Buying equipment on the basis of “more” is relatively straightforward. If you need more microphones then regardless of which mics you add to your stock you can at least at the end of the process comfort yourself with the simple satisfaction of having added something tangible to your kit list. You can now put seventeen mics on a drum kit when last week you had to scrape by with only sixteen.
Buying better on the other hand is much trickier, first you have to define what better means? And as for those extra facilities or options, are they really going to be worth the money. After all the ability to do something “better” implies that you can already do something, so there’s always the nagging worry that you could have got the much more easily quantifiable “more” rather than the spending hard earned cash on the intangible “better”.
And hardest of all is to stump your cash for something that “sounds better”, in other words it doesn’t really add to your facility count it hangs on the most intangible of intangibles – the sound. Having worked in HiFi I know just how thin a line this can be – did the Linn Ittok really sound better than the Syrinx PU2 – I hope so I bought the Ittok. Though was it ever possible to arrange a rigorous enough comparison? And by the way if anyone out there still cares (or has both) I’d love to hear about it.
Which brings us to analogue summing amps, which commend themselves to us purely (probably a slight exaggeration) on the basis of sound. The idea is that summing your digital signals within your DAW is a process that degrades the sound and that a better sounding mix can be achieved by maintaining the channels in discrete form through D to As and then summing the mix in the analogue world.
In this review I’ll be looking at two summing amps the x.sum from Speck electronics and the Phoenix Nicerizer 2. And before we delve into the whys and wherefores of summing amps lets start with unadulterated experience. I packed my bags for Newport’s Not-In-Pill Studio where Martin Ford (ex Dub War and Skindred) does the business. Not-In-Pill has just upgraded to ProTools HD so my visit was nicely timed. First up was the x.sum.
Speck don’t have a huge profile in the UK though there website lists a very impressive list of users and their densely featured line/keyboard mixers have attracted a lot of attention. I was first drawn to their products when looking for the most mixer I could find in 1U to carry around with my Tascam MX2424 to use as a monitor mixer. The x.sum fitted the bill. It has 16 stereo inputs and you still get decent knob spacing – you heard it here first. The x.sum has a depressingly large external power supply, however given that we spend our lives or at least some of them worrying about power supply oomph I shouldn’t complain. But what I can complain about is the power supply plug which is big on compact but didn’t inspire me with solidity. Offering more facilities than most summers each x.sum channel has a full range gain control, a mono button and stereo pan and offers two stereo outputs (Mix B is available when used with a small break out box).
With a modicum of replugging the Speck was strapped across the ProTools outputs so that a quick AB of digital mix/summed mix was possible. The instant response in the room was really positive with Martin, Gethin Pearson and my son Andrew convinced without any doubt that the x.sum was doing the business. I wasn’t sure. So we set up a “blindish” test and as I was the least convinced I was to be the hamster. In this unscientific environment I was able to pick out the x.sum five times out of five. No kidding, I couldn’t tell you what it was that I was hearing so I merely state the fact. Time to do some homework.
I decided to call in some expert opinion. I began at Swiss digital supremos Merging technology. Merging are reponsible for a whole raft of hardware and software including the Pyramix DAW and high end converters in partnership with Digital Audio Denmark. Merging have also been at the forefront of DSD recording and processing and issues of audio quality are high on their agenda. Claude Cellier is the prime mover behind Merging and we began discussing the issue of internal word lengths. One reason often suggested for the perceived improvement in sound quality when using analogue summing amps is that the workstation is running into problems due to insufficient internal capacity to represent the mixed signal. So if you add two 16 bit numbers together you need at least a 17 bit number to represent the sum of the two signals. 32 channels of 16 bit audio will need 21 bit precision – according to my maths you need x + n bits to mix 2n channels of x bit audio. I really should have checked that with Claude. In that case a 24 bit internal mix bus looks sufficient after all we’ve got 3 bits to spare. But that is before we undertake any signal processing like EQ, if you want to mess with that 16 bit signal and you need more bits. You can see that 24 bit precision is beginning to look a bit shaky. Like some (but not all) DAW manufacturers Merging have tackled this problem by using a 32 bit float format (http://en.wikipedia.org/wiki/IEEE_754 is a good place to start) as their native format. So there should be plenty of headroom for lots of channels and lots of processing. And on the other side of the summing coin when using an analogue mixer, Claude pointed out that adding together the outputs of multiple D to A converters means adding together the noise and distortion present in each channel of conversion, maybe not such a good idea.
My next stop was at PrismSound, probably the most famous name in high quality A to D conversion. I nabbed Graham Boswell and thrashed the issue around in some detail. I think the conversation is best summarised in five bullet points:
- Good design is possible in either analogue or digital.
- With good practice it’s possible to get good results from analogue or digital.
- Digital techniques including anti imaging filters have improved significantly in the last ten years. Bit depth is not the be all and end all of digital design.
- Well designed equipment depends on engineering finesse not magic.
- Manufacturers shouldn’t dictate how customers use their products.
So loaded with this wisdom we return to our summing amps. The Phoenix Nicerizer 16 is pitched a fair bit upmarket from the Speck x.sum. It features 16 balanced Class A inputs summing to transformer balanced Class A outputs. The build of the Nicerizer is significant and cries “quality” from the rack. It’s 2U and the lovely red knobs are a joy to behold and yes even to twiddle. The only gain alteration is a +8dB switch on each channel those controls are pan pots. If you want to play with level – do that in the DAW. The Nicerizer offers comprehensive monitoring options though the LEDs seemed a bit sluggish, how’s this for a quote from the manual:
Although the metering only shows a maximum level of +6dB and the
LED’s are Red, this in NO WAY indicates that the unit is clipping!!
Ignore those meters, you’ve got another 20dB of gain beyond the red. But what about the sound? Plugging the Nicerizer in wowed the room – Martin was very, very impressed. I think the top end of the box was the first area to be remarked on – the crispness of the hf being something people homed in on. Such was the overwhelming joy brought by the Nicerizer that my pleas for blind testing fell on deaf ears. We spent the rest of the afternoon enjoying the sound of mixes through the new toy. At the end of the session I asked Martin if he thought it was a bit of kit they would buy if the cash was available, “definitely” was the reply.
An instant thumbs up then but how would that response fair over a period of time and would that response by universal? I left Martin the Nicerizer for a month to see if his reaction would change and also to find out what other people passing through the studio would make of analogue summing. Martin’s partner at Not-In-Pill Jeff Rose is the hard headed guy who handles the accounts and I was relying on Jeff to be a tougher nut. A month later I called back to find Martin still in love with the Phoenix box and that Jeff too had succumbed to it’s aural magic. The extended listening had left Martin enthusing about the warmth and punch of the midrange and still totally sold on the top end. What I found interesting was Martin’s comment that these were elements that had been in the original recording but somehow were getting lost in the mix, literally, when the mix down had been done inside the DAW.
What’s a skeptic to do? Well in my case I’ve been listening to the x.sum for the last month for what I originally wanted it for, a monitor amp, and it’s the bottom end that impresses me. Is it better than …. insert your choice of technology here. I don’t know and I really don’t care. More than ever this exercise has reinforced several simple rules of audio. Judgments have to be made on what you hear. And that means you, because different people hear different things. And finally if you do mix for a living in the digital domain you should get a demo of an analogue summing amp just in case. And depending on your budget either of these would be a good place to start.