimbrality and
    rack mountability, and availability of sounds) but the most important 
    factor for me is REALISTIC sound.
    
    I do not expect to be creating many samples of my own, nor do much
    extensive editing.
    
    I know that there is a slight advantage for me in getting the Mirage
    in that I can use its disk to store data from my ESQ-1 but that
    advantage does not weigh heavily against most other factors for
    me.
    
    So, enlighten me on samplers.  Consider me to be a total
    novice/illiterate when it comes to sampling.
    
    	db
    
T.RTitleUserPersonal
Name
DateLines
938.1WHERE IS IT !?!?!?!?!?!!CANYON::MOELLERWed Sep 09 1987 13:2615
    Dave, you might check out a topic & replies in this conference titled
    "Assiduously Seeking Sampler". I originated it and had much the
    same set of criteria as you - didn't want another keyboard, wanted
    a large library readily available, etc. I knew whatever unit I got
    I'd keep for a loooong time, so it's sort of high-end consumer
    time..     Considered the Mirage, the Prophet 2002 (rack), AKAI S900, 
    E-Mu Emax, etc. I settled on the Emax. 
    
    The info is about a year old and doesn't discuss some
    of the newer/smaller samplers, Akai S700, Roland S10, etc., but
    all the units considered are still on the market.
    
    karl moeller
    
    p.s. ... about COMMUSIC III, Dave.. uh, how's it coming ?
938.2Don't leave zone without it...JAWS::COTENote stuck? Try Kawai...Wed Sep 09 1987 13:2712
    
    
                    M U L T I S A M P L I N G
    
    ...unless you're into Darth Vader and chipmunks. Stretching one
    sample over 5-6 octaves doesn't cut it, no way, no how. You want
    to put multiple samples on the keyboard, each stretching only
    part way. 
    
    Don't by a machine without it.
    
    Edd
938.3Stream of consciousness tutorial.ACORN::BAILEYSteph BaileyWed Sep 09 1987 13:4485
    You forgot 12-bit.  Most of the semi-pro current samplers are 12-bit:
    The Akai, Prophet 2000+, Roland S-10, Emulator-II I know are, and I
    vaguely recall that the E-Max, and DSS-1 and the Ensoniq are also.
    
    8-bitters are dead.  You can get 12-bits worth of electronics cost
    about the same as 8-bits worth these days.  The Emulator-I was 8-bit.
    
    12-bits is enough for most applications, but usually when they make
    a 16-bit sampler, they also spring for higher qualitiy components
    all around, so it's hard to say ``16-bits definitely sounds smoother.''
    All 16-bit samplers I have heard (FZ-1, D-50, Fairlight) sound
    noticably better than 12-bit ones.  But seem to cost at least two
    times as much--it's more than bits you are paying for.
    
    One useful spec associated w/i samplers is sampling rate.  For most
    applications this is more important than dynamic range (bits). 
    Obviously, 40+ KHz is good to shoot for as an upper rate (most samplers
    have several sampling rates), but most units give about 35 KHz max.
    
    The best thing to do to sort out sample rate and dynamic range
    questions is to listen to the beast.  Get a range of sustained
    sounds--cymbals, strings, vocals and pianos are usually what the
    salesman can produce for any given sampler.  Ignore ``sound effect''
    samples for this purpose.  You can't learn much from them.
    
    For other features, you have to decide what your application is.
    
    If you want to simulate real instruments, multisampling is a must.
    Multisampling is the provision to play different samples for a set of
    ranges on the keyboard.  Playing a single piano sample, transposed over
    an 88-key range sounds horrid (as I'm sure you can imagine).  Further,
    you probably want velocity cross-fade.  This is like multisampling w/i
    key velocity as the input.  For a piano, you would like to get as many
    ranges as you can.
    
    And to cope with all these samples you want MEMORY.  Just multiply the
    length of the longest (non-repeating) sample you need by the number of
    multi-samples, by the number of different velocity samples by the
    sampling rate.  Simple.  Say 4 sec * 4 sample * 4 velocity * 40000Hz
    * 2 bytes/word = 5 Megabytes.  Aaaagh!
    
    But you can get by with much less by finding a point in the sample
    at which you can loop until the release of the note, and use the
    envelope generator (yes, samplers have EGs, and VCAs, or whatever
    you want to call them) to provide the decay.
    
