What happens when you mix technical chatter on the Cakewalk forum, Samuel Beckett, and The Matrix? I’d wager you get something like the surreal video above. Prompted by the posting of technical details for a new update to Cakewalk’s SONAR production software for Windows, and empowered by a strange, new tool that generates eerie virtual reality from typed text, we get banter like this:

The arpeggiator is now on every track, so you are supposed to use it. It is one of the new rules of recording.

Yes, I came from the days of one-finger piano playing. This is a total blessing to me.

I’m going to take that as a challenge and base my review of SONAR 8.5 on using an arpeggiator and step sequencer on every track. And I’ll have to pronounce all those hard g’s in the voice over, clearly.

And no, this is not some twisted viral campaign on the part of the folks of Cakewalk; I’ve been assured that this came from a user.

Okay, what was this post originally about? Oh, yeah – the actual technical details of the SONAR 8.5 release. Noel Borthwick talks about all the details of the new SONAR release on the Cakewalk forums. Apparently, some people care deeply about whether this is SONAR 9 or 8.5 or some conspiracy theory there, but what interests me is the technical details of the software itself.

SONAR 8.5 Fine Print

Noel goes down to a code level. Interesting tidbits: working with Intel, Cakewalk was able to do a demo of SONAR running an absurd number of tracks, instruments, effects, and live video without pegging the CPU, with a tiny 2 ms of latency. The Cakewalk engineering effort also has put together what may be the most highly-optimized VST support and richest 32-to-64-bit bridging on any platform, anywhere.

Whatever the opposite of “marketing speak” may be, I think that’s what Noel has achieved, getting into a sort of developer-to-developer level discussion. It is still readable, and worth digging through.

See also: Intel Developer Forum details and video on the Cakewalk blog

I could talk more about that, but let’s just leave it at step sequencers and arpeggiators on every track, okay?

Windows matters. It’s what roughly half of CDM readers use, and – for all the attention Apple gets – it’s a big part of the computer music world. Windows today also faces many of the same under-the-hood challenges that other operating systems do, so even if you’re a die-hard Linux or Mac user, you may want to pay attention.  You don’t need to love Windows, and you certainly won’t be hosting a Windows 7 launch party. You want to know if the OS will get out of your way and let you get to work.

Windows Vista proved what happens when an operating system’s many interconnected pieces are out of alignment. Even a graphics driver out of sync with underlying changes in the OS could render audio unusable, because just one missed sample can produce an audible glitch or dropout. Part of why I’m optimistic about Windows 7 is that Vista today is a radically different picture, thanks to many, many fixes delivered by Microsoft in updates and more mature audio and video drivers. But that means not just whether 7 is better than XP, but whether 7 is also better than Vista.

Vista wasn’t entirely alone: Mac and Linux have all had their share of growing pains in recent years. The devil is usually in the details. So, I again turn to one of the best guys in the business for sorting out all those technical details. Noel Borthwick, the CTO for Cakewalk, probably has a better big-picture view of how music and audio work in Windows than anyone on the planet. He’s a person hardware and software vendors outside Cakewalk often rely upon as a resource. Noel kept us technically honest on Vista, and he’s doing it again on Windows 7, with some exclusive information for CDM.

Those details get mighty technical, so here’s the punchline: Windows 7 is an OS Noel would use himself. It was hard to get anyone to recommend Vista over XP; loyal Windows-using developers I know still largely stick to XP. But would Noel switch from XP to 7?

Yes, absolutely. Windows 7 finally delivers on the stability and performance that users hoped for from Vista. The kernel changes and optimizations for large scale multi-core processors make it very attractive to DAW users who are interested in better low latency performance. I will be building a new DAW soon and Windows 7 X64 will be my OS of choice.

What’s new in Windows 7?

