I saw that Native Instruments is now selling Traktor 3 the DJ sofware for $119 ($99 over at Guitar Center) so I wanted to check it out.

I have this strange relationship with Traktor. Every time before I tried the demo version, it crashed at some point, so it was not auspicious. If there’s something a DJ software has as a requirement, it is rock solid performance, not fun rebooting computers during a big.

Anyway, this time it worked just fine. This was just a 15 minute check. My litmus test was to open up various 80ies and earlier dance tracks and see if it was easy to mix and match those.

Traktor 3 finds suitable locations for the beats along the way, so it’s seldom a need to do pre-warping. Or, the automatic warping works really well. That’s a plus. However, the starting point for the beats were always very much something that I didn’t like or want. It means that there’s always a need for cue:ing, and I’m so used to pre-warped tracks in Ableton Live that I know where they start from, so I don’t even need to cue that much today. So that is a nuisance.

The warping algorithms in Traktor sounded really good, even with moderate pitch changes. With extreme pitch changes with same tempo there were indeed artifacts, I think even more than in Ableton Live’s complex mode. But I must say that the default warp quality values were somewhat more pleasing in Traktor.

As for the user interface, well it took a while to find out how to change the global tempo. The UI is still cluttered, even if it has been cleaned up since the early days. There are too many semi-modal boxes to jump back and forth with.

The integration with iTunes was a good point, Ableton should introduce that in their next version, too.

All together, I’m still happy with Ableton Live. They should just introduce a better quality complex mode warp algorithm.

This is what the waveforms look concerning importing and autowarping George Clinton’s Atomic Dog.  This was a huge hit for George Clinton back in 1982 and it’s a good representation of the P-Funk sound, not to speak it’s pretty hilarious, too. I think I even DJ:ed that track once very long time ago, like in the early eighties.

Anyway, this is a good example of the dynamic range of tracks up to about late nineties when the dynamic range wars started.

What I do across all my DJ tracks, now covering from obscure funk 45 rpms done in the sixties to whatever today is to make sure that the gain level is the same across all the tracks. I defined 7dB headroom as a nice compromise — seven is a lucky number and it does not really matter as long as you keep a standard. The dynamic wars tracks I just decrease the headroom so it’s in the 7dB range.

The tool I use is Sonalksis FreeG plugin. It’s a free plugin and it monitors the levels. I place this plugin on my master track and set a nice loop of 24 bars or more in a place that seems to be the loudest part in the track. Then I observe the RMS values, this is the average mean that is more interesting than peaks here and there. Looking at the existing values I could decide how many dBs to increase or decrease in the clip view (don’t use the plugin itself to increase the value.) When done I just save  this information as part of the .asd file saved for the track/clip. 

I might have written about this before, but my iTunes and Ableton integration has evolved over time, so here’s the latest work flow. The reason I’m using iTunes with Ableton Live is that even if Ableon Live’s browser has evolved over time, it still does not have a lot of features that are really good to have when working with multiple tracks, especially when scaling up to hundreds and hundreds of tracks.

To start with, you could select a specific iTunes library when you hit the option key as part of starting up iTunes. This will give you a dialog box to either create a new library or select an existing one. This makes it possible to have a separate DJ music library. I have actually two as I changed my playlist to have old seventies/eighties funk and good disco tracks, but I keep the other one around in case I suddenly want to play tech house and so on. Anyway, this is not necessary, you could have one huge DJ library, or even use the one and only library you use for iTunes. Anyway, it also makes it possible to keep the DJ library on an external disk, only copy this library to a laptop and so forth.

Secondly, you could just drag and drop audio files from iTunes to the Ableton Live browser. I wish they will fix the issue of using the clipboard for copy/paste between the applications in Ableton Live 8. Then this would be even easier, including copying over multiple entries.

Next thing to remember is that Ableton Live picks up the .asd files with warp info from a corresponding .asd file with full same name as the original audio file. If you have already warped your audio file and drag this file into iTunes, iTunes does not know anything about this dependency so it does not copy this file over. If a future Ableton Live embedded the .asd information inside audio container formats that support meta information, such as MP3, CAF, AAC and so forth, then this would not be an issue.

Anyway, what you need to then do is to select the audio file in iTunes, and with a right-click or control-key click get the context popup menu. One of the entries there is to Get in Finder. This will get you a window with the location where the iTunes managed file resides. You could now drop in your .asd file and things are fine. Just remember that iTunes sometimes renames the audio file so double-check that your .asd file has the exact same starting file name including any extensions such as .mp3 and that the .asd is added to the end.

Now, another approach is to first just get the audio file into iTunes. When you want to warp, drag this file into Ableton Live. When you save the .asd file via the clip panel it will save the .asd file into the location where iTunes has the audio file.

