Soundscape DAW review for Summer’98

Includes Model 8 I/O and Version 2.02 software

ã 1998 by Eddie Ciletti
 
Soundscape Digital Technology Inc.
4478 Market Street / Suite 704
Ventura, Ca. 93003
vox: 805-658-7375
fax: 805-658-6395
http://www.soundscape-digital.com/
e-mail:  us-sales@soundscape-digital.com
 
APPLICATION
The un-crash-able Digital Audio Workstation
SUMMARY
Powerful yet easy-to-use.
STRENGTHS
Ten channels of HD record per unit. Twelve tracks of HD playback plus ten "live" inputs can also be mixed. Analog, AES/SPDIF, TDIF and MIDI interface. (see "Accelerated System," below)
ACCESSORIES
"8 I/O" provides 8 analog I/O plus ADAT optical. $X,xxx.
PLUG-INS
In-house plug-ins ( The Audio Tool Box ) include dynamics, delay-chorus-flange ($325), time-pitch-sample-convert ($199) plus third-party: TC Reverb ($699) & Dynamizer ($799), Wave Mechanics Reverb ($349) and Synchro Arts VocAlign ($995). 
The TC Dynamizer has 3 band compression, expansion, soft clipping, with look ahead peak detection.  Like the Time/Pitch module, the Synchro Arts VocAlign  is tweaked for dialogue synchronisation and music applications.
WEAKNESS
Can new features ever come fast enough? 
Options

(see below)

List Prices and Features
Accelerated "System"
$4,495 (consists of HDR-1 base unit + AC-1 accelerator)
Twenty-four track 

Package # 1

$12,400  for two "systems"  with ALL plug-ins, but not including Dynamizer ($799) and Vocalign ($1000) or Aphex Big Bottom (Sept). 
Hardware I/O option #1 20-bit "8-I/O-1"  $ 1,995  ADAT & TDIF, 8-XLR & Word
Hardware I/O option #2

available September

24-bit "8 I/O-2"  $  600  8 in/out 8-RCA with ADAT & TDIF, Word & internal clock. 
Allows clock selection from all sources.
Hardware I/O option #3 20 bit "8 I/O-3"   $  600  8 in/out 8-RCA with ADAT & TDIF, Word & internal clock.
Twenty-four track 

Package # 2 

 $12,700 with two "8 I/O-1" with no plug-ins except basic mixer

Ever wonder whether reviews are "slanted" or not?  This one is. I’ve been a Soundscape user ever since reviewing the workstation four years ago. Back in the day, when sound cards were "mere toys," I started with eight tracks in real time. Since then, I purchased two units — which Soundscape has since accelerated — allowing a life-long dream to be fulfilled: twenty-four tracks in real time with real-time effects.

Before you begin to think this is just an advertisement in sheep’s clothing, however, consider that I’ve also seen Soundscape naked. I know what still turns me on in the control room even though I also know what’s hiding in the closet.  Nothing too scary, but in the transition from version 1.18 to V2.xx, the CD software lost some of its functionality.  Automation exists in a very basic way but not via moving faders.

STRIP TEASE

All new technology is a bit of a tease at first. At under $1,000, a sound card is appealing for the price alone.  "When used as directed" sound cards will deliver a few audio tracks to compliment a MIDI sequence.  They are perfect for stereo editing and mastering.  They can also import tracks from a digital recorder for "surgery."  But really mixing multitracks in real time with processing on every channel is another story.  Sound cards rely on the host processor and every other link in the computer chain PLUS they must share resources if slaving to a sequencer.  By contrast, Soundscape can play a 24 track mix with Photoshop running ( for last-minute tweaks to CD cover artwork ) and while Microsoft Word is open to this document!  Can you do this on your system?