• Better multithreading: Improved performance of highly-multithreaded software and hardware by removing a significant bottleneck, especially relevant to a tool like SONAR
• Better memory management: Improved memory management when working with multiple threads
• Less nagging: More customization over UAC prompts (meaning they don’t have to nag you more than you want)
• More lightweight: Fewer system services run by default on a stock system, plus a leaner footprint of the OS
• Media support: More native media format support, including QuickTime MOV and H.264, plus drag-and-drop media transcoding
• Composite devices: More logical display of hardware with multiple functions (like audio and MIDI).
• FireWire: Enhanced FireWire support, with IEEE 1394b
• Multi-touch: Multi-touch display support
• Usability improvements: An improved user interface, task bar, and Libraries for managing files

If you’re ready for all the gory details, read on – including a frank appraisal of how all of this compares to XP in real-world performance, and what compatibility issues to look out for if upgrading from either Vista or XP.

Noel Borthwick of Cakewalk effectively wrote this story in response to my questions, so these answers all come from him. Microsoft has not responded to my requests for a review copy, so I’ll be able to evaluate this on my own system – albeit far less scientifically than Noel can – closer to launch.

WARNING: Extremely geeky details of the inner workings of Windows 7 follow, in keeping with our “never dumbed down” policy. If you’re a developer, you can likely get some leads on how to better support Windows 7 in a single point, something even Microsoft doesn’t provide as completely. But if you’re willing to dig, you get a rare view of the OS from a developer view – no marketing speak, no cheerleading, no fanboyism, no platform wars, no writing for the lowest common denominator.

What Actually Improves Audio Performance

Peter: In terms of performance for audio production, what are the significant differences in Windows 7?

Noel: Windows 7 on the surface is very similar to Windows Vista. It has the same audio driver support and same audio system infrastructure as Vista. However, it’s some of the under-the-hood improvements that are more significant for audio production. There are some interesting innovations and optimizations in the Windows kernel, making the OS more scalable for concurrent processing. This makes it attractive for highly multithreaded applications like SONAR. Additionally there are various new API’s/SDK’s that may be of significance to developers. Some highlights are below:

Multi-threading: Removal of the kernel “global dispatcher lock”

In Vista and earlier, on a highly multi-threaded system (e.g. SONAR running on an 8 core hyper-threaded Intel Core I7 PC), you have many threads all processing tiny audio buffers at low latency. All these threads are ultimately waiting on the dispatcher lock when it comes time for them to be managed by the Windows scheduler. This global lock becomes a bottleneck in the system and prevents efficient multi-core workload distribution and scalability. This problem gets magnified as you increase the number of cores since they are all gated by a common lock. In Win 7 the kernel team changed the logic in the Windows scheduler to abolish this global dispatcher lock and use per object locks. This effectively removes this age old bottleneck and allows Win 7 to scale better even under workloads of 256 processors.

This change means a lot to applications like SONAR that rely on multithreaded processing of very small workloads. Initial benchmark results have been promising in this regard. SONAR performs more efficiently at low latency on multi core machines.

Improved Memory Management – PFN database lock

The PFN (page frame number) database lock was used by the memory manager to lock pages of memory in the working set. Like the dispatcher lock above, this would gate memory access from different threads causing resource contention. Work in this was first done in Windows server 2003 SP1 and Windows 7 has now has this optimization as well, improving asynchronous access to memory.

Power Optimization: Core Parking

Windows 7 has a new feature called Core Parking. Core Parking is a power saving optimization that shifts processing load to one or more cores and puts other less busy cores to “sleep”. The objective is to let other cores idle if workload levels allow for it. This optimization had us scratching our heads when we ran a benchmark test on a Quad Core I7 machine. At any point in time, we would notice that some cores were idle in task manager. The reason for this turned out to be Core Parking. Core parking can be useful to save battery life while running projects on laptops.

Better WaveRT Performance

Unlike Windows Vista, Win7 now uses event mode internally. This is good news, since it will help guarantee that HDAudio drivers in Win7 support WaveRT event mode properly. Additionally event mode is now part of WHQL logo certification for driver vendors, so any WAVERT device must support this to get a Win7 compatibility logo.

Ed. note: The plain-English translation here is that WaveRT, Microsoft’s own real-time audio driver facility, now is more likely to work the way you expect. Cockos, makers of REAPER, actually provided the ability to turn off WaveRT Event Mode at the end of last year because of unpredictable results. Windows 7 should resolve these issues.

