Ancient History

Update 1 27/2/2014: Thanks to Arthur for new info
Update 2 09/3/2016: New info/insights from FY15 financial report + clearer grammer. I’m still getting regular hits to this old and rushed article! What do you, the reader, hope to learn? Share with me in the comments!

Creative Technologies (CT) became famous for their SoundBlaster soundcards in the 1980s. They were the first company in the world to combine FM Synthesizer (MIDI music) & PCM Codec (Voice/Sound effects) + a MIDI/Game port in a single ISA peripheral card. The all-in-one card was sold at a price that basically undercut the competitions at that time. From my memory, their main competitors were Adlib card (also an ISA card) & Covox Speech Thing, which is an external device that plugs into the parallel (printer) port.

Perhaps more importantly, the driver support was open to all PC & games developers, who started the multimedia revolution. This got a huge kickstart when Microsoft announced plans for the Multimedia PC (MPC) and SoundBlaster was the only card in town to meet those specifications.

From there, enhancements were made rapidly as new components (Integrated Circuits) with new capabilities became available from chip suppliers. CT was also very smart in ensuring new hardware and drivers were 100% backward-compatible so every new generation of soundcards provided an almost irresistible urge for gamers and enthusiast to upgrade! Even if the games at that time didn’t yet support the new features.

There were many competitors from Singapore & abroad but an aggressive marketing campaign + aggressive product roadmap soon see them to the pole position. Needless to say, CT became a darling on NASDAQ & SGX. Every new product launch will see a corresponding spike in share price & a nice dividend payout.

They were doing very well in the Soundcard industry in the 90s, especially with the acquisition of E-MU & then Aureal, which turned their soundcards so powerful, it can be used in industrial applications due to the E-MU DSP. Games were specially written to take advantage of all these 3D positioning sounds as Aureal APIs allow game and multimedia developers to place sound/music at specific spatial locations relative to the listener. Games written for Aureal 3D sounded truly immersive!

For music/video playback, only QSound encoded audio sounded more convincing with distinctive sound playing in and around your head. SRS & CMSS-3D extracted audio cues while stereo audio is streamed so these algorithm only expanded the soundstage and was not suitable for gaming. Aureal 3D was not suitable for music/video playback because you have to tell the algorithm where to place the sound.

DirectSound actually heralded the downfall of CT’s dominance. Microsoft introduced DirectX for multimedia & game developers to standardise on the PC platform. Intel also released reference motherboard designs to OEMs to copy. Microsoft followed up with Direct3D & DirectSound3D which brings arcade level gaming to PCs.

Being in pole position for so long, CT became so big & bureaucratical that they fail to notice the development of inexpensive soundchips being built & integrated into motherboards. Granted these early generation soundchips sounded horrible due to multiple issues like the inability to play & record at the same time + grounding issues on motherboards causing all sorts of noise & interference on the audio output & unstable drivers causing crashes.

CT was still selling SoundBlasters like hotcakes & everyone would switch off the onboard sound. They didn’t realize that after just a few generations of soundchips, these ICs were gaining in capability fast, like the ability to support multiple DirectSound3D streams + 24bit sound output + 108db sound quality + 7.1 sound output.

With each generation being less than 9 months, motherboard makers quickly learnt from their mistakes by isolating the sound circuits with better grounding & by moving high noise data pathways from the North & South Bridge chipsets away from the soundchips.

Windows drivers also got much better with Microsoft support. DirectX was also rapidly improving in capabilities, so much so that in less than 3 years, integrated sound became good enough to compete with SoundBlaster! $0 vs $300+ for a high-end Audigy Gold card or $100+ for a AWE32.

Consumers flocked to the cheaper solution of course, driven not just by price but the irrelevance of the main feature of the AWE-32 & Audigy. That feature is wavetable-synthesized music. With the Audigy, you can get studio quality music from a PC Soundcard that cost $300+ versus a professional card/deck that cost thousands. CT was slow to respond to this trend, probably blindsided by the professional market so it wasn’t till it was too late that CT has integrated Soundblaster in a few select motherboards. But due to cost from integrating the much larger Soundblaster IC, most OEMs didn’t support it. The market just dried up and now the Taiwanese company, Realtek is the undisputed leader.

In the past, if you wanted music in your software or games, the only way was through FM synthesizers. These chips modulated a base frequency into a shape that somewhat resembles real instruments. The best early example is the Nintendo Entertainment System. A simple script called a MIDI file has information on when to play a note and for how long on a selected channel. Every MIDI file has multiple channels each assigned an instrument. When all these channels were played together, a surreal synthesized music is heard. The main advantage of MIDI is these files were extremely small so were favoured due to disk & memory constrains in the past.

