Cakewalk Pro Audio 9

PLEASE NOTE: This article has been archived. It first appeared on in September 1999, contributed by then Editor-in-Chief Rip Rowan. We will not be making any updates to the article. Please visit the home page for our latest content. Thank you!

Cakewalk recently released the new version of its flagship product, Cakewalk Pro Audio. The new version, Version 9, reflects some new thinking on the part of the design team at Cakewalk.

To be sure, this legacy application looks and feels almost exactly like its predecessors. It shares a virtually identical user interface to versions 7 and 8. Only a few new notable features and enhancements have been included. And all major enhancements are under-the-hood additions.

So what’s new?

What’s new in version 9 is a renewed commitment on the part of the Cakewalk design team towards three critical concepts: stability, usability, and performance.

Repeat after me: Stability. Usability. Performance. When software companies ask what you want in a new piece of software, recite your new mantra. Stability. Usability. Performance.

Stability refers program uptime and predictability. Does it work, or does it crash? If I change my hardware or add new software, does the system get all fouled up and require a reinstall or a call to the help desk? Do features work sometimes but not at other times, with no indication as to why they are not working? Can you trust the tool in a critical situation?

Usability refers to the program’s predictability and implementation. Usability is more than feature sets and user interfaces. Usability describes whether the software supports your process. Does the program perform in a consistent manner from function to function? Are features implemented in a logical manner that reflect the way work actually gets done? Are important functions buried deep into the user interface? Does the user have to perform gymnastics to get routine tasks done?

Performance refers to the system’s efficiency and capability. Performance refers to the ability of the system to process your commands, and is easily measured for audio software in terms of throughput and latency.

To be completely frank, I was not at all happy with Pro Audio’s previous two versions, Version 7 and Version 8, because I did not feel that their stability, usability, or performance was up to par.

However, Pro Audio Version 9 reflects major improvements in all of these areas, and, perhaps, makes Pro Audio a tool worthy of consideration for demanding, mission-critical applications.



The feature set of Pro Audio 9 remains largely unchanged over Version 8. Many minor-to-semi-important new features have been added, but anyone familiar with Version 8 will feel right at home with V9.

Pro Audio 9 – Track and Console Views

plugins, MIDI plugins, full drag and drop implementation, built-in digital video, SMPTE / MTC sync, volume and pan envelopes, a fully automatable virtual mixer, and the list goes on and on. Read Cakewalk’s website if you want more details. Suffice to say, if you can’t get the work done with PA9 it’s probably not because you’re missing a feature. Everything you need to write, record, and mix music is here, except the talent.

So let’s look at what’s new.

Stereo Tracks PA9 now supports stereo audio tracks – great for recording and mixing. Stereo tracks are great for people like myself who routinely submix groups of tracks to simplify the mix. Sometimes I will consume 48 audio tracks for a single 4-minute pop song! With 24 bit audio, it makes good sense to bounce tracks down to a small handful of submixed tracks for easy handling. Tracks can be mixed to mono, stereo, or pair-of-mono tracks. Stereo track support also allows makes Pro Audio useful as a post-production and mastering host.

Stereo track support also allows users to make full use of stereo plugins. Now, instead of running two DirectX plugins – one for each channel – users can run one stereo plugin, improving efficiency and allowing access to special features such as stereo-linked compressors, sidechained compressors, vocoders, and the like.

Mixdown Mixdown features have been enhanced, allowing users to mix down any part of a song either to Windows WAV file, or to a submixed track. This enhancement makes it brain-dead easy to mix down your song to your hard disk (in standard WAV format, or in Real Audio G2 or MP3 Internet formats). You can also use the command to bounce a group of tracks to a new submixed track within Pro Audio. Many options are provided to include or exclude the real time effects, mixer automation, aux sends, etc.. Former V8 users will be very glad to see these improvements, as Mixdown functionality has been treated poorly in past versions.

AudioX Cakewalk introduced a new technology – AudioX – in this release of Pro Audio. AudioX is a convention that allows any AudioX software package to control DSP functions on AudioX-compliant hardware. Yamaha’s DSP Factory, along with products from Sonorus and Digital Audio Labs support the new format. AudioX exposes the audio hardware’s built-in DSP to the audio application, letting you control the hardware’s mixing and DSP capabilities without the need for an external control application.