There’s even a third approach, iTunes could refer to audio files outside it’s library folder, see iTunes documentation for more information in case you don’t want iTunes at all to manage your audio files.

So now we are at the point to discuss what is so good about iTunes… To start with, as part of warping I figure out the key, major/minor and so forth. I also register the BPM speed after the initial  auto warping. I then add this information into the file info inside iTunes. Again right or command-click and select Get Info. You will get a dialog box such as the in the image above. BPM has its field. I use the Grouping field for the key information as there’s no default key field in iTunes. Also add or modify any other information as you wish, including creating your own Genre by just typing into the Genre field (in case the pre-existing ones are not your liking.) If you don’t like to register BMP for a lot of existing tracks inside your iTunes library, check out Tangerine.

When this is in place you could do sorting in the iTunes view based on these values, provided you activate any missing columns — again right or command-clicking in the column header section is your friend. You could sort based on key, or bpm, or artist, genre, date added and many other values.

I even use the rating star system; my approach is to register the level of ‘energy’ in the track from one star (ballad), two (moderate dancing), three (normal), four (sweaty) up to five (energy-drink level.) This way I could also do sorting or searches narrowing down the style so I don’t place a slower paced track in the middle of a sweaty section.

What else? I could make play lists with recommended sets and save those in the iTunes database, do all kinds of interesting searches, just by typing a couple of letters I could quickly find all my Isley Brothers tracks and so forth. I could even make smart play lists that find things based on key, style, date added and many other values.

Also, it’s fun scrolling through the entries using the cover flow and see covers… Data backup of the file to DVDs spanning multiple DVDs is also easy. And that’s just a tip of the iceberg when you could do all kinds of interesting things inside iTunes itself before you get the track to Ableton Live.

Now, if Ableton Live in 8.0 could synchronize to iTunes libraries in forms of showing playlists and smart playlists there might not even be a need to drag-and-drop any longer…

I’m sure there’s even more to it. Yes, it takes a little bit time to add tracks to the database including registering meta information, bu I see more benefits from this all from this work. Feel free to add your own comments in case you have other cool ways of using iTunes and Ableton Live, I could always pick up a new trick or two.

Here’s another technique I’ve used when not being sure where to place fixed Ableton Live warp markers. In this case it is Janet Jackson’s What Have You Done To Me Lately. First time I heard this track was at a huge club in London when I visited the place while taking company courses in Birmingham… That track over a huge and good PA system made a big impression, indeed.

Anyway, the actual warping of this track should be very straight-forward, as there’s a very definite beat and Jimmy Jam&Terry Lewis as producers definitely has the beat locked down.

There are, however sections, where’s this big reverse cymbal at the end of the fourth beat, and it will make the waveform look very muddy and unclear concerning where the border beat for the fixed warp marker should really reside.

What I’m doing in such cases is just to make sure that the previous line, here 118.4 in the image, is aligned to a visible transient. By the help of that the next section, 119, will naturally align at the right position.

Remember that the warp markers in Ableton Live are like rubber band knots, if you move something around, the rest between the fixed warp markers will stretch and re-align, so you could always see how it all fits together looking at all the transients. Some of them will make visible blocks of audio, and if those block fit well into  the lines, then the warping should be OK.

Old tracks have very different problems concerning warping with Ableton Live. To take a test case; Chic’s Everybody dance is a very famous disco/dance track. I purchased the 12″ extended version (from Amazon’s MP3 download site) and thought that this would be a 10 minute job concerning warping.

After two hours or work spanning two evenings I think I got it nailed down, for the time being. There were many issues.

To start with the rhythm section, the drum and bass,was tight but not so tight concerning the beat. When you listen to the track it sounds tight, but when analyzing it all with Ableton Live you start to see all kinds of misses and hits here and there. This was especially true at the end of the track. I assume the rhythm section got tired after playing the patterns for over six minutes.

The other interesting thing I learned an evening later is that the downbeat for specifying warp points was masked by a early triggering clap sound. I put warp markers on the kick sections, and usually the kick and snare/or clap aligns nicely. However in this case the clap was far too early (see the image). Even more, the early clip triggering variated across the track.

I could finally figure out where the kick was happening, there was this very thin peak that repeated itself in all these problematic sections, so I placed the fixed warp marker on this spot and got it all under control.

Warping electro house tracks and techno is easy. It takes a lot of effort to warp old tracks, indeed.

So yes, I’m really interested in classic dance music today, all the way to tracks I listened to as a kid on Radio Luxemburg, that radio station that covered covering most of Europe on the AM band until 1992 (1440 AM to be precise). I wonder who listens to AM radio today, you got these strange amplitude bumps now and then and the Hi-Fi quality was not there, but hey you heard a lot of interesting music. Especially in the 1972-78 range when I sometimes listened to the service until 6am in the morning and the station faded away, they played a lot of post disco dance music, Jackson 5, Ohio Players and so on.