    In absolute terms, with current technology,  decent amount of memory is
    1 Meg, but look at what this means in terms of sampling time and
    number of samples.
    
    If you're doing drums, you probably want separate outs.
    
    If you want to do covers of Genesis songs (with spike-driving noises),
    you won't need many special features capabilities.  Just sample
    and play.  This works great for serendipidous ``real-world-noise''
    as new instrument stuff.
    
    Polytimbral capability seems intellectually pleasing, but I can't
    really see the need for it unless you want to simulate an orchestra.
    That is, with multisampling, you can get many different timbres
    without having to receive the notes on different channels.  Seems
    like six of one/half dozen of another to me.
    
    Now, personally, there are several things which I would consider
    musts for a sampler that I buy:
    
    	1) Support of MIDI sample standard.
        2) NO quick disks (see note on MFD-1 for comments on this)
    
    Note that Roland give its owners all new samples for free.  The
    Roland upgrade policy is also very nice (it cost $5 for the latest
    S-50 upgrade).  Their stuff is rack-mount.  You should probably
    wait for the rumord S-220 if you want a ``lower-cost'' unit.  The
    S-50 and S-550 cost a blue mint (but sound awesome).
    
    Of course the Mirage is the DX-7 of sampling in terms of availible
    sounds.  Billions and billions sold.  I'm sure you can even get a
    sample of the kitchen sink for it :-). 
    
    Hope some of this is relevant.
    
    Steph
    
    
    
938.4start collecting that literature, guyCANYON::MOELLERWed Sep 09 1987 19:1350
    This table might help organize the various features of the various
    available samplers. It was obviously used only to distinguish
    between my two (then) front-runners, but could indeed accomodate
    more models and features. Some of you know what MY winner was.- karl
    
    Note 603.28             Still Seeking Sampler Assiduously   
    I'll do this by category, and attempt a 'score' per category, with 
    total.                       ***  0=low, 5= high.  ***
Score/Feature    
    Voices:
[5]	EMAX: 8. 
[5]    	S900: 8.
[5]    	EMAX: stack two voices, no polyphony loss. 
[3]    	S900: stack two voices, polyphony loss.
    Availability:
[1]	EMAX: 3wks-2months. Real Soon Now. 
[5]     S900: NOW. Like, IN STOCK.
    Library:
[4]	EMAX: factory disks already in stores. Emulator ports likely.
[1]]    S900: Crappy, highly dependent on individuals. 
    MIDI Implementation:
[5]    	EMAX: each preset can be assigned to a specific MIDI channel
    		for sequencing.
[5]	S900: each preset can be assigned to a specific MIDI channel
    		for sequencing.
    Audio Outs:
[5]    	EMAX: 8 outs, 2 stereo w/programmable panning, mono
[3]    	S900: 8 outs, 2 stereo, programmable L/R only, mono
    MIDI Note range:
[5]    	EMAX: 128 notes logical, 88 notes physical
[5]    	S900: 88 notes logically/physically
    Price:
[3]	EMAX: $2700 no discounts incl library
[5]    	S900: $2400 discounted   incl library    
    User Groups/Support:
[4]    	EMAX: anticipate major factory support/ user group
[0]    	S900: after one year, nothing anywhere.
    Users Manual:
[4]    	EMAX: ~200 pages, by Craig Anderton. Very thorough. 
[2]    	S900: ~45 pages. Adequate. 
     Memory:
[5]    	EMAX: 512k of 8-BIT words is just the same as	
[5]	S900: 750k of 12-BIT words   
     Unique Features:
[2]    	EMAX: Full feature arpeggiator. 
[3]    	      Internal sequencer can take entire sequence from external 
    	      unit in one pass. Good performance feature.
[2]    	S900: auto pitchbend .. great for sax/string bass, etc.
    
    The envelope please :  EMAX: [51], S900: [42] ! ~20% gap... 
938.5Esperen!!! Me too!!!ESCUDO::FASTOrlando Saez SGO/#5 AME Soft GuruThu Sep 10 1987 19:5821
        Hey 'yo tambien tengo la misma duda...!!!'
    