It remains to be seen whether AudioX will become a viable standard in the audio hardware world. The current implementation can best be described as “clunky but promising.” Will competing manufacturers jump on a common bandwagon to make their products easier to use? I sure hope so. The ability to use the recording host software to automate and control the hardware’s mixing, routing and DSP is a potent capability that all products would benefit from. Imagine – an AudioX-compliant digital mixer….

MIDI Plugins Cakewalk continues to add to its list of MIDI plugins. Cakewalk’s MIDI plugins are nondestructive tools that let the user process MIDI tracks by just dropping the plug into a track’s effects bin. Version 9 includes Arpeggiator, Chord Analyzer, Echo / Delay, MIDI Filter, Quantize, Transpose, and Velocity plugins. Also included is the NTONYX Style Enhancer plugin. Style Enhancer (reviewed in our January 1999 issue) is a cool tool for adding a human feel to MIDI tracks.

Session Drummer Plugin

New in this release is the Session Drummer plugin. Just drop this nifty gadget into the effect bus of a MIDI track, point the MIDI channel to a drum box, and go to town. By arranging any number of patterns to fit your song (the plugin includes hundreds of preset patterns), you can quickly create a drum backing for your composition. When you’re done with your composition, you can render the Session Drummer’s output as a MIDI track for detailed editing. Session Drummer is a convenient way to hack out drum grooves that completely avoids the hassles of either syncing and programming a drum module or recording MIDI drum patterns manually.

Global M/S/R Mute / Solo / Record buttons now appear in any window where they might be applicable – for example in the Audio editing views or the MIDI Piano Roll view. V8 users will again rejoice – no more changing views just to solo a track.

Scrubbing Audio scrubbing now works similar to an analog tape – just drag the scrub tool over the audio you wish to scrub, and the audio plays back at the rate you “scrub” it. This is a mixed blessing – some users will undoubtedly miss Pro Audio’s “looping” scrub feature that looped the portion of audio the user scrubbed. I for one don’t use scrubbing, since the visual display pinpoints clicks, blips, and the beginnings and endings of takes better than any scrub tool.

Guitar Features Guitarists will appreciate two new views which display MIDI data as guitar tablature or even as finger positions on a fretboard. Also, a chromatic tuner has been added to simplify tuning. These features were lifted from the sister product, Guitar Studio, and added to the PA9 feature set.

Staff, Tablature, and Fretboard Views

Some people would argue that a tuner is a useless feature in a “pro” recording platform. I disagree. First of all, no two tuners are exactly the same. If three musicians all use different tuners, chances are they are not exactly in tune, and sometimes the difference is enough to be a problem. So a studio ought to keep a “master” tuner available at all times. Secondly, with the tuner built into the recorder’s UI there’s no plugging and unplugging – just play a note and tune it up, no matter how the signal’s getting into the board. You can stop in the middle of a take, tune, and start a new take right away without unplugging anything.

Unfortunately the tuner does not work with bit depths and sampling rates in excess of 16/44.1, and wouldn’t work at all with my 24-bit test hardware, a Mark of the Unicorn 1224 24-bit I/O system. More thought should have been taken to ensure that the tuner would work correctly with all of the supported hardware. This is what I meant earlier by stability – does the software work with my hardware? Fortunately, this was an isolated problem and sure to be corrected in future versions. At any rate it is far from being a crippling handicap.

Performance Meters Pro Audio now offers performance meters to monitor disk and CPU activity, as well as a “dropout” light that indicates when playback has stopped due to system bottlenecking. I found these to be somewhat misleading and rather unnecessary – sort of like a warning light on your car’s dashboard that lets you know that you’ve just hit a tree. In most cases my system will stop well before either meter reaches 100% – and sometimes, the meters will reach 100% without a dropout!

If you’ve recently built or upgraded your system, and you’re trying to tweak out the performance bugs, it can be useful to know whether disk or CPU activity caused your system to stall. In that case the meters are useful. But, on a stable system, I usually know why my system stalled – in my case usually because I tried to run more DirectX effects than my CPU can process. According to Cakewalk, the meters consume almost no CPU cycles, so I guess if they’re free, and out of the way, then I don’t care if they’re there or not.