Other Improvements

Peter: Noel also assembled some other improvements worth noting in Windows 7. They’re subtle, but useful: you may finally be able to avoid installing QuickTime/iTunes just to play some video files, interfaces with audio and MIDI jacks don’t have to show up separately any more, there’s improved FireWire support, usability improvements, and multi-touch on mainstream computers is now nearly here.

Noel:

Additional File Format support

Windows 7 adds native playback support for media in MP4, MOV, 3GP, AVCHD, ADTS, M4A, and WTV multimedia containers. It has native codec’s for H.264, MPEG4-SP, ASP/DivX/Xvid, MJPEG, DV, AAC-LC, LPCM and AAC-HE

Yes you read that right – QuickTime MOV file support is now natively available in Windows 7 so you don’t need to install QuickTime. Another big plus is that this is supported under the X64 version of Windows 7 as well, something you cannot do with Apple’s native QuickTime itself!

All media files using these codec’s should play in Media Player. It appears that these new codec’s are exclusively available to Media Foundation applications and not via other legacy API’s such as DirectShow etc.

File format transcoding

File format transcoding of many popular formats is now built into the Windows 7 shell. I.e. dragging and dropping files onto a device automatically performs the necessary format transcoding if the format is supported. This was primarily done to copy formats to portable devices like cameras but should be useful in other scenarios as well.

Multi-function devices and Device Containers:

Prior to Windows 7, every device attached to the system was treated as a single functional “end-point”. While appropriate for single-function devices (such as an audio interface), this does elegantly represent multi-function devices such as a combination audio/MIDI interface. In Windows 7, the drivers and status information for multi-function device can be grouped together as a single "Device Container", which is then presented to the user in the new "Devices and Printers" Control Panel as a single unit.

http://www.microsoft.com/whdc/Device/DeviceExperience/ContainerIDs.mspx

Note: this should not be confused with device aggregation as is available with Core Audio on Mac OS. On the Mac, you can treat multiple audio interfaces as though they’re one interface, so, for instance, you could get extra outputs by combining a couple of audio interfaces, and your software will see them as if they’re just one box. But SONAR provides this capability on its own, so if you’re a SONAR user, you can get the same functionality.

FireWire/USB

Windows 7 contains a new FireWire (IEEE 1394) stack that fully supports IEEE 1394b with S800, S1600 and S3200 data rates. According to reports, USB 3.0 may be supported in a future Windows Update. It was initially planned for Win7 but is not supported in the shipping version of Win7 due to delays in the USB 3 specification.

Multi-touch

Windows 7 includes integrated support for multi-touch displays.

Libraries

Libraries are user-defined collections of content including folders. It’s a handy way to categorize and create shortcuts to samples, music, etc. Special shell folders (Documents, Pictures, Music, and so on) are now Libraries.

Accelerators for Windows

Windows 7 Accelerators provide a way for learning more about selected text, optionally using voice control.

Virtual hard disks

The Enterprise and Ultimate editions of Windows 7 incorporate support for the Virtual Hard Disk (VHD) file format. VHD files can be mounted as drives, created, and booted from.
An installation of Windows 7 can be booted and run from a VHD drive, even on non-virtual hardware, thereby providing a new way to multi boot Windows.

Leaner Footprint

Win7 has a leaner footprint and has been tweaked to work well on less powerful PC’s, laptops and Netbooks. I have heard reports of Win7 working more smoothly on machines that would be slow under Vista.

Listen Mode

Another nice touch in Win 7 is that they now have a listen tab in the audio properties. Turning on "listen mode" basically routes input to the default output device allowing you to monitor an input device in Windows itself. Sadly this runs via the Windows audio engine which is always running in WASAPI shared mode, so it’s subject to a 30 msec delay. Of course you can always load an application like SONAR and route the audio inputs to an output for low latency monitoring.

Compatibility: What to Watch

Upgrading from Vista

Peter: Relative to Vista, are there any changes that are likely to introduce new compatibility issues with hardware or software?

Noel: With any new OS there is always the potential for compatibility issues. Win7 is built on the Vista foundation and one of its goals was better compatibility. As such most applications that are Vista compliant should work as well or better in Windows 7. UAC in Windows 7 has been improved so this might also help with general compatibility problems with some applications.