With 16-bit CD quality sound becoming the norm, the music portion needed to keep up and Wavetable Synthesis was brought over from the professional audio equipment market. Instead of a frequency modulated waveform being generated, the note from the MIDI file being played actually comes from a pre-recorded sample from a real instrument, called a patch.

At that time, if you wanted the most compatibility for wavetable sound, you had to go for the SoundBlaster AWE-32. But if you wanted the BEST music, you got the Gravis Ultrasound, which also supported a newer even more powerful music scripting format called the MOD(ule). This allowed more reverb, variation, cross-channel mixing, patch manipulation and unlimited patch size.

In the past, Voice & Sound effects were stored in PCM wave format whose file size was huge in comparison to MIDI. As with all inventions, necessity was the main driver of innovation. People wanted to compress CD-quality wave files to make them easier to transport & work with in games & multimedia applications, especially over the new media called the Internet. MP3 was quickly adopted & formed the basis of all music in games, media & movies over the Internet. This basically spelt an end to all synthesized music (both MIDI & MOD) in the consumer realm, rendering all Soundblaster AWE/Audigy & competitor standalone cards overkill. The integrated soundchips with their ability to handle 64 channels of PCM sounds were all that’s needed by consumers. From my memory, the first local casualty of this was Aztech Singapore with their Sound Galaxy range.

Unfortunately, CT’s trouble didn’t end there. A few misstep along the way soon confounded their recover. First was the CD-ROM drive. In the 90s, there were quite a few companies in Singapore making CD-ROM drives. However, once again, cheaper alternatives from overseas (Taiwan mainly) soon convinced consumers to buy those white-box OEM drives instead of the nicely boxed CT drives, which always seem to be a generation slower than these white-box drives. Wearnes Technology was almost wiped out during that episode. CT had a much better balance sheet so they were able to secure Singapore government assistance in the ensuing write-off.

However, the assistance comes with unintended consequences. According to (unverified) insiders of the deal, the agreement was for the Singapore government investment branch, Temasek Holdings, to take over a majority stake of the equitable shares of CT basically nationalizing it, & Sim Wong Hoo in effect lost the company he fought so hard to build. It could be this reason that CT was forced to delist from NASDAQ. The fund was injected through the Economic Development Board of Singapore. Sim Wong Hoo did eventually bought back much of the shares but at great costs.

Now, we all know that government top-down bureaucratic management style doesn’t work in a hi-speed hi-tech industry & that may be what happened to CT, where bottom-line was the overwriting concern rather than taking risks by pumping millions of dollars into R&D to create new products.

CT did have a second-wind with their Zen MP3 players which looked nice, had different form factors/colours, supported all common music formats, can be expanded with external memory & had great software that can transcode one format to another easily. However, Apple came along & stole the thunder with the iPod. CT’s lawsuit against Apple only secured a lifeline, not a victory.

CT’s foray into the 3D graphics card industry also met with poor results. The 3D GPU industry moved at an even faster pace than the soundcard industry. I’m not sure why CT acquired 3Dlabs, but I doubt it’s for their PC potential since CT used nVidia chips before calling it quits.

CT bought Cambridge Soundworks in 1997 with the hopes that the resulting synergy can reinvigorate the company but the gaming surround speakers failed to take off & CT had to sell the division away.

Along the way, they also acquired some gaming peripheral companies which didn’t work out and had to be divested. At this point in the 2000s, CT was completely off my radar and I no longer tracked any of their product range.

In 2009, CT launched a well-publicized but poorly executed (in consumer-space at least) campaign for their (poorly-name) StemCell CPU with “a 100-fold increase in supercomputing power over current technology”. No one understood what the fuss was about. CT’s only real showcase of the technology was an Android powered tablet that few people had a chance to try & a backplane-like cascade of ZMS-05 modules doing 1-teraflop but rendering something rather mundane.

As far as I can tell, the ZMS-05 platform is simply a dual-core ARM-based CPU with 3DLabs GPU component capable of performing 100 concurrent streams. The only difference may be the development platform allows the GPU to handle general programming tasks rather than simply graphical tasks. This may be similar to OpenCL for PC GPU. At that time, this capability was not available for the ARM architecture.

Unfortunately, it was not marketed enough & competitors like Qualcomm (Snapdragon), Samsung (Exynos) & NVidia (Tegra) soon signed up all the tablet OEMs leaving very little for CT.