Console View

Console View Finally, the Console view now sports a slick new 3-D brushed-metal-and-smoked-glass look. Very sexy. When musicians record in my studio, seeing the Pro Audio console is always a “gee whiz” event for them. I certainly don’t place such improvements high on the priority list, but they do count. Especially when trying to sell projects.

If new features were the only change in Pro Audio 9, then I’d recommend that you save your money. Although the new features are good, none of them are “gotta have it” functionality that will revolutionize the way you record. As I’ve indicated, however, it’s the under-the-hood changes that make V9 so appealing to me.


The one improvement that just leaps right out and demands to be noticed is system latency. In case you’ve never heard the term before, latency refers to the time delay from the moment you make a change onscreen until the time that change can be heard in playback. Versions 7 and 8 of Pro Audio suffered from severe latency problems – on my system, running Pro Audio V8, latencies of up to two seconds were not unusual when playing a complex mix. That means I hit a track’s solo button and – “one one thousand, two one thousand” – now the track is soloing. A pain in the butt when mixing. And forget about trying to sweep an EQ to find that annoying snare drum ring – with a two-second delay between your EQ’s frequency knob and the sound coming out of the speakers, it’s almost impossible.

For years now we have heard that standard Windows multimedia drivers cannot achieve low latencies. This “problem” has been the major selling point of products that use non-Windows drivers, like ASIO.

So you can imagine my reaction when I discovered that Pro Audio V9 will run with very low latencies – in many cases below .1 seconds! This radical improvement in performance comes from a technology that Cakewalk has named “WavePipe.” WavePipe simply refers to the way Pro Audio communicates with the standard Windows multimedia drivers, and isn’t some kind of creepy proprietary code. Call it “good programming.” The simple fact is that Pro Audio achieves world-class latencies and throughput with the standard Windows drivers. No ASIO needed. In fact, with even the largest, most complex mixes, I was able to achieve latencies under .2 (that’s two-tenths) second.

Now, I have not benchmarked Pro Audio V9 against software that uses ASIO drivers, so it would be unfair to say that Pro Audio V9 has latencies that are lower than – or as good as – an ASIO-compliant program like Cubase. The simple fact is: I don’t know, and I don’t care. For mixing applications, .1-second latency is plenty good. It’s doubtful that you’d notice if it was faster.

How good is .1 second latency? To find out, just take a stopwatch, and try to start and stop the watch as fast as you can. Chances are, if you’re really fast, you’ll be able to start and stop it in .1 second.

The user’s experience is that changes take place “immediately” – hit solo, and the track immediately solos. Quickly and easily sweep an EQ to find trouble spots. Move an onscreen fader and the audio immediately follows your move. This is good stuff.

The other benefit that comes from improved latency is that the entire program becomes more responsive. Meters and faders move more quickly and fluidly, allowing easy setting of input levels and compressor thresholds. The time counters scroll smoothly. Buttons react responsively. The thing just feels fast.

I would be complaining loudly if this improvement in latency came only at the expense of track count. However, all of my V8 files will play well in V9 without tweaking. Some of them require a slightly higher latency setting than the default – up to .2 second for a really maxed-out mix. Most of my V8 work plays great with latencies of .1 second.

I doubt you’ll see significantly higher track counts or effects counts with V9 over V8. I don’t. If you do, chances are that your V8 install was hosed or wasn’t properly tweaked. But what you will find is equivalent track and effects counts with extreme sub-second latencies.

What kind of performance can you get from V9? From the standpoint of sheer disk throughput, a lot. On my system, a 450-MHz Celeron with inexpensive, 5400 RPM Western Digital EIDE drives, I can get up to 64 tracks of 16-bit / 44.1 KHz audio, or 48 tracks of 24-bit / 44.1 KHz audio – easily enough to handle the most demanding recording applications.

With the advent of cheap, powerful CPUs and 24-bit submixing, I tend to use fewer disk tracks and a lot more DirectX effects. A typical mix for me in V9 used 16 tracks of 24-bit audio, with 10 EQs, a TrueVerb, a Hyperverb, 4 C1 compressor / gates, 5 Renaissance Compressors, an FX2 Tape Simulator, and a De-Esser. All running in real time with under .2 second latency.

Maybe I impress easily, but I think that’s pretty cool.