We have run into only a couple of compatibility issues in Win7 during the course of our development/testing of SONAR 8.5.

The MMIO API in Win7 (typically used for writing RIFF wave files) has a compatibility issue with the mmioDescend API with LIST ‘WAVE’ chunks. This caused our code that reads audio bundle files to fail and read scrambled audio data. We worked around this problem in 8.5

In WASAPI exclusive mode under Win7, the minimum latency you can achieve is now unfortunately 3ms and the code reports an error if lower. The fact that Vista has no such limitation has been reported to Microsoft. Hopefully its a mistaken fence in their code and this issue is fixed via an update, since it’s a step backwards for low latency in WASAPI mode.

Ed.: That last issue is an interesting one for anyone really pushing the envelope with low latency, so I’ll keep in touch with Noel if there’s any update.

Upgrading from XP

Peter: What hardware and software compatibility issues should users be aware of if they’re thinking of migrating not from Vista but from XP to Windows 7?

Noel: The compatibility issues that typically affect users migrating from XP to Vista/Win7 are:

UAC problems: Many applications and plug-ins are not built to handle the newer security settings in these OS’s. For example, if an application relies on something that requires administrative access it will fail when running as a limited user in Win7. This is a serious issue since in Vista/Win7 even if you are running from an administrator account; programs are launched by default with limited user privileges. Unlike XP, you have to explicitly run as an administrator to use such programs. To be Win7 logo-compatible, all applications need to should support running as a limited user.

Drivers: Although for most practical purposes audio drivers in XP and Windows 7/Vista are similar (you still need to write WDM drivers) there are sometimes quirks in specific drivers may cause problems. Most typical driver issues here are caused by installers that make assumptions about the OS version. In many cases this issue can be solved by the end user by setting the “compatibility mode” to Vista in the file properties for the appropriate driver installer file. (Right click the setup exe file to set its properties)

Ed.: I don’t feel either of these is a deal-killer, as I’ve been living with Vista for some time, but they’re still worth watching out for if upgrading from XP. And it means if you have an older machine that’s still working properly, you’re just likely to leave it on XP and worry about sorting the upgrade on a new box.

Less Nagging?

Peter: We talked when Vista came out about User Account Control and particularly audio-specific tasks that required elevation or different handling of permissions in Vista. I know UAC has been streamlined in W7. Do these changes impact audio apps at all? Are there corresponding under-the-hood changes?

Noel: The UAC changes in Win7 are primarily to allow more customization over the UAC elevation prompting process. There are no changes to the fundamentals of how UAC itself works that I am aware of. The classic problem with audio applications with UAC is when programs or plug-ins write to areas of the registry or file system prohibited from standard user access. Even when you are running as an administrator, by default when you launch a program (or the program itself launches a secondary process) Windows 7 will run that process with standard user privileges. If a program or plug-in attempts to write to an area which it doesn’t have write privileges for, virtualization will kick in. While this may allow the program to work, in general it is bad practice to rely on virtualization, since it can cause many unwanted side effects and behaviors in applications.

There are now four customization settings for UAC:

1. Never notify (least secure). The user is not notified when a program tries to install software or make changes to the computer. The user is not notified when they make changes to Windows settings or when programs try to do so.

2. Only notify me when programs try to make changes to my computer. The user is not notified when a program tries to install software or make changes to the computer. The user is not notified when they make changes to Windows settings. However, the user is notified when programs try to make changes to the computer, including Windows settings.

3. Always notify me. The user is notified when a program tries to install software or make changes to the computer. The user is also notified when they make changes to Windows settings or when programs try to do so.

4. Always notify me and wait for my response (most secure). The user is notified when a program tries to install software or make changes to the computer. The user is also notified when they make changes to Windows settings or when programs try to do so.

Customization and Tuning Advice

Peter: How much customization would you advise people do to their OS? That is, you’ve just installed a build of Windows 7 for working with SONAR on a test machine. Do you run the stock configuration, or start turning off services, disabling disk indexing, etc.?