OUT OF THE POOL

From real world experience, I’ve seen both sides. On the cheap, some dive into the shallow end of sound-card pool only to stumble out with a headache. Still in a daze, they go to the other extreme dropping big bucks on that "other" system that doesn’t run under the Wintel flag.  Soundscape falls in-between, appearing less sexy partly because of the price (when compared to a card), perhaps because it lives outside the computer and unfortunately because that "other" system gets all the press.  Knocking the competition is a cheap shot.  But for about $10,000 street price — twenty-four tracks of Soundscape, with plug-ins, seems like a great deal. Hello?  Is everybody missing something out there?

So, while frustrated users have been futzing with sound cards or aging-on-hold for customer service, I’ve been living large.

I do long-form multi-track mixes, a real challenge for a card-in-a-computer, but not for Soundscape. Its dedicated external hardware talks directly to affordable IDE drives (4GB for under $200) relying on the computer for interface not for power.  Soundscape is not only "surround ready,"  the sample rate conversion
algorithm from its "Time Module" can already convert 48kHz files to 96kHz, so that a 24-bit WAV file export will be possible for DVD (if and when
required).

Soundscape works so well out of the box it is surprising that people haven’t caught on.  My multitrack music application may not be the norm — what with MIDI and Sampling  — but I also know several users who take advantage of Soundscape’s ability to sync to picture.  All of the latest software can be downloaded from their web site; the EDL (edit decision list) software being updated on a regular basis.  ( Click for Sidebar provided by "post" user, Alan Silverman.)

MAJOR HARDWARE

Soundscape Version 2 begins with the HDR-1, the original two-space rack unit with an AC-1 accelerator. In addition to a T-DIF port (for direct digital connection to a DA-38, -88 and –98), there is a stereo input and four-channel outputs — both in analog and digital flavors per unit. A second two-space box, the 20-bit "8 I/O," houses ADAT optical, TDIF plus eight analog ports in and out.


Image One:  Soundscape's own Dynamics Module

THE SOFT TOUCH

The most current software is Version 2.01 (final) and Version 2.02 (alpha). The mixer includes built-in EQ, delay and effects sends. Users have complete control over the mixer including number of channels, bands of EQ, sends, etc. The in-house plug-ins include dynamics, pitch, time and delay modules. (Image OneTC Electronic have ported low-fat versions of their Reverb and Finalizer (Image Two); Wave Mechanics make a very powerful reverb plug-in (Image Three).


Image Two: The very beautiful low-fat TC Electronic Reverb Plug-in

If you haven’t noticed already, the really sexy plug-ins suck digital petrol like a big-finned fifties Cadillac. The clever DSP designer can trim the fat and keep the meat. Soundscape does this pretty well, but with third-party plug-ins it becomes obvious that the taste is in the fat. That's because dedicated hardware — a  reverb unit, for example — does one job well without distraction.  When ported over to the "virtual world,"  workstation DSP faces regular "interruptions" and must perform multiple operations all within 1 / 44,100 th of a second!  (I don't think multitrack workstations with 96kHz sampling will be practical and affordable in the near future.)  Mixing, cross-fades and edits are easy compared to EQ, Dynamics and Reverb.  In a "real" hardware-equipped studio, users have to determine when the "free" effects are "good enough" and when to rent the serious outboard gear.


Image Three:  The extremely geeky but powerful Wave Mechanics Reverb Plug-in


FACING REALITY

The two currently available reverb plug-ins could not be more disparate, a plus for users!  The TC Reverb sounds and looks good, with an easy-to-use interface to parameter tweaks.  The reverb "tail" is split up in 3 frequency bands each with adjustable decay time.  There are seperate predelays for initial reflection and the tail. A high frequency filter for the input signal and six visible room shapes.

The Wave Mechanics Reverb offers many more parameters — described in raw DSP terms — which can only be described as "completely geeked-out."   While the essential features are "similar" between the two products, you could hardly tell by looking at them.  The Gate provided on the Wave Mechanics reverb is great for those hard-to-get "short, tight room" sounds for drums.  It is, however, quite a power hog.