More good news. After hundreds of hours bashing away during the heaviest of recording and mixing loads, I have not had Pro Audio crash or misbehave. Not once. This is easily the most stable version of Pro Audio since V6 – and maybe the most stable version ever. Its extreme stability makes it the most stable hard disk system I’ve ever used, including all of the major packages out there, Mac or PC.

There’s not much to say about a program that works well. You can’t put a sticker on the box proclaiming, “Now Crashes Less!” or “Even Less Frustrating!” or “It Actually Works!” So you can’t promote it when you get it right. I’ve worked in a software development environment before. It can be quite hard to justify an intensive bug-fix effort, especially when the “other guys” have more “feature bullets” on their box. So my hat is off to the Cakewalk team who had the guts to take the time and chase the bugs out of this software, rather than adding more gadgets and widgets. I hope the trend continues and I encourage you to let Cakewalk know you care about stability by “voting with your dollar.”

Now, the caveat. Whenever we test software or hardware, and claim something is reliable, the inevitable user shows up who can’t get the damned thing to work at all, and they blame us for recommending the software. I can’t guarantee that the software is going to work on your system at all. And no software is crashproof.

What I can say is that Pro Audio 9 offers the best performance and stability right out of the box of any previous Pro Audio release and is at least comparable in stability to any other audio recording software on the market at any price. I can also say that the experience of the V9 users I’ve talked to is unanimous on this issue.

As for me, I’m just tickled to death that I can run this thing for days – days – without a crash or other failure.


Many of you know of an ongoing compatibility issue between Gadget Labs Wave/8*24 (reviewed here on ProRec) and Cakewalk Pro Audio V8. Although most users have experienced no problems whatsoever, some Wave/8*24 users (myself included) experienced a chronic stuttering when working in Pro Audio V8, especially when running a large number of DirectX effects.

Cakewalk and Gadget Labs worked together on this release of Pro Audio to solve the problem. With the release of Pro Audio 9, and new drivers from Gadget Labs, I have heard positive results from some Wave/8*24 users. Comments from these users have ranged from “a definite improvement” to “problem solved.” So I was eager to put V9 – and the new 8*24 drivers – to the test.

Bad news for me. With the latest versions of the Wave/8*24 drivers (version 3.00) and Pro Audio 9, I was easily able to reproduce the stuttering problem on my machine.

Since other users have reported that the new software has helped or even completely solved their problem, I strongly urge current Cakewalk / Gadget Labs customers to try the new driver and V9 to find out if the combination works for them.


Here I have a mixed report. One of the reasons that performance and stability are so strong is because the developers at Cakewalk have decided to forgo some usability improvements that would have necessitated more sweeping changes in the software. The more you change, the more bugs you create. So no major changes were made to the way the user interacts with the program, but many small fixes were put into place that make the existing tools work more consistently and logically.


On the plus side, many little fixes and improvements have been made to the Console view. The Console is supposed to remember its initial settings – to set it, just rewind to t=0 and make your changes. This implementation was hosed in previous releases, and you actually had to take snapshots of the settings to get the knobs and faders to stay put.

I am pleased to report that the Console now works according to its design. And, recording any fader moves is a snap: just hit the Record button on the Console and any changes you make will be remembered and automated. In the past I have been immensely frustrated with the Console’s many glitches and inconsistencies. Now, it seems to do what it’s supposed to do. I would, however, like to have an option where the Console is totally non-automated, and always remains “hardwired” to the Track View, instead of having to be rewound to the beginning of the song.

Of course, if you know anything about me, it’s that I hate mixing with onscreen mixers. It’s like sitting at 72-channel SSL with your hands tied, mixing with your nose. I have been complaining – to no avail – to eliminate onscreen mixers from all recording and mixing software. Some companies listened. Cakewalk didn’t.

The sad fact is that the Console remains a necessary part of doing business with Pro Audio. It is the only place in the program where you can make adjustments to Aux sends, Aux returns, the Master outputs, and the post-insert (master) track levels (the pre-insert levels can be controlled with the clip envelopes). The only usability change versus previous versions is the new ability to add DirectX plugins from the Track view. A good change, but not enough to let me put away the Console and be done with it. Too bad.

