Update: PDF now available on my Designs page going through how to build your own GWM4 and everything that goes into it.
In this blog post I’d like to talk about some of the background behind the GW Model 4 speakers I’ve developed and cover some of the design process.
Over the past few months I’ve seen many engineers in a number of social media groups and forum threads asking for recommendations for smaller fairly inexpensive speakers which are appropriate for mastering. The Barefoot Footprint 01 which are priced at $3950 USD are a popular choice for this. I decided to see if I could make something for a similar price that offers at least an equivalent level of performance. The other major motivation that I had behind this project is that very few speakers fully use the potential of DSP available nowadays and I’d like more engineers to be able to experience the benefits. Far too many engineers think that their passive speakers are the ultimate solution. More on that later…
There were a couple of design goals I wanted to achieve in order to have a fantastic speaker for use in smaller mastering studios which are increasingly common. These are certainly not the ultimate speakers but for their “class” of speaker I think they are impossible to beat. I wanted a speaker priced under $4000 USD with extended low end (down to 30Hz), low distortion, great off-axis performance, SPL levels appropriate for mastering, compact size, use at 1m - 2m (nearfield), and fully utilizing the power of modern DSP. These are essentially a small version of my main speakers. I can’t incorporate all of the technology from my main speakers into something at this size and price but the result has exceeded my expectations.
Speaker design is all about tradeoffs and these are no different. Some of the main parameters which need to be juggled are how low the speaker extends, the size of the cabinet, the sensitivity (how many dBSPL produced per unit of power), how much power is required to drive the system, and maximum SPL levels. As the cabinets get smaller, it’s increasingly difficult to produce more low end. You essentially need to put more power into it in order to get the same low end performance which means higher power amplifiers and woofers which can handle that amount of power. This also ends up sacrificing the maximum SPL levels. If I were designing a similar system which only went down to 60Hz then I could design it in a way that it could play significantly louder. Based on the fact that I was designing a compact system and that it goes down to 30Hz (-6dB point) means that the max SPL levels had to be sacrificed. While this system may not play as loud as similarly sized speakers, I ensured it can handle the levels that I master at which I believe are fairly standard. I would not recommend using these speakers for tracking where you’re dealing with loud uncompressed transients and trying to fill the room for clients. Having said that, in comparison to the average 7” studio monitor, these can play louder and have lower distortion thanks to the incredible Purifi woofer being used. If this woofer did not exist, these speakers would not be possible.
There is a lot of debate around sealed vs ported speakers online and much of it is based around misinformation. One system is not inherently better than the other. Both can be designed very poorly or very well. If sealed were inherently better then the top speaker manufacturers like JBL, ATC, and B&W would not be making ported systems. The way that I like to think of it is as a relocation of the low end energy. With a sealed system you get a gradual rolloff whereas with ported you’re taking that lower rolloff energy and moving it up higher in the frequency range so that you get more output but a steeper rolloff below that. Unfortunately an important part of ported speakers which is overlooked in many systems is applying a high pass filter below the tuning frequency. Below where the port is working the woofer essentially moves without producing sound which just causes distortion. Removing this unnecessary energy cleans up the sound but unfortunately isn’t possible with passive components.
One important factor when designing a ported system is the air velocity travelling through the port. If it is too high, it will cause distortion from not being able to move through the port and will cause chuffing noises. A great ported system should have low air velocity at high SPL levels and this speaker was designed with that in mind which did result in sacrificing some of the compactness of the cabinet to have a longer port.
Another common debate is passive vs active crossovers. Again, both can be implemented very poorly or very well. A great passive crossover can make a fantastic speaker but that speaker could be made even better with an active crossover (and better yet with a DSP crossover). A passive crossover introduces all sorts of parts into the signal chain which degrade the signal, waste power, and make amplifiers misbehave. I believe there are 3 reasons why they are still prevalent -
Many swear by passive and aren’t willing to have their minds changed
In general active systems have the amps built into the speakers and audiophiles want to be able to switch out amps
Cost - extra amps and extra DACs in the case of DSP. Most people aren’t willing to make this investment
A DSP crossover introduces no components neither between the DAC and amp nor between the amp and drivers. It also allows for many more intricacies in the shaping of the slopes of each driver that allow them to perfectly cross over. It would require hundreds of parts to reproduce this with passive components. I believe that any passive speaker will be improved by removing the passive crossover, using one channel amplifier per driver, one DAC channel, and being calibrated correctly.
The GW Model 4 were designed to take full advantage of DSP which very few speakers currently on the market do. This is similar to having a DEQX or Trinnov built into each speaker. When I first started this project I was having a hard time putting together a system with would allow this as there aren’t any widely available that offer the processing power and I/O that I needed. Fortunately I was referred a company which offered a unit that had every feature I could’ve hoped for. At the moment I’m not willing to share which company this is but I can say that they are using the latest technology on the market offering more processing power than anything else around. Each speaker has a SHARC chip in it with excellent ADC and DAC sections.