Soundscape has two Motorola processors: one for record, playback, editing and synchronization (56001) while the other handles the Mixer (56301). When the Mixer Window is open, the available resources are displayed as "P" (DSP Processor resources) and "M" (DSP Memory resources).  Realize that for a twenty-four track configuration, the "display" indicates the combined resources of both systems in a "worst-case" scenario.  As such, the displayed values are not a precise reflection of "power consumption."  My "typical twenty-four track mix"  — with lots of EQ and dynamics — uses up 92% of the P-DSP and 48% of the M-DSP. Table One makes this easier to visualize that, in this case there is enough headroom for one fat-free TC module but not for the Wave module.

It’s the same in the analog world: our beloved tube gear is the least efficient way of turning power into sound.
Click here for a comparison of directX and hardware-dedicated plug-ins.
 
 
 
Plug-in use of resources
P-DSP
M-DSP
w/o Reverb w/24 active tracks
92 %
48 %
TC Reverb w/24 active tracks
99.5 %
56.2 %
Wave Mechanics Reverb w/24 active tracks
103 %
71 %

Table One: Comparing plug-in use of resources and exceeding same!

THE IMPORTANCE OF BEING CLEVER

The short term work-a-round is to either turn off tracks or create a temporary mixer sole for the task of creating an effect.  Either reduces the system requirements long enough to record the effect and sub-mix if necessary.  There are three long-term solutions: MIXTREME, a new PCI card from Soundscape, a DSP expansion chassis and a third complete system which, for some, would be unpractical.  Another option (for the Wave Mechanics plug-in only) is to process the reverb "in the background," that is, in non-real time.

MIXTREME ($549 with V2 Mixer software or $699 also with S/PDIF in/out) uses the same DSP and mixer software as the Soundscape AC-1 Accelerator for the HDR-1.  It features two TDIF ports for 16 channels of live mixing.  It was not available at the time of this review.  Soundscape does provide an expansion port on the AC-1 but has not yet created a host chassis for DSP modules. Either would be welcome because the mixer really is great.  When combined with automation and an external controller this system will be hard to beat. At present, I love the fact that a detailed mix can be saved and Totally Recalled with effects!   (Mixtreme can also used with Logic Audio, Samplitude, Cubase VST, Sound Forge, Cool Edit etc.)
 

RECENT ADDITIONS

The most recent software features 99 separate Solo and Fader groups plus channel modules that employ Equal Power Panning (3 dB lift from center instead of the "Equal Voltage" lift of 6dB.) A complete manual, written in HTML form, includes hyperlinks for each chapter and many related features. Any browser can be used to open and navigate the manual, which is quite detailed.  I have only recently cracked the manual because the software has always been so easy to use. But after a full year of Version Two, there are many new features that are making my life easier — the simple reward of reading!

CONTROL SURFACE

Soundscape has always supported both the JL Cooper CS-10 and the Penny & Giles DC-16 MIDI controllers. I was surprised at how easy the CS-10 came online. It connects to the MIDI ports (in and out) of any sound card.  Clicking on a Soundscape / JL Cooper ICON opens a tiny controller window that confirms operation (faders move, knobs turn, buttons respond) and voila, you’re ready to take control.  Not all features of Soundscape’s mixer are available via controller, a downside because my right hand is beginning to suffer. (No Viagra jokes, please!)  EQ and editing are two features I would surely like to access with traditional knobs and a scrub wheel.

There are many new controllers, the Mackie HUI and the JL Cooper MCS 3000 series to name a few. The latter includes the CS-10 command subset as well as the HUI/Pro Tools "personality file."  At the moment, however, Soundscape barely takes full advantage of the CS-10.