CT did have a design win when Intel acquired the Ziilabs team for the StemCell architecture in 2012, probably to bolster Intel’s inadequate GPU. I wonder if it was integrated in the new Haswell Core-i CPU thereby giving that CPU the graphics boost it demonstrated in 2011. Haswell chips were eventually released in 2013.

However, for CT, it may not be a good deal because all the best engineering minds from 3DLabs have gone to Intel. Even though CT still owns the IP & Intel is technically licensing it, an IP that doesn’t generate new IP and/or products will be obsoleted in just a few years.

Lastly, CT also have a “Hanz” line of Chinese language hardware platform & software education system but those didn’t really help the bottom-line as shown in the Financial Report 2015.

So what’s left?

Not much actually. They probably ODM certain products, ship the design to China for manufacture & probably rebrand China products for sales elsewhere. Their revenue and R&D spendings have been reducing YoY & they posted a USD $33m loss in FY2015. Their longtime Creative Store at Marina Square is gone. CT is still holding on to USD $99m in cash + cash equivalent down from USD $140m in FY2013.

Where did it go so horribly wrong?

Like the fall of Rome, the collapse did not happen overnight. There must have been many events that lead to disaster. We could point to the failed investments in the past or the lack of direction or even top-level mismanagement. Some may even blame the Singapore Government in meddling in the private section.

I think it’s because CEO Sim Wong Hoo is 61 this year. With no succession plan in place (that I heard of), there’s not likely to be any change in CT’s direction nor will there be a change in their fortune or share prices.

Whatever it is, CT is a pale shade of what it used to be.

Is it a lost cause?

It really depends on whether they have someone to succeed the CEO position and bring fresh perspective & fresh blood into a still pond. It’s currently trading at SGD$1 on SES.

They’re trying to sell a high-end soundbar called X-Fi Sonic Carrier, which admittedly looks good and support Dolby Atmos 15.2 sound but carries a $2800 Pre-order price tag and ships towards the end of 2016. Most sound-bars comes bundled with a mid-range SmartTV so I’m not sure how well this will sell, especially since the launch price is an eye-watering $7000! I believe I can get an LG 50+” OLED UHD SmartTV with sound-bar + sub-woofer for that price!

Is Creative trying to do high-end audiophile market like Bang & Olufsen?

What can they do now?

In 2013, I thought AV Receivers may be worth investing in. I don’t think so anymore, what with most young people abandoning huge TV for huge phones with hundreds dollar headphones. For older folks who still prefer huge TVs, the AV Receivers & box speakers are replaced with a SmartTV coupled to a multi-directional sound-bar with a powerful sub-woofer.

The only gadget trends that’s big right now, with sustained interest for the next 3 years, is Virtual and Augmented Reality (VR/AR) goggles and Internet of Things (IoT.)

I doubt they have the engineering talent left for VR/AR, so maybe CT can start with small IoT projects based off ZMS-05?

They have a lot of great IP which they seem unable to tap. Soundcards (both PCI-e & USB) are essentially dead. With ARM CPU makers integrating DSPs inside their SoC, the DSP cores from E-MU are no longer special.

Without an aggressive marketing strategy like Beats (Apple) or Sonos, I’m not sure how much the Sonic Carrier can add to the bottom-line, even if the sound is awesome. CT really needs to find a niche where they can sell/bundle truckloads of these sound-bars. Maybe small auditoriums or lecture halls?

With that in mind, I have always wonder why CT never went into the AV Receiver market. That’s a market that people are willing to spend on great sound & video, including speakers!

The 3DLabs GPU can be put to great use by upscaling lower res video source from RCA or Component and outputting through HDMI 1.4 for HDTV. The streaming engine can easily be programmed to do deblocking, de-interlacing, Motion interpolation & sharpening of digital video, or 3D-comb filtering, smoothing of TV & other analog signals. Then to top it all off, perform a 3D-conversion!

For Audio, just couple that ZMS-05 module with a ton of balanced Class-AB OP-AMPs and they’re good to go!

The Android OS can be used for onscreen navigation with integrated apps for things like YouTube, Vimeo, Netflicks etc., similar to media boxes from Western Digital & A.C. Ryan.

Conclusion

As a Singaporean, it’s sad to see CT fall from grace but at least they’re still around. I hope someone takes over CT or let CEO Sim Wong Hoo retire from daily running. CT definitely needs a direction so they can recover soon before the cash runs out.