The first area which the DSP deals with is linearizing the response of each driver. This includes both the frequency and phase responses which results in an incredible amount of resolution similar to what a DEQX provides. The second area is the crossover filters. This perfectly tailors the response of each driver to the desired crossover slope and also carries out linear phase processing for no phase shift. The last area is filtering for room correction. No studio is perfect and so I believe that every system benefits from correction EQ although it must be expertly implemented. It’s not simply a matter of inverting the response curve of the room.
With all of the processing running, roughly only 70% of the processing power is used plus less than half of the memory. In the future I may further tweak the filters in order to take full advantage of the processing power available but for now it is running everything that I’d like it to. Another potential future use of this processing power is to offer a digital output on each speaker in case a user wants to add a sub on each speaker (to extend below 30Hz and have more output power). This would provide the crossover and response correction for whatever sub is used for perfect integration with the system. Currently this digital output is not on the speaker but could be added by simply installing an additional XLR jack and wiring it to the DSP board.
All of DSP on the speakers does add latency. For mastering situations this is not an issue. Regardless, a switch on the back of each speaker allows it to perform in a low latency mode. This changes some of the filtering which results in an almost identical frequency response but sacrifices some of the sonic resolution. Other switches on the back allow the selection of analog vs digital input and a flat curve vs a room correction curve.
Following are some measurements showing the on and off-axis response of the speaker. These are windowed measurements meaning that they are essentially anechoic. Because of the windowing, these are only accurate down to around 150Hz. Below that the measurements are not correct. Unfortunately I would need to either measure the speakers in an anechoic chamber or outside at a higher distance off of the ground to get meaningful measurements below 150Hz.
The important thing to note in these measurements is not the flatness of the curve as that can easily be EQ’d to suite any user’s personal preferences. The point of these measurements is to see how little the response changes off-axis. It isn’t especially hard to make a speaker which is flat on-axis. Improving the off-axis response is where things get trickier. The DSP as well as the use of a waveguide allow this speaker to have more consistent off-axis response than most speakers around. I was very surprised by the performance which these measurements revealed. Intuitively it seems that only the on-axis sound would matter but it’s the off-axis sound which reflects in the room so it ends up greatly affecting the sound quality.
Horizontal at 15 degree intervals going from 0 to 60 degrees -
Vertical at +/- 15 degrees -
Part of the package included with these speakers is a calibration session with me. This gets the speakers tuned to your room and preferences. Proper setup is paramount to getting a great sounding system. Many people have questioned why I’m offering this instead of letting people do it on their own or including a copy of Sonarworks. I believe that the system would not sound anywhere nears as great and so I will always include this personal remote setup service with any speakers that I sell. These are skills that I’ve acquired after tuning a number of studios around the world. If that means I can’t process as many orders since I only have so much time then so be it. The calibration process typically takes around 3 - 4 hours and a follow up session to make any additional changes that the user may like. I’m also willing to offer an additional calibration session if the user moves to a different room.
Why Model 4? Where are models 1 to 3? Those are larger systems which are mostly still at a conceptual stage but can be built as a custom order. They are all designed around the same principles with the same DSP and amplification units. The difference is that they’re larger 3-way systems for use at 2m and farther. There may also be a Model 5 released at some point which will be a smaller speaker with a more limited frequency response range.
When it comes to comparing these speakers to similar units available on the market, it’s easy to see that these are being sold at a very low price and will outperform anything at a similar price. The first 5 pairs are available for $3300 USD per pair and after that the price will most likely go up to $4000 as I would have to prepare for a larger scale production run of these.
One commercially available system which is very similar to mine are the Jones-Scanlon Baby Reds. They use almost identical parts (although set up in a different way) except the tweeter although they use a similarly priced one. That system sells for around $5000/pair. Another similar system is the Ex Machina which is $8000/pair and again use a very similar DSP system and other similarly priced parts. Of course those are 3-ways but in terms of the parts cost that would add perhaps $300 per speaker in parts cost to my speakers if I went that route so there’s still a decent comparison to be made.
In terms of other systems on the market which employ similar DSP, there are the Kii Three, D&D 8C, and Meyer Bluehorn all of which are considerably more expensive. With the price that I’m offering these Model 4’s at I’m hoping that it allows more engineers to experience high end sound and the potential of the DSP which is only starting to entre the market.
If you’re interested in trying these out, get in touch and I will add you to the demo list. I’m willing to cover some shipping charges within North America but keep in mind that I’m prioritizing those who are willing to cover shipping and are keen on potentially buying these in the near future.