RIGHT 4 U

If Soundscape works for me, it should definitely handle anything you can throw at it.  I have taken the liberty of "pushing" a dual system which takes up a mere "4-U" of rack space.  You will likely use less tracks and have plenty of room for plug-ins. You don’t have to be a super-geek or super-rich to have this kind of mixing and editing power.  Soundscape has never let me down and has survived four years without becoming out-of-date. What other company can make that claim?

One other note about Soundscape.  Software upgrades have always been free.  Expect to see an eight track 24 bit upgrade in October. So has most service.  Consider that some people pay "extra" for an extended warranty, a gamble in many cases.  A workstation will always need new software.



 
 
 


SIDEBAR: POSTING WITH SOUNDSCAPE

ã 1998 by Alan Silverman

I’ve been using Soundscape’s SSHDR1 to edit and post-mix audio for TV music specials such as "An Evening With Harry Belafonte and Friends,"  "A Judy Collins Christmas, " and "The Tunes of Tommy Dorsey – A Sentimental Journey."  I have two units which comprise a 24 channel system with 16 channels of Tascam TDIF digital I/O.   The TDIF ports are especially useful since it allows me to keep all audio transfers in and out of the system in the digital domain.  I receive elements on DA-88 at 48K sampling, and deliver the final mix plus Dialog / Music / Effects ( D/M/E ) splits on DA-88 for (hopefully!) digital transfer to the final 48K broadcast master, usually D-2 or Digibeta.  I really find that staying digital all the way results in a much better on-air product.  The DA-88 has become a standard for audio post, so having a TDIF friendly workstation turns out nicely.

For working against picture, For working against picture, the SSHDR1 is an extremely fast system to use and locks to incoming SMPTE time code from a Betacam or ¾" video deck master in an instant. (SMPTE to MIDI conversion is via MOTU's Midi Timepiece AV.)  Although the system can work with a digitized AVI video file from the host PC’s hard disk, I generally prefer to work with a video tape.  The picture quality is better to look at while mixing, and it’s easier to keep up with picture revisions that have a pesky habit of showing up while I’m mixing.

I find the mixing tools already built into the Soundscape system are comprehensive enough so that I no longer need to use an external automated console for gain riding.   The tools treat the audio as "clip based."  Any sound file on the system’s drives can be divided into parts.  A part can be positioned anywhere in the master time line and given it’s own volume level and a variety of fade curves.  It’s a simple but elegant way of working that’s very fast and flexible.  Once you get a dynamic roughed in, it’s really easy to adjust by sliding or trimming the part boundaries.  A real benefit to this method, is that if  (when!) a picture change is made during the mix process, it’s a simple matter to insert or delete time and have all the down stream levels and volume moves follow along.  Accomplishing this with an external SMPTE based mixer requires a lot of careful off-line editing of automation data and is very time-consuming and error prone.

The system also has a complement of audio processing tools that work well. Multi-band EQ is part of the basic package, and features such as time/pitch processing, reverb, and compression/expansion are available as plug-ins.   I also use the SSHDR1’s built-in digital mixer and AES digital I/O to get audio to an  external Weiss digital parametric EQ and compressor / limiter / de-esser.  This is superb gear that I use primarily for CD mastering work, but it’s nice to be able to bring that level of quality signal processing to TV sound tracks as well.

Both of my Soundscape’s are fitted with removable drive bays that use inexpensive, off-the-shelf IDE drives available for well under $300 for a 6GB drive.  It’s a simple matter to have multiple large scale projects going on at once when all you need to do is turn a key and swap a drive or two!

At completion of the edit/mix, I’ll switch the master time code source over to the SMPTE output of the DA-88, lay off my mix and splits digitally via the SSHDR1’s mixer page and TDIF port, and pack off that neat little tape for digital layback.

