Click for Data Doubler kits!
|Accelerate Your Mac! |
Bring in the Noise
Mac Audio Column
by Thad Brown
12/12/98 [link revisions]
|I've Seen the Pulsar
Thanks to the mighty marketing juggernaut of the local famous-national-music-chain-retailer affiliate, I was the lone student in the Pulsar classroom. The Creamware rep was EXTREMELY knowledgeable and helpful, he has a degree in audio engineering and synthesis, so he oughta be at least knowledgeable, but he was even helpful as well. The Pulsar is still a very new product, and it showed a bit, but the basics of it are everything that the hype would lead you to believe. A few notes . . .
-The synths are a much bigger focus than some audio people like me might think. They all sounded excellent, and are generally in the techno and retro camps. They are very processor intensive, but since they run on the card DSP and not the host CPU, you can use as many of them as the card supports. Also, they didn't skimp on quality to get polyphony, and attitude I think is admirable. They are not of the "utility" variety, like a General Midi set of something, so I don't think that you would want more than a few running at once anyway. A bass and a few leads or pads would probably be enough.
-The whole system is modular. You hook up MIDI inputs and outputs to and from synths, and do the same with audio ins and outs. So, when you start the app from scratch, a mixer isn't even loaded, instead, a mixer can be selected from the list of choices (mixers, synths, I/O) and dragged into an active area. It's quite logical and easier than I make it sound.
-Perhaps the coolest surprise is how easy it is to automate the system. Right-click on a button (probably option-click when the Mac version comes out), select it for automation, then move any MIDI controller, and the Pulsar "notices" which controller you are using and maps the automation to that. VERY hip. The demo guy had a Kawai K5000S synth, which has a bunch of real time controller knobs that are hardware mapped to certain controller numbers. No problem with the Pulsar, just select the knob you want to control, twist a knob on the K5000 and off you go.
-I didn't get to see it do any real hefty audio DSP, like a big reverb patch. Don't know how many of those it would handle, but the rep said that four band EQ on each channel and four inserts were possible for the full 32 channel mixer simultaneously. That probably means delays and time based effects, but in any case, it's more than any host CPU currently available.
-Now for the bad news. Mac version has been pushed back to Q3 1999. I'm hoping that Mac demand might convince them to move some more resource our way, but that's kind of a bummer. It will mean that more third party modules, synths, and plug-ins will be available, which is cool, but I want the damn card now.
-Price/performance is predictably astonishing. The sampler alone is a full blown Akai S1000 compatible model that would set you back over $500. Add to that some incredible software synths, stereo 24 bit analog and 16 channel of lightpipe I/O, and more DSP than any host CPU currently in any desktop Mac or PC, and it looks like quite a bargain. In fact, with a pair of 8 channel A/D converters or a few ADATs or DA-88s, it looks like a really fantastic system for doing full band production and mixing.
The real make or break for the Pulsar, however, will be third party support. The audio geek rumor mill has that support being promised, but not implemented. If they come through, this will be something else. Finally, the rep said that all future Creamware products are planned to be cross platform, Mac/PC. Let's hope next time it will be in the same year.
Another Software Sampler
The folks at Koblo haven't been hanging around eating Edamer and drinking trappist ales. Or if they have, they can code when full and drunk. The Stella software sampler is a new entry in the line of apps built on their Tokyo engine. The have a demo available for download, along with the apparently free and fully functional Vibra 1000 synth. It's a super shrunk down mono version of the higher end Vibra 9000, but it still sounds cool and has an arpegiator. The sampler follows the "one window" philosophy of Koblo, and at first blush seems pretty cool. I'll write more when I get a chance to use it a little more.
Just In Case You Have Any Money Left . . .
. . . go buy some new converters. In my constant carping about the new audio cards and interfaces out there, I actually missed the introduction of a number of interesting looking new analog to digital and digital to analog converters. The converters are without a doubt one of the most critical points for audio quality or the lack of it. If whatever you are recording through now sounds weak or thin or noisy or all of above, there is a good chance that converters you are using are not up to the task. In fact, in my experience, the two spots where the biggest "bang for the buck" improvements can be made for digital recording is the mic preamp and the converters. A very good, clean mic pre and good A/D can make an SM-57 sound like a very different creature.
The inside of a computer, as any audio dweeb can tell you, is a firestorm of electromagnetic, acoustic, and radio interference. Audio cards have improved dramatically in the last couple of years, and shielding certainly helps, but the best thing to do is move the converters completely out of the computer and into a rack somewhere else. If the cabling going from the outboard converters to the computer is good, this is the best and safest way to go from analog sources to your computer or recorder.
The three additions to the world of converters of note are from Midiman, Sonorus, and Lucid Technologies. All share some characteristics, most notably that they are all stand alone converters, not converter/card combinations, and that they use an optical interface for transfer. Keeping the converters stand alone means that they can be used with any system, and can work with something you buy in the future, so you are not tied up to a proprietary system that may not do what you want in a few years, and the optical connections are blessedly resistant to most of the kinds of interference that plague coaxial cables. So, here are this years models and my take on them.