    It happens that I have the same concern about sampling keyboards
    as Dave. The only thing I'll like to add is: 
    938.4, .3, .2, .1 agree with Emax, S900, Mirage... (Please not 
    the Mirage DSK). Well what about the market price in terms of
    differences over the time. I'm not sure about this, but I think
    if a keyboard for prices for $2,000 and after a year I can find
    it for $1,300, it isn't a good reference. I would think that its
    technology ran out very fast, or simply that there is other that
    offer better features for the same lower price. Don't you think?
    
    Well, let me know... remember I'm in the same knowledge stage about
    sampling that Dave.
    
    ROBOMUSICMAN
    
    Orlando
    
    
938.6Sampler Feature Chart...FSBIC1::DDREHERE/P in trainingTue Sep 15 1987 19:3857
Back from vacation.  My wife just had a baby boy, Jonathan,  9lbs 15oz.

I'm now entering the great sampler hunt also.  I've been waiting on the
Roland S-550 but I haven't seen it around yet.  This chart came from the
latest Roland Users Group.  This issue is just full of info on
Rolands latest stuff (D-50, D-550, S-550, S-220, MT-32, TR-626, new 
MC-500 and S-50 software, etc.).  Check it out.


I think I want a rack mount unit.  The serious candidates so far have been
the S-900, Emax, and (hopefully soon) the S-550.





Product          S-550      S-50         Akai S-900    Emu Emax        Korg DSS1
--------------------------------------------------------------------------------
Voices           16         16           8             8               8
Keyboard/Rack    R          K            R             K,R             K
No. of samples   64         32           32            122             16
Sampling rate    15k,30k    15k,30k      3 to 40k      42k             15,32,48k
Sampling time    7.2x4 sec. 7.2x2 sec.   12 sec.       28 sec.         5.5
Freq response    13k        13k          16k           19k             20k
Resolution       12-bit/w 16-bit proc.   12-bit        8-bit companded 12-bit
Video interface  Yes        Yes          No            No              No
MIDI channels    8          4            16            8               1
Outputs          8          4            10            8               2
Digidesign       Yes        Yes          Yes           Yes             Yes
Velocity X-fade  Yes        Yes          Yes           Yes             No
Velocity switch  Yes        Yes          Yes           Yes             Yes
Mono mode        Yes        Yes          Yes           Yes             No
Sounds provided  Yes        Yes          2             15              4 systems
Library          Sound Bank Sound Bank   Small         Large           Small


But wait, there is more...

Product          Mirage        Casio FZ-1      Emulator II       Sequential 2002
--------------------------------------------------------------------------------
Voices           8             8               8                 8     
Keyboard/Rack    K,R           K               K                 R   
No. of samples   8             64              64                16
Sampling rate    33k,50k       9k,18k,36k      32k               20k,31k,40k
Sampling time    2x2 sec.      14.56 sec.      17.5 sec.         6.2 sec.
Freq response    21k           16k             13k               19k
Resolution       8-bit         16-bit          8-bit companded   12-bit 
Video interface  No            No              No                No    
MIDI channels    8 (mono)      8               1                 8 (mono)     
Outputs          1             9               8                 2     
Digidesign       Yes           No              Yes               Yes   
Velocity X-fade  Yes           Yes             Yes               Yes   
Velocity switch  Yes           No              Yes               Yes   
Mono mode        No            Yes             No                Yes   
Sounds provided  17 disks      2 disks         10                4   
Library          Large         Medium          Very Large        Medium 
938.7SALSA::MOELLERShe's my Black Magic MarkerTue Sep 15 1987 20:4219
    Welcome back, Dad�. A small embellishment to the chart regarding
    Emax' variable sample rates and time available... also, multiple
    samples in memory can have separate sample rates.. a cymbal needs
    much higher a rate than the bass drum/toms, for example.
            
       RATE        TIME     ~FREQ.RANGE
    10kHz	52.0 secs	5kHz
    16kHz	33.5 secs	8kHz
    20kHz	26.0 secs       10kHz	
    28kHz	18.8 secs	14kHz
    31kHz	16.6 secs	16kHz
    42kHz	12.4 secs	21kHz
    
    Also.. an important point for me, the ability to 'stack' two samples
    under one key with NO loss in polyphony... switchable via Positional
    Crossfade (PXF) or velocity. The S900 and 2002 can stack, but suddenly
    are only 4-voice machines.
    
    km
938.8BARNUM::RHODESWed Sep 16 1987 09:419
Product          S-550      S-50         Akai S-900    Emu Emax        Korg DSS1
--------------------------------------------------------------------------------
Freq response    **13k**    **13k**      16k           19k             20k

You mean to tell me that the S-550 and the S-50 only have a 13K bandwidth????