The good news is: at least now it’s usable. With the improvement in latency, and the improvements in usability, the V9 Console is as close to a usable onscreen mixer as you’re going to find. I’d personally rather not use it at all, but at least now I don’t have to curse it when it misbehaves.

And in all fairness, no software yet has mastered the “mixer-free” paradigm. Some, like Pro Audio, offer envelopes to control most, but not all mixer functions. Others, like Sonic Foundry’s Vegas Pro, eliminated the mixer but failed to provide all the necessary mixing elements (just try to do a razor-cut on a reverb swell in Vegas – or even a master fade – and you’ll see what I mean).


The enhancements to the mixdown functionality were a welcome change over V8. In the past, the “mixdown” command always mixed the audio to a new pair of tracks in the current song. So getting a final two-track mix out of Pro Audio always involved two steps: first, mixdown to two tracks; then, export the two-track mix as a WAV file. It’s as if nobody at Cakewalk ever thought that far ahead. And in V8, if you were submixing (bouncing) tracks, the resulting submixed audio was always the full length of the song – even if the part you wanted was only three seconds long. A real time-killer.

Now, the mixdown command lets you directly mix to a WAV, Real Audio G2, or MP3 audio file. And you can now select a time region and only mix the audio between the FROM and TO markers. These improvements resolve many old issues with PA8. I found the ability to select and export a specific region invaluable when extracting loops from old mixes for a loop collection I’m building: just solo a track, select a measure or two, and hit export. Instant loop, no waiting.

But there’s still some room for improvement. Currently, a mixdown bug causes 64 extra samples of silence to be added to the beginning of the export – no biggie if you’re just doing a mixdown, but a real pain if you’re exporting loops that have to be dead-on. This known bug is expected to be fixed in an early maintenance release.

And no matter how long I use this thing, each and every time I mix a song down, I inevitably wind up with just a little sound bite because I have left a time region selected. So I have to deselect everything by clicking in the “space” outside the audio clips, and mix again. Very counterintuitive. And no option has been made available for exporting at a sample rate or bit depth other than the project’s default settings. This means that projects that use 24-bit audio will always export 24-bit audio. You will have to convert the audio in an editor to get it ready for CD.

User Interface

How many times have you gotten ready to cut some lumber, or eat a meal, or work on your car, and immediately reached for your Swiss Army Knife?

Well, why not? After all it does include a saw – and a knife, fork, and spoon. And a nifty little set of pliers.

It’s hard to be the best at everything. Even if you could be “the best of everything”, it’s even harder to wrap “the best of everything” in a single, convenient package.

On the one hand, Pro Audio’s dockable toolbars and clean look represent a great user interface. Wherever possible, Cakewalk has opted for clean, usable controls, rather than the glitzy, shadowy, overly sexy user interfaces of some other software packages. The controls are practical: the right controls with the right size and shape; with clear, simple icons.

On the other hand, this is a legacy application which suffers the problem of all legacy applications: scopecreep.

The feature list has grown to behemoth proportions, including a lot of baggage I’d like to see stripped out. A programming language (CAL). An object-oriented controller design environment (StudioWare). A live performance arranger (Virtual Jukebox). Even the simplest project frequently requires three or four windows to be open simultaneously: track view, audio view, console view, and an audio plugin. It is not unusual to have more windows open.

The problem with features is they are easy to introduce, and impossible to remove. Maybe only 2% of Cakewalk users can write a CAL script. But take it away, and those 2% will become a very vocal minority! Without careful pruning the program becomes a jumbled collection of features, rather than a steamlined problem-solving tool.

Case in point: the “+3 dB” and “-3 dB” audio editing tools – relics of the earliest versions of Pro Audio. Destructive editing commands that have no place at all in a tool with envelopes that allow precise, non-destructive volume control. Why are these commands still present in the package?

Great improvements in usability could be made – but only by sacrificing some old luggage and some tired paradigms. For example, the “one task – one window” paradigm has got to go. You start with the Track view. Mixing? Add the Console view. Doing MIDI? You’ll need to open the Piano Roll. Editing Audio? You’ll need to open some Audio Views. Plug-ins? There’s a window each. And we haven’t even started to talk about the Tempo view, Event List, Lyrics view, Staff view, StudioWare panels, Video window…. I’m getting tired just listing them. If you want to make full use of all that Pro Audio has to offer, get ready to open and close some windows.