The Midiman interface is an 8 in 8 out system with 24 bit converters and lightpipe for digital I/O. It can dither the output down to 20 or 16 bits and also includes a proprietary method for splitting a 24 bit signal across tow 16 bit tracks, in effect making a 16 bit ADAT into a 4 track 24 bit recorder. Such a recording, though can only be played back via the same hardware. Sample rate is selectable and the box has world clock in and out.
Sonorus is the folks who brought you the StudI/O interface for Macs and PCs. Many of the well heeled types I know who do serious multitracking on computers use one or two of them in their machines and external boxes for conversion. Sonorus has announced a high end 24 bit/96k converter for more than I would pay for a car, but bless their hearts they are making a cheapo model as well. They are splitting the job into separate boxes, which is a fine idea, so buy eight channels of EITHER analog to digital OR digital to analog. That way if you know you'll never need more than 8 channels for monitoring or mixing, but want 24 channels coming in, you only have to pay for what you need. The budget models have 24 bit converters, but only work at 44.1 or 48k. Not a problem for most of use. Sonorus has an impeccable reputation for pro quality and reliability, so this one may be a real winner.
If you only need two channels, or you really need 96k sample rate but can't spring three grand for an 8 channel system, Lucid has a very attractive new toy. They also decide to split duties up with A/D and D/A in separate boxes, but add a few twists. First they are stereo boxes only, not 8 channel systems. Second, they will use 44.1, 48, 88.2, or 96k sample rates as you choose. And last (but certainly not least), they will put the digital signal on three different stereo output pairs at the same time; one coax S/PDIF, one optical S/PDIF, and one AES/EBU on XLRs. This last feature solves a host of potential hook up problems and means that you could take a rack of them anywhere and know that the digital signal would be possible to get into anything that had a digital in.
Good quality converters also look right now like a fairly safe place to put your money. I doubt there will be much call very soon to record at higher word width or sample rates than these boxes offer. The Lucid box in particular looks like a nice one since it will even do 96k, albeit in stereo. The 24 bit.44.1 converter may be a fairly safe "sweet spot" place to make an upgrade. You know, like it'll be at least a few months before you are upset at how much they cost.
Home Mastering Tips
Though I repeat I am not a mastering engineer, I got a LOT of emails asking for tips on home mastering. I would always recommend getting somebody else to do it if you are not confident, but many of us can't afford that step, so it's ourselves or no mastering at all. So, here are the official "XLR8YOURMAC I'm not a mastering engineer but I play one on the web" mastering tips.
Listen to the whole song and the whole band. The intro may be really soft or the backing vocals may not come in until the second chorus. If you set the compression and EQ for one part and don't listen very carefully to the rest, you may wind up with things not working at all with the other parts of the song. Listen to the whole band to be sure that there isn't one part that overwhelms the whole, even if it is the coolest part. Guitar players always want to be too loud, it's in the DNA, I know because I am one, and I had to learn to turn myself down before my mixes ever sounded any good. Make sure everything can be heard, not just the parts you like.
I try to EQ subtractively if possible. Boosting signals can increase noise and sounds more invasive to me most of the time. When working with a compressor, you are lowering the volume of the loud sections automatically, and then increasing the final gain of the part to give an apparent increase in volume. When dealing with full bandwidth source material, like a rock band, it's likely that one or two things will trigger the compressors more than anything else. If a part is too loud in one frequency band, often a very modest cut BEFORE the compressor in that band will keep the compressors from clamping down too hard when that part is present. This will make the compression sound smoother and more natural, and will keep that one part from standing out too much. Also, if a part needs boosted, sometimes a slight cut just above or below that instrument may help it stand out more than boosting the frequency that makes it stand out. This one doesn't always work, but try sweeping through to find the band you want to boost, then go up, say, a half an octave and put in a small dip. It may get better results than boosting the other band.
Spend $50,000 on managing room acoustics and monitoring systems. Or be damn careful about how your room is messing with the sound. I broke down and bought a new power amp (which I needed anyway) during the Jellyshirts mastering job. With each song I tried to listen to it both at rock club and whisper volume levels, with the louds for smaller mistakes and distortion, and the super low for balance. If it sounds good both places, it's probably pretty close to right. But at the super high volumes, I couldn't tell if I was distorting because of level problems in the digital realm or if it was my underpowered amp starting to choke. Out comes the credit card and in goes the Mackie M-1400i. Combined with a basic understanding of my room sound, the power and decent monitors let me be pretty sure when I was on the right track with EQ and compression decisions.
Send Thad Feedback or new links at: firstname.lastname@example.org
Your Source for the best in CPU/SCSI/VIDEO card reviews, daily news, and more!
and do not necessarily represent those of the site publishers.