Noel: Optimization and customization is a topic that can’t be fully discussed in the scope of a brief article. In general you need to optimize a system when you have known bottlenecks. Otherwise you can spend a lot of time tweaking things that have little effect on the end goal. In fact, you may even end up destabilizing a perfectly working system. A stock Win7 machine is not optimized for audio necessarily but it appears MS put some thought into trimming out unwanted startup tasks to cut down on startup time. For example there are now “Triggered start services” in Windows 7, so out of the box you can have fewer services running after a fresh boot. There are probably many background services in a modern DAW that could be suspended if you don’t need them but they should be evaluated on a case by case basis depending on what you use the machine for.

Peter: A lot of users were advising running Vista with Aero off, certainly in the early days. Do you think it’s now advisable to leave Desktop Window Compositing switched on for audio work? (Note: I am aware that there’s actually no way to *completely* disable the Aero windowing environment in a way that it reverts to XP, as even in Class mode with no compositing settings the engine has been altered.)

Generally speaking, turning off Aero will free up some resources on your system, since it uses more costly 3D graphics rendering and transparency a lot. However on any modern graphics card, Aero offloads a lot to the GPU so unless your DAW is also competing for the same GPU resources, turning it off may or may not make an appreciable difference to performance. Most applications that are not graphics intensive use GDI for rendering to the screen and since GDI doesn’t take advantage of DirectX hardware acceleration it’s normally not contesting with the GPU. If you are using plug-ins that use Direct 2D or Direct3D, you are probably better off disabling Aero.

Launch Party, After All?

Thanks, Noel. So, the big news behind all of this is that a move from XP to Windows 7 is finally advisable.

I would still caution, as I did recently with Mac OS Snow Leopard, that you typically don’t want to upgrade to a new OS the day it launches. You’ll want to verify compatibility with your software and hardware before making the jump.

That said, this is an unusual upgrade in that it appears to resolve more issues than it introduces. I actually haven’t been able to find a single user out there testing Windows 7 who has found any issues with audio or music production. Of course, when it launches, we’ll have a much larger test base, so I expect we’ll find something – even Windows Service Packs and point releases of Mac OS have been known to create some issues. As we get closer to launch, I’ll review how you would backup your existing XP or Vista system to ensure that if you do choose to upgrade, you can revert to a previous version.

I am, however, cautiously optimistic. And now is an especially good time to make the jump to 64-bit. It’s easier on Windows than any other OS at the moment, and easiest in SONAR, because SONAR allows you to easily migrate 32-bit plug-ins into the 64-bit environment. You’ll need a 64-bit machine and enough memory to make 64-bit worthwhile, but if you’re building a new workstation, as Noel is, the timing could be perfect.

I also think there’s plenty of room left to talk about issues that go between operating systems, particularly how audio software can better support multi-threading and processing on the GPU, multi-touch, as well as emerging I/O standards like USB3. (OpenCL, much-touted in Snow Leopard, is also supported on Linux and Windows, and Linux actually beat both Mac OS and Windows to the punch in providing a first implementation of USB3.) Correction: I should also add that the excellent Reaper has also added this feature. With full 64-bit support in Cakewalk’s own Dimension and other instruments, NI’s Kontakt sampler, and the bundled 64-bit-native plug-ins in Reaper and SONAR, that means you can build a really capable 64-bit rig on Windows.

With fixes getting the OS out of your way, we can return to issues that really matter, many of which apply to every OS.

Music is, as always, the perfect place to talk about these issues. We push our machines harder than just about anyone, and in ways that are the least tolerant of timing discrepencies and glitches. I’ve said it before, and I’ll say it again: if you want to look into the future of computing, ask a musician.

And that calls for a party.

Previous coverage:

SONAR 8.5 and how it can smooth the transition to 64-bit (8.5 is the build that includes Windows 7-specific improvements)

Vista Tweak: Use the Audio Profile Cakewalk’s CTO Uses

Optimizing for Vista: Inside the Mechanics of SONAR 8 with Cakewalk Engineering

Adieu, XP; How Vista SP1 is Doing, and Why This OS Generation Has Been So Tough

Vista for Audio, 1 Year Later: Talking OS Plumbing with Cakewalk’s CTO

Vista for Music + Pro Audio: Exclusive Under the Hood with Cakewalk’s CTO

And yes, I think Noel deserves an Honorary Contributing Editor position for all he’s done giving us absurdly-precise inside details for how Windows works.