Cross them off the list...

Todd.
938.9Happy Baby!LOLITA::DIORIOWed Sep 16 1987 11:2211
    
    re: .6
    
    Congratulations Dad! 
    
    You listed the Mirage as having 8-bit resolution (implying 8-bit
    linear) when in reality, isn't it also 8-bit companded (like the
    E-Max)?
    
    Mike D
                         
938.10Worthless Data, for your AmusementCTHULU::YERAZUNISdepleted uranium speaker cabinets?Wed Sep 16 1987 18:5614
    From Ye Olden Rumor Mille and Catahhr Cure: 
    	
    	Expect a new sampling product from Ensoniq soon.
    		
    	price: about 1.7K.  Features: @kitchen_sink
    
    	Probably have a Mirage compatibility mode, but it's *at least*
    	16 bits internally.  Yes, the rumor is that it may be more than
    	16 bits.  Why?  I dunno.  Maybe it shares the sample memory
    	with a �VAX-II PMI?  :-)
    
    	sorry, no other data.   
    
    	I've already put myself in line for one.   
938.11a sceptic in the systemJON::ROSSMicro-11: The VAX RISCThu Sep 17 1987 12:0911
    no way. Price too low. They already have that market. And
    why compete with yourself.
    
    Seems right that its time for another entry from them tho.
    
    16 bits is hot. 
    
    I buy it. 'cept for the price....who said? where info from!
    
    ron
    
938.12SALSA::MOELLERShe's my Black Magic MarkerThu Sep 17 1987 13:249
>You listed the Mirage as having 8-bit resolution (implying 8-bit
>linear) when in reality, isn't it also 8-bit companded (like the
>E-Max)?     Mike D

    I'd like to hear from our several Mirage owners, but in my own
    listening tests, there are (read my lips) M A J O R differences
    in fidelity between the two units.
    
    karl     
938.13Maybe I got ca-ca speakers...JAWS::COTEIt's A Glamour Profession!Thu Sep 17 1987 15:343
    ...but are the differences attributable only to the format?
    
    Edd
938.14oops!LOLITA::DIORIOFri Sep 18 1987 14:2612
    re .12   It was never my intention to compare the fidelity of the
    Mirage to that of the E-Max, and I certainly don't want to start
    a Mirage vs E-Max war here! I was merely asking
    if the Mirage had an 8-bit *companded* format as opposed to an
    8-bit *linear* format. I probably shouldn't have mentioned the
    E-Max at all... Sorry for any misunderstanding.
    
    PS  for the price difference between the E-Max and the Mirage, there
    had better be (and I'm sure there are) M A J O R  differences in 
    fidelity (and everything else) between the two units!!
                              
    Mike D.
938.15let's you & him value differencesSALSA::MOELLERFri Sep 18 1987 14:528
    Well, Mike, sorry if I jumped on your case. The newer Mirages MAY
    indeed have some kinda 8bit companding scheme. E-Mu have been quite
    coy about the Emax' real resolution, repeatedly stating it is the 
    'equivalent of 14bit linear'.. whatever that means. We do all agree
    that there is LOTS besides byte resolution and sampling rate 
    differentiating various samplers.
    
    karl
938.168 bt cmpndng...JAWS::COTEWould I lie to you, Honey?Fri Sep 18 1987 16:085
    To my knowledge, ALL Mirages use 8 bit companding. They claim
    'effective 12-bit resolution', whatever THAT means. I take it to
    mean 'better than linear 8'.
    
    Edd
938.17lets resolve this.JON::ROSSMicro-11: The VAX RISCFri Sep 25 1987 10:1412
    
    And Kurzweil is 18bit floating point (Orin?)
    
    Companding is a way to stretch the dynamic range but
    I think you do pay in resolution....the range is 
    broken into segments, say 3 bits for 8 segments, then
    you have each segment resolving to 5 bits...
    
    It works but you dont get everything for free...

    rr