Anyway, there’s something special about dance music performed by musicians, drummers, bass players, guitar players, keyboard players and so on. They have a nice groove that is tough to reproduce as a one-person loop producer.

As dance music today is more and more stereo-type due to the ease of producing it, old style musician-centric dance music is for me now very refreshing and interesting. It pops out from the mass of dance music out there. Plus there’s something human about it.

So this is the reason I’m mostly focused on putting together set lists using older, band-focused dance music. But, if there are contemporary tracks also done by musicians, that’s fine as well.

Recently I’ve been purchasing and warping a lot of dance tracks from the late seventies and early eighties, Chic, Jackson 5, Sister Sledge and so on. Why is another story. Anyway, I tested with the latest version of Ableton Live (7.0.7) in case auto-warping has improved since the last time. The test case was Jackson Five’s Can You Feel It.

The initial check using the built-in metronome actually sounded like auto-warping did a proper job, to my positive surprise. The second taste case was another Jackson 5 track, however after three minutes the auto-warping got out of whack.

Going back and listening to the first track critically I noticed that the alignment of drums was not that perfect, after all. So we are back on the same state as before, Ableton should really focus more on auto-warping and make it more robust. I do believe that it’s possible to design a proper auto-warping algorithm as it’s not rocket science to analyze various bandwidths and figure out where the drums align. It might be one of those cases where before it took too much CPU power, but with Intel cores sitting idle with the new chips this should not be an issue.

Anyway, when doing warping by hand it was really interesting to see how things were aligned. The track Get on the Good Foot by James Brown was super-tight, his background band was so on the pocket it was incredible.

When warping a special 12″ version of Chic’s Good Times, the drummer and bass player at the end part, after seven minutes, kind of collapsed and the beats were not aligned at all. Now, this was in the age of no copy/paste of loops, rather having musicians play the same thing for up to nine minutes, so it’s understandable but interesting to notice now afterwards.

When I warp old tracks, there’s the issue of swing taken into account or not. A lot of old songs are performed with a lot of swing. However, personally I think for DJ performances it’s good to warp the tracks to be very tight, makes it possible to mix the track with any other track that is also warped in a tight matter, as well as introducing other contemporary loops and so on.

Next more about my new DJ musical direction concerning dance music.

I just finished another BioWaves episode (010) using Ableton Live and I used a couple of non-standard transitions that I thought would be fun talking about.

To start with, BioWaves the podcast is not done live, of many reasons. I’m always running out of time to get the episodes done, there’s a lot of work, selecting tracks, doing artwork, finding urls to the tracks (sometimes that takes a while as the MP3 files or web sites don’t have enough information.) There has been many cases where I’ve pulled out one track and added a new one, a day before release. There’s a certain pattern I try to keep concerning track selections (more later.) And I try to play each track three or four minutes without any editing. To do the actual episode live would just add another level of complexity. Anyway, I will mention how these techniques could be done in a real live situation, and if I play live, I indeed play live, no cheating there. Anyway, enough excuses.

The first one was to just suddenly stop one track and keep the other one up, in other words no what-so-ever fade-outs. This is a very dramatic and cool effect, especially if the ending is a big reverb washout or something that ends with a big band. To do this in a live situation, either use a reverb on a bus and let it ring while you quickly take down the volume for that specific track — and meanwhile the other track is playing.

In my case it was just pure luck that the first track ending I’ve selected randomly had a nice washout sound, while the other track then continued happily with its beat.

The second technique has to do with how to handle very dramatic BPM changes. In my case it seems I have nailed down a certain pattern on the one hour shows where there’s a build-up to the 50th minute. From that forward I want to take the pace down, play various chill and otherwise non-standard dance music tracks, and those are usually down to 110BPM or below. While the build-up was up to 130BPM or higher.

What to do? Well, my standard trick is to take down the bpm levels as part of the ending track where there are no drums, and hopefully no fast patterns such as arpeggiators. Slow melodies are the best. By taking down the BPM levels — sometimes even 20 BPM or more — few would notice it. If there’s still something unnerving about that dramatic BPM drop, I could add in ping-pong delay och a big reverb effect that further masks out the big drop.

You could also easily do this in a live situation, just find the right spot. If you don’t have one handy, make a loop or a clip with something that could be used to taking down the BPM levels and loop this around.

Which leads to the third trick I use from time to time. I really want certain tracks to play together. Usually it’s easy, but there are always some problem cases. The trick I use in those cases is to make a short four-bar or so loop of the ending track, and let this one loop for a while. Meanwhile I fade out this track and let the other come in, either via a filter, or just by plain volume fade outs/ins. By finding a good loop point that works across the whole fade makes this much better. This is especially useful for tracks with very complicated endings, and where you even can’t fade out/in as it starts to suddenly sound strange at some section.