Microsoft product screen shot(s) reprinted with permission from Microsoft Corporation.

Quick: you’ve got to sell UMPC (Ultra Mobile PC’s) to a mass market! How to do it? Well, Intel decided to show off pro audio and music production on the Linux-based Transmission, from Trinity Audio, as we saw earlier this week. I’m not entirely sure what got Intel thinking our geeky way, but I’m going to enjoy it while it lasts. And in all seriousness, Linux really an ideal OS choice here, because of its ability to be customized to the application.

The other flipside: low-power is the future. Computers now suck up 15% of the electricity in the US — electricity that produces a lot of our pollution and greenhouse gases. You do the math. A lot of that power gets used up in data centers, but the aggregate of all those homes counts, too. That will impact the future of all end-user operating systems.

Trinity has sent us some photos of the Intel booth at CES. Yes, Linux audio is getting some wider exposure. And even if you’re attached to Mac or Windows as your desktop/laptop platform, a mobile Linux device could be an ideal companion in the near future. We’ll have a chance to look at Trinity’s own device next week at NAMM and see how it stacks up.

MacBook Pro knows the way to Santa Rosa

One thing I was a little unclear on in my previous story is what matters in Santa Rosa, Intel’s latest architecture platform. (They didn’t call it Core 3 Duo, but then, consistent branding and Intel don’t generally go together.) As with Core 2 Duo over Core Duo, we’re getting incremental performance enhancements relative to the previous generation. Each step is relatively small, but they start to add up — hence, Apple quotes 50% gains over the original Core Duo. (And that’s why they dumped PowerPC, which in the mobile space was starting to practically paddle backwards.)

The key differences as far as raw performance: faster front-side bus (800MHz instead of 667), which for audio is a big deal, faster clock speeds on the models themselves at the same price, and fast RAM, plus a faster GPU for GPU-related tasks. (And, um, any day now we’ll start to see audio on the GPU — it’s tough to program, and GPUs are only now becoming the norm, and CPU cycles are getting cheaper, but it will happen.)

Also, none of this was meant to say “eBay your MacBook Pro.” PowerBook G4, maybe, but the first-gen MacBook Pro is still a terrific audio machine, with a GPU that’s no slouch. My main laptop right now is a first-gen MacBook (no Pro), and it blazes through everything I throw at it.

What is multi-core, anyway, and what does it matter for music creation?

For some time now, high-end PCs have come with the option of multiple processors. Traditional single-core computer architecture features a single CPU which does all of the work. In a DAW application, the computer needs to do lots of work in a very short amount of time. There are two ways to get more work out of your computer: use a faster CPU, or use more CPUs. Today’s multi-core processors offer the best of both worlds. Each core can do its own independent processing, but because they are on the same physical silicon chip, they can do it even faster.

Before we get too deep, I’d like to clear up a common misconception about multi-core processing (or even multiple processors). There is a common assumption out there that if you simply throw more cores at a task, the task gets done faster. Let’s remember that the reason multiple cores yield more power is that they can do independent processing. Also, applications (like DAWS) will only see a real benefit if they are built in such a way that important tasks can be done in parallel. That is to say, the DAW has been designed to intelligently balance the processing load across all available resources.

Fortunately for musicians, digital audio workstations (DAWs) are the perfect type of application to take advantage of the parallel processing provided by multi-core and multi-processor computers, BUT there is a significant amount of coding that is required to take advantage of it. At Cakewalk, we have put in the time and effort to take the maximum advantage of this technology.

If you think of the work a DAW must do, you can break it into two major areas: the UI (the windows, sliders, clips, etc.) and the audio engine (the low-level streaming code responsible for pushing all those bits through to your soundcard). If this was all there was to a DAW, the benefit of multi-core would be noticeable, but not profound. You would find that no matter how busy the audio engine got, your UI would still seem very responsive, because each core was doing one of those tasks independently without waiting for the other to finish. This is fine as far as it goes (and is as far as some DAWs go), but what we really want in a DAW is a system where the Audio Engine work itself can be divided up over multiple cores. In SONAR, we’ve taken the effort to build a truly multi-processing Audio Engine which takes full advantage of as many cores as your system provides. It is a truly scalable solution to handling bigger and more complex projects full of plug-ins and virtual instruments (in other words, “typicalâ€? projects).