The project can then be archived by a couple of means.  The system will back itself up to DAT via the AES port using a proprietary data format, but this ties up the hardware.  I prefer to transfer the files from the SSHDR1 drives to the host PC drive using a fast built-in function.  The PC can then back up to a SCSI data DAT drive in the background while continue to use the SSHDR1 for other work.  In fact, I find this kind of multi-tasking to be a major advantage of a PC based system.  I can edit audio (MOTU's Midi Timepiece control panel will also be running in the background), check my e-mail, keep a session log in a word processor or spread-sheet, and be backing up a prior project all at the same time.  You know, it really works!



 
 
 

Return to where you left off?

Go to the Top?



 
 
 


On Plug-ins and Direct-X

Since TC Works (a subsidiary of TC Electronic) is writing plug-ins for Soundscape, they have a good handle on the system's internal performance, something us "mere users" can not fully appreciate.  I visited the Soundscape Users Forum in June'98 and found some dialog between users and DSP designers.  The information that follows was based on that dialog.  It was graciously provided by Thomas Block, Ralf Schluenzen and  Sven Duwenhorst - all of TC Works - and edited by Eddie Ciletti of Manhattan Sound Technicians and EQ Magazine.  The material is not to be used without permission. ã 1998 by Eddie Ciletti, Thomas Block, Ralf Schluenzen and  Sven Duwenhorst.

So what is DirectX?

DirectX - part of the Microsoft Windows system architecture - is in charge of the communication between software and low level hardware components. DirectX is structured into different parts like DirectShow, DirectDraw and DirectInput, each taking care of one specific area like video, audio, etc.  DirectShow is designed to stream any kind of multimedia data such as Video and Audio via MPEG software decoder, including video for the internet.

Lately, DirectX has established itself as an application independent "Plug-In interface" on the PC - with the big advantage that any Plug-In written to DirectX specs can be accessed by any application which incorporates the interface. So it doesn't matter if you use Cubase, Logic, Cakewalk, CoolEditPro, SoundForge, WaveLab or Samplitude.  All applications will be able to use all DirectX Plug-Ins!

A look at DSP-based processing versus DirectX "native"

So you might ask, "Why bother buying a dedicated-hardware system like Soundscape, if you can do it cheaper by going native?"  (Native means using the host PC's processor to perform audio calculations.)  Well, plug-in compatibility isn't everything.  "Dedicated Hardware" and "Native Processing" have their pros and cons.  With Soundscape, you get nearly 0 ms latency and wonderful stablility because it is independent of processor activity. The latency of DirectX is like playing guitar through a 200 ms delay.  But one "pro" and one "con" each is not enough to give either the brush-off, so read on!
 
 
 
 
DirectX PRO's:

The internal audio processing is performed in 32bit floating point format using the host processor. 

Advantages of this format: 

  1. The levels in the audiostream do not hard-clip at 0dBfs, instead they're specified up to +6dBfs. Remember this is "virtual processing headroom" unlike hardware, so there won't be any audible distortion in the processing stream as long as the level at the end of the processing (plug-in) chain is scaled back to 0dB. (Of course, the A/D and D/A levels must be properly set.)
  1. The internal processing on an INTEL Pentium processor is performed with 64bit accuracy.   Theoretically, that results in an amazing signal to noise ratio.  The real SN depends on the signal applied and is not constant. But a general rule in audioprocessing is that a close estimate of the SN can be
 
calculated by: 

Number of relevant bits * 6dB = SN in dB.  A 16-bit system * 6dB = 96 dB SN 

So much for the theory.   In the real world, executing multiplications creates noise due to the finite representation of numbers. Algorithms with feedback-type filters (not FIR) and reverbs can create a lot of rounding noise because the signal is circulated thru multiple operations. Now say that the noise we have created is at -106dB. In a fixpoint system this noise will stay at -106dBFS and be relatively independent of the signal. In the floating-point system, the noise will move with the signal (though, it will always stay below the -106dB in our example), and a moving noise floor is more apparent to the ear than a fixed one. This is just one of the reasons for using 64bit floats internally when calculating filters and stuff, as we do.