There’s an easier way. I’m talking about panes, and some easily configurable, customizable panes would go a long way towards simplifying the user’s job. Panes – and other interface options like tabs – allow the interface to contain a greater variety of functionality and still feel cohesive. Currently the Cakewalk UI is based around functions. Redesigning the user interface around the process is going to be a difficult, painful process, but sooner or later (hopefully sooner) it will have to be tackled.

Surely there’s a way to integrate the functionality of the Track view, the Piano Roll, the Tempo view, the Console view, and the Audio view so that one window with context-sensitive panes can replace stacks of windows. Some users do mainly audio. Other users do mainly MIDI. Solution: offer some preset configurations that place the most common items in one window at the same time, and make use of a context-sensitive tool palette that presents the right tools depending on the task.

To their credit, the folks at Cakewalk have done a tremendously impressive job of packaging this vast feature set into an intuitive and easy-to-use program. You can productively live your whole life in Pro Audio 9 and never even know about some of its more innovative features. But, of course, therein lies the problem.


The new design of Pro Audio V9 makes it easier to optimize playback on your hardware. On my system, the “out of the box” settings worked pretty well, with only a few modifications needed to really max out performance.

The buffering and queuing options have been simplified and split into two areas in the “Options – Audio” dialog box: “Mixing Latency” (found on the General tab) and “File System” (found on the Advanced tab).

On my system, I found the default “Mixing Latency” settings to be dead-on: 4 playback buffers, and a Latency setting of 92 ms. Increasing the Latency slider to higher settings like 300 ms usually did not result in a greater track count, and decreasing it to values like 23 ms resulted in poor track count. I found the 92 ms setting to be offer good track count and still offer “instantaneous” latency.

For “File System” the default settings of “No Read Caching” and “No Write Caching” work best on my computer. However the default “I/O buffer size” of 64K was too small. The best overall performance on my system was at 256K. I am using high-speed EIDE drives formatted with large block sizes (for more information on file system optimization, please read Jose Catena’s seminal article here on ProRec). Try different values for I/O buffer size on your computer: 64K, 128K, 256K, 512K – it makes a real difference. At settings that are too low, you’ll see a decrease in track count. At settings that are too high, you’ll find that track count doesn’t increase, and latency starts going through the roof.

Cakewalk has updated its Wave Profiler to include the best settings for most of the current pro audio soundcards available. I found that Wave Profiler correctly profiled both the MOTU 1224 and the Gadget Labs Wave8/*24 I tested the system with. Using the 1224 with an Opcode MIDI Translator PC, I had dead-on MIDI / Audio sync right out of the box. Great stuff.


Cakewalk Pro Audio Version 9 is the best release yet in Cakewalk’s Pro Audio software line. It is the first version of the product to truly live up to its “Pro” moniker.

Version 9 is super-stable, high-performing, full of cool features, and well-suited for musical tasks from composition to commercial scoring and everywhere in between. At an upgrade price of only $79, Version 9 is a no-brainer for any current Pro Audio user. And at a street price of about $300, Pro Audio packs a serious musical punch at a More-for-Less price.

I’ve heard it said that the “pro-sumer” recording products (including Cakewalk Pro Audio) are too “amateurish” to get the job done. People argue that “Product X has the Expandalizer” or “Product Y doesn’t have Frequencification.” Witness the rush of the “pro-sumer” companies to include 96 KHz sampling rates in their software products a full year before the supporting hardware was even widely available. Hype, hype, hype. As if any of these gee-whiz technologies are what’s really important when you’re making music.

The fact is that a decent producer / engineer can make a hit record with just a fraction of the technology from any “pro-sumer” recording package. On a Mac, a PC, or – *gasp* – without a computer at all (I can hear you now, thinking “is that really possible?”).

What makes a product “pro” is almost never the gee-whiz feature set. It’s performance, stability, and usability. If a product has good performance, and if it reliably and conveniently solves most of your problems, then it’s “pro”. Even if it doesn’t support 64-bit, 192 KHz audio or whatever the market starts telling us we need next.

Don’t be fooled by the naysayers. If you can’t produce the music you want with Pro Audio V9, chances are that no amount of music technology will solve your problem.