Why walk when you can take the highway?
A simple analogy would be to imagine that a dual-core processor is like a four-lane highway—it can handle up to twice as many cars as its two-lane predecessor, without making each car drive twice as fast. In layman’s terms what multi-core processing means to you is “more.â€? In SONAR you will experience more tracks, more simultaneous effects, more virtual instruments than ever before. And you will also experience faster screen redraws of complex waveform activity and capabilities to work with higher quality digital video.

State of the art
Multi-core is just part of the equation; the news is good on all development fronts. Windows XP is a stable operating system for music production, and Windows x64 and the upcoming 64-bit version of Vista provide even more power for music production. Processor designs have advanced to the point where really anything is possible with the right application.

The dream of doing it all in one box is a reality; there has never been a better time to create digital music on a PC! And, with the mass deployment of multi-core processors, nearly everyone who makes music will see an immediate and significant performance boost in their favorite DAW—as long as their DAW has been designed to take advantage of parallelism, like SONAR.

Besides being the Dir, PR for Cakewalk, Steve Thomas has many years of real-world experience making music on both sides of the glass as a musician and audio engineer.

CDM Responds

Having tested SONAR 6 even just a little bit, I’ve in fact found that you can throw a whole lot of intensive audio processing tasks at it without blinking, and I’m using a comparatively low-end processor. My AMD 3800+ Athlon x2 CPU on the desktop PC I built is now trading for as little as $150 (if not for the 3800, certainly for a comparable CPU). I’ve gotten similarly nice results out of Ableton Live 6, which also features multi-core enhancements; to me SONAR and Live make a really nice Windows music-making combination. Notably, you will see similar performance gains on multiple processors; I’ve performed benchmarks on Logic 7 and Live 6 on both multiple cores and multiple processors (like the Core Duo and dual-core G5 Mac) and found similar gains from multithreading.

Steve, aside from being a marketing guy, is also a musician; he and I recently talked about the work he does in his studio as a guitarist; just a few hundred dollars for a Dell PC got him an attic studio that does everything he needs. So thanks, Steve, for some insight into SONAR and multi-core processing.

The first response came over the weekend from Adobe’s John Nack on his personal blog, waxing largely philosophical about why it made sense to support the newer Intel Macs instead of the PowerPC platform Apple themselves had abandoned. Now, I’ll be the first to concede Mac users can be hotheaded, but I think the better response would be to cut straight to the technical reasons why Adobe’s developers made this choice. Mac users assume, because they’ve been told so repeatedly by Apple, that creating universal applications is a “checkbox-clicking affair.” You can see a comment to that effect in the extensive discussion Mr. Nack triggered on his site.

Adobe audio product manager Hart Shafer chimes in today with the simpler technical answer:

Soundbooth and PowerPC Chips [Hart's Audition]

Basically, Soundbooth contains lots of Intel-specific code that would be inefficient to port to PowerPC, and the additional QA testing required for an additional CPU architecture was deemed an unworthy investment. (Note that the flipside of this argument would be that, as some Mac users had hoped, Apple’s switch to Intel makes development cheaper for software that’s heavily reliant on the processor, like audio apps.) Now, I’m in no position to evaluate that argument. Since a significant number of our readers are programmers, I’ll let you read his entry and tell us if you think this is a significant issue. What is interesting here is that Shafer never says Adobe’s can’t also port Audition to the Mac. I’m going to keep hoping this is possible; Peak can’t meet everyone’s needs, the excellent Spark is long gone, Apple discontinued their standalone editor product, and I think Mac users would welcome Audition with open arms. In the meantime, I’ll continue enjoy using Audition on my PC.