  1. It is easier to implement powerful filters - like 56 Bands of EQ in one Plug-In - much easier than on Fix-point DSP systems,  because the developers do not have to struggle with rounding error effects on filter coefficients, which would cause filter instability such as clipping and all sorts of other not-so-nice things.
  1. DirectX plugins can run on any application that supports the specified interface.

 
 
 
 
DirectX CONs:
 
  1. DirectX is not specifically designed for hi-end audio processing. It is a part of the Microsoft OS, designed to handle all sorts of multimedia streams (Audio, Video, compressed, uncompressed, etc.,).  Naturally, this "open-ness" introduces quite a  heavy programming overhead - having features we don't need while lacking others.
  1. Audiostreaming and signal processing share the same processor (CPU) with the operating system. Since Win95/WinNT are designed as multi-purpose operating systems, they are not specialised to handle real-time tasks or to prioritize audio processing at all given times.
  1.  I/O Latency: Due to the points mentioned above, the audio signal path is pretty long.  (For example: the soundcard input + soundcard driver + audio application + plug-in + audio application + soundcard driver + soundcard output) and sometimes reminds us of chinese whispers - however not quite that bad. (Would someone explain this analogy to me?  ec)  But the result is a latency (delay) of roughly 200ms. When using bad cards with bad drivers this can easily double.
  1. Even though Windows is running on most PC compatibles, the actual performance on each user's computer is sometimes unpredictable - it is nearly impossible to tailor the performace requirements of a plug-in so that it runs comfortably on all potential configurations.

 
 
 
 
DSP PROS 
  1. DSP, as the name implies, is a piece of hardware specially designed to process digital signals.  The whole architecture of the chip assists the developer in realizing highly optimised algorithms. It's a given system with a given performance: the chip architecture and performance are known. An algorithm can be perfectly adapted to work in this environment at all times. 
  1. The whole hardware environment of a DSP (in audio systems like Soundscape, for example) is designed and optimised to handle the processing in real time. This means that the only latency introduced in the audio signal path will be the conversion time of the AD/DA converters. 
  1. The DSP's internal operating system is tailored to that specific piece of hardware and optimized for a special purpose (audio and/or video). Generally, the DSP's OS is reduced to the absolute miminum, making it much easier to develop, test and debug than a Windows NT system - with it's massive overhead. That's why external, DSP-based audio systems are generally very stable in terms of crashes and other unpredictable behaviour. 
  1. Well implemented DSP based systems can be slaved to external digital systems with sample accuracy.  This is not the case in the "native" environment, unless hardware expansions and special drivers allow for it. 
  1. An external audio system like Soundscape uses the computer just as front end, minimizing the workload to such a degree that it can be run on a relatively slow computer ( a 486 for example).  Soundscape users will note that on the occassion that a system crash occurs, Soundscape will keep on running.  The system is that stable!

 
 
 
 
DSP CON's

A modern, fast DSP does the internal processing in 24 bit fixed point format. 

This format has two drawbacks: 

  1. On a Fix point DSP, the audio signal clips mercilessly at 0dBfs and introduces digital distortion (even when there's just a marginal overload). So keeping the levels under control is a crucial part of development for this platform. 
  1. All DSP systems (so far) are proprietary systems - so plug-ins created for one system won't run on any other.  Development is relatively expensive, as we have to adapt each product for a specific hardware environment. 

In conclusion, we at TC follow the principle of trying to go for the highest possible quality on any supported platform.
If we feel a platform isn't good enough, we wouldn't support it.

*****************************  TC|WORKS  *******************************
                          Flughafenstrasse 52B * D22335 Hamburg * Germany
         Phone +49.40.531.08 30   Fax +49.40.531 08 31  E-Mail ralfs@tcworks.de
************************  A TC Electronic company  ***************************
 
 
Return to 
"Plug-in Reality?"
e-mail 
Eddie
Return to 
the REVIEW List?
Eddie's Home Page
Go to Eddie's Directory