2014/02/09: An edited version of this post was recently republished on DJ TechTools.
DJ software: complete with the kitsch in sync
There's a view — one I guess is primarily propagated by people who were DJing long before the likes of Traktor and Serato1 came along — that using software on laptops doesn't really "count" as DJing because you don't need to beatmatch. Algorithms for beat detection usually work well enough these days that software can usually have a good go at this for you.
It's a slightly odd opinion though, and one that I actually think does DJing a disservice. The implication of beatmatching being so important is that there's little more to DJing than that. Just get the beats matching up or push the sync button, and, that's it, congratulations, you are now a DJ and your membership card will be in the post!
What's strange to me is that you'd think, as DJs, they'd be aware of the importance of everything other than beatmatching that goes into even a simple A-B mix. Track selection and sequencing, transitions between tracks, EQing, harmonic mixing are things that sync won't handle. Software's not really going to easily handle this for you2 as a lot of it is down to personal preference and choice, as well as the type of music you're playing. (Oh, sync is entirely optional too: no-one forces you to push that button.)
Why beatmatching is worth your time
Having said all that, there are still good reasons why it's still worth learning:
Lets you use other equipment. Imagine you only ever rode a bicycle with stabiliser wheels; great, you can ride a bike all day long. Except the time you have to use a bike without stabilisers instead. It means it shouldn't take be too much of a struggle to adjust if you end up playing on CDJs or vinyl/DVS.
Sync is good, but not always perfect. Annoyingly, Traktor still uses a fixed beatgrid. (I think Serato actually has warp markers, but I'm unfamiliar with it, so feel free to correct me on this.) Provided a track has a well-defined bpm, which most electronic dance music does, it does pretty well at detecting it.
However, a simple beatgrid can break down with tracks with live, unquantised percussion or where there are particularly unusual rhythms or switch ups.. Knowing how to handle cases where sync fails completely or where minor adjustments (e.g. with tempo bend) are necessary is essential to keep beats from drifting.
No need to analyze and beatgrid tracks in advance. This isn't a bad thing to take the time to do: with Traktor: tempo-based effects and looping can sound off if the track does not have the correct beatgrid. On the other hand, being able to throw a new track into software and immediately play it without worry is really convenient.
Mixing into the end of someone else's set. Some DJs will start afresh; some will beatmatch into it. Choice is a good thing. No DJ software beat matches to an external sound source as far as I'm aware. (I'm willing to be corrected again though; it's not just me that's thought of this.)
Beatmatching is good practice for listening closely, especially when you're trying to selectively "tune" in and out from the track being faded in. This is useful regardless of how you're beatmatching (e.g. for EQ purposes).
What Traktor offers
Traktor has several features to aid beatmatching. If you're learning to do this manually, the natural inclination is to think, "OK, if I'm learning to do this myself, I shouldn't use any of these tools ." But, when you're learning to manually beatmatch, these features are still really useful. So what do we have?
You can practice with the phase meter to see how far off you are after you think you've beatmatched correctly, or just check how far off the beat you are when hitting play or scratching in a track. The phase meter shows you how far ahead or behind the current track is relative to the track that's currently the master. If the bar is in the middle of the phase meter, the track is in phase with the master. In the example above, the track is over a quarter of a beat ahead.
(Note that both tracks you're playing need to have correct beatgrids otherwise the phase meter is likely to be misleading.)
Of course, you can always use your ears for this, but the phase meter gives you a quick visual indication of how well you've done.
You can check BPMs immediately when you think you've matched tracks as well as possible3. One easy way is just to click the GRID option directly below a deck (you can also add a permanent BPM counter in Preferences > Track Decks > Deck Header.)
This gives you a direct indication of how far off a match you actually were; if you're only a couple of hundredths of a BPM out, you're doing pretty well.
Importantly, you can see that you're improving even if you haven't quite got the hang of it. To reuse the bike analogy, the learning curve is pretty steep. First, you fall off a lot, then you start riding albeit a little wobbly. I found beatmatching pretty similar: for quite a while, I couldn't manage it, then all of a sudden I could (though quite badly). You then become steadier and grow in confidence with practice.
(One gotcha with beatgridding is that if Traktor may well have detected a BPM half or double what it should be, e.g. tracks at 140 BPM might be detected as 70. If the BPMs of the two tracks are reportedly wildly different, but you seem to have beatmatched them nonetheless, that's probably why.)
One exercise often mentioned in beginner beatmatching tutorials is to mix two identical tracks. When I was starting out, I found this sounds cluttered and confused me a little. Unsurprisingly, it's sometimes really hard to distinguish the two tracks as they sound the same.
Using different tracks might be easier. However, the tracks you're using might not be the same speed which leads to the chicken-and-egg problem of how can you beatmatch before you've learnt to beatmatch? In this case, the sync button should handle it for you with one push of a button.
If you're not sure what a "perfect" mix of the current two tracks should sound like, the sync button will give you that too. You can then nudge one track slightly out of phase with the other using tempo bend to easily hear the difference, say, when a monitored track is ahead versus the live track; also useful in developing beatmatching skills.
It's not always easy to judge how well your beatmatching is while you're actually doing it. Recording your practice and listening back to see how it actually sounds to a listener is pretty simple with one click. Are the beats matched together perfectly or not? Did you let the beats drift at all through a blend?
A comment on MIDI controllers
One design choice that I would criticise about many MIDI controllers is that their pitch faders tend to be on the small side. This makes it difficult to accurately beatmatch, since precise adjustments (hundredths of a BPM) are tricky. One workaround is to map a knob as a fine pitch control with a much smaller range.
This lets you lock in to the rough BPM using the pitch fader, then use the knob to make very small corrections.
As long as software doesn't perfectly get beatmatching correctly with zero preparation, there's still a need for beatmatching. Not only that, software can actually help you practice and improve too: I self-taught myself beatmatching with Traktor and a MIDI controller. Let me know if you've done the same.
Also see open source alternatives like Mixxx and xwax. Mixxx in particular looks really promising. Unfortunately, my controller isn't really supported (the pitch faders don't work correctly when I last tried, using a old VCI-100 running DJ TechTools v1.4 firmware to improve the resolution of the jogwheels and pitch faders.) ↩
Note that there are key detection algorithms that can help guide harmonic mixing. There two big problems though. One is that they're not perfect yet; the estimates there (on a small set of songs, admittedly) is around 60-70% accuracy. And, if they were, all that does is tell you the current track's key. Even if you select your next track on the basis of key, it's not going to decide this for you and there are several choices. Do you stick to the same key? Do you pick the relative major/minor, or the parallel major/minor or even ascend/descend by a perfect fifth? Or completely eschew all that and perhaps choose something in a different key, maybe bringing it in a non-melodic section? ↩
Again, you need to make sure that the tracks you're using for practice have been correctly beatgridded. ↩