But, regardless of whether Adobe made the right call here (and it’s their choice to make), there is one conclusion that’s safe to draw here: cross-platform development isn’t always as easy as it might seem. Some newly-coded apps can be easily ported to multiple platforms, although (speaking as someone who routinely runs apps on Mac, Windows, and even Linux side by side) not always with equivalent performance results. Others would be so difficult to port that the time would be wasted. The irony is, the ongoing march of computer technology may mean the easiest way to use software on different platforms is to keep an extra computer handy. (Hint: rescue a computer from someone who’s going to throw it away, repair it, and laugh heartily.)

UPDATED: I’ve gotten additional details straight from the source, with specifics on specs, design concept, and software from the Trinity device’s creators; see our updated report.

LinuxDevices.com has an extended report on a new all-in-one recording device built with Linux; it’s been met by skeptical readers at Engadget and Music thing. There’s little point in spending a whole lot of energy now worrying about this product as it’s right now only a series of product renderings and a largely vague website. It’ll probably appeal to someone, with a large, built-in LCD screen, portable form factor, integrated XLRs, and the ability to run Linux audio software, and while US$999 MSRP sounds high, there are other recording workstations in that ball park without the software features. But would-be buyers are understandably unenthusiastic given they can simply opt for a laptop or tablet. Yes, that’s less mobile — but it’s also far more capable, and for many users, that’s well worth a little extra space.

Then again, you might want to look beyond this product alone. There are many reasons to be skeptical about the Trinity box, but the more interesting story here is that this could be the first of many mobile music and audio devices to run chips from ARM and TI and the Linux OS.

Trinity vs. Other Devices

Lately, product creators from large and small manufacturers seem to like to ignore existing products that work, like, say, laptops. Trinity Audio’s Ronald Stewart says to LinuxDevices, “What if you don’t have $300,000 to make a song? We want to let you record and mix a masterpiece on the bullet train in Tokyo.” It seems like a laptop would be a natural tool for the job, and indeed, while I’m not sure it’s a “masterpiece,” at least one Black Eyed Peas song really was recorded on a Shinkansen, with a Mac laptop and Reason.

The Trinity box will have to compete with dedicated field recorders starting much cheaper (albeit without XLRs) around US$400, more expensive dedicated recorders with proven reliability and pro features like SMPTE, new ultra-portable Windows boxes with better computer specs and compatibility with a range of audio interfaces and software, and, most of all, laptops, which provide easier interfaces and broader software compatibility with Windows, Mac, and Linux. I’ll be interested to see this actually ship, but it will face an uphill battle. October is AES season, so maybe we’ll see it then.

Other Mobile Devices in Store?

That said, I think the real story here isn’t the Trinity per se, but some growing tech trends. First, many devices are likely to forgo standard Intel chips for processors geared for mobile devices. The Trinity uses an ARM chip but still runs Linux; Texas Instruments, whose chips are in many audio devices you already use, is making a big play for Intel’s business. Add to that free embedded Linux licenses as opposed to pricey Windows licenses (and the fact that Windows is generally less flexible when it comes to embedded and mobile applications), and I think many more interesting products are in store. We’ve already seen the Korg OASYS running a modified Linux; LinuxDevices points to a new monitoring device for musicians from Glyph and upcoming Linux-based video recorders from Neuros. (I talked to Neuros’ CEO and he confirmed they’re making the move to Linux.)

So here’s the missing piece: software. Audacity is a mediocre wave editor to begin with, and even less suited to a mobile device. It seems to me that there’s a real opportunity for someone to start developing new apps with the mobile audio market in mind, from podcasters to musicians. Build a Linux app with a clean interface for mobile work and some creative features, and you could have this market to yourself. In fact, even if no one delivers dedicated hardware that people like, the app could run on custom laptops and ultra-portable PCs.

What would your dream mobile device look like? Or are you happy enough with laptops?

In other news, this whole product has unfortunately led to mostly-pointless debates over the merit of the Linux OS. I don’t think the Trinity hardware is the best device to use as an example in a Linux audio debate, but it has led to hilarious quotes, like this one from Music thing: “Linux is for people who buy Habitat furniture, listen to James Blunt and vote for the Liberal Democrats.” Sounds like the start of a great Linux switcher ad.