Again, in a live situation you just loop a part, or have a handy Ableton Live clip around and use this one when fading out one track. You could also use a loop for fading in a starting track. Or loops for both cases. As you see in the picture above, I even used two very short loops for the fade in case and two longer loops for the fade-out.

I’m afraid someone might protest about all these artificial transition techniques. That’s fine. Ultimately it’s really the audience that should make any judgments  if they like the mixes or not. Actually my hidden agenda with this posting is to inspire other producers and DJs to use non-standard transitions — the fade in/out is so boring and predictable nowadays…

DJ transitions are those parts between each individual track where the DJ is connecting one song to another.

The classical transition is to merge two tracks so the audience does not notice where one ends and the next one starts. In the progressive house movement this was taken to the extreme where it could take minutes before the new track took over.

In this new age of easy beatmapping tools such as Ableton Live, various CDJs, Traktor, and so on, long, seamless transitions is something that anyone could do, as long as they select tracks that easily merge together — either the key or the production sound should be similar to achieve this.

Which leads to my point. I don’t really see that special to even have long, orchestrated transitions nowadays, of many reasons. First, it’s just so boring, as anyone could do it, there’s really nothing creative behind it. Secondly, with the tools we have today one could swap in and out various tracks or sub-tracks all the time, so there’s maybe no need to have a special transition, just a continuous show with some breaks here and there.

Thirdly, it might even be in a DJ session that such transitions will steal energy from the dance floor, maybe it’s better to do something like ‘kicking in the third gear’ now and then to suddenly chance the pace, and the audience mostly likes to be positively surprised compared with being slowly bored.

One reason I started to think about this during my morning bike ride to work was that when I did the rough mix for the next BioWaves episode, the transitions were somewhat different, not the normal ones. I was thinking about if I should change the ordering and line up another set of tracks. But then the light went in my head: it’s good to show different kinds of transitions instead of the normal fade out/fade in ones. So I will stick to the current order and ‘think different’ transitions — I’m sure someone will complain and state that they are not professional, but they are indeed different. I rather have some interesting surprises here and there in a mix, compared with a longer period of mixed-together uniformity.

PS: Here’s a creative challenge, take eight tracks or so, totally random ones, if possible different styles, and either live or by using time line tracks make transitions between those totally different tracks. By being forced to do unexpected transition techniques you might get insight into new tricks you will then use later.

Livid instruments just announced the Ohm DJ Controller, and it sure looks tasty. You could check out the specs yourself at their site. It was a clever idea (and they try to patent this) to have a 36-button matrix in the middle of the controller view with two sides for both hands.

It’s also good that they focused on the aspect of lighted controls, it’s darned hard to see controls in a dark DJ booth from time to time.

Anyway, this is part two. There has been all kinds of variations concerning making the ultimate DJ controller, or live controller. They all have good ideas. But I think the hardest issue is that every DJ and laptop performer has their own ideas how the layout should look like. Sometimes you could do compromises, but if you want to pay a lot of money for a controller, you really want it to behave as close as possible to the vision you have.

For example, in my case I would like to have sliders for various channels, and on top four rotary pots, and one of these should switch the behavior of the other three. For example, in one mode the other three behave like a normal 3-band EQ. In another mode, they are used to activate and control a beat splicer. And so on.

I would not mind to have a touch pad, as it will take a while until I think Ableton Live will have non-mouse support for all the features I need, especially concerning finding clips and inserting them to tracks.

And that’s just the tip of the iceberg. Every DJ and performer has their own small things they would like to have. If any of the big companies, M-Audio et rest, would just go ahead and provide a modular approach to controllers, that would be better. There’s always the option of building your own controller, but I’m afraid that it takes time and effort, and meanwhile most of us are busy doing tracks, so a commercial solution would be better.

Even better, such a solution would most likely have customers constantly buying more parts.

Yes, I know of Mawser. They never seen to ship their product, the look is industrial, and their prices are too high. Maybe if some other company purchased Mawzer and finalized their product…

Something that I use from time to time in mixes, and also in live DJ cases with Ableton Live, is to put in a very fat house dance loop on top of material that really does not have any strong beats, let’s say world music.

I really like the Kaskade Sound Library Volume 1 that is available from various places, and I saw that Volume 2 is now also available. The only drawback I have with those loops is that they are 16-bit, but they still sound brilliant.

Anyway, with those loops, the drum ones, you could get anything to swing really well, even when doing DJ work.

Yes, I could build my own drum programming, but it takes a while, and it’s not always that interesting to do yet another classical house drum loop. Doing odd drum loops is another story.

Now, if you still want variation, use various settings in Ableton Live with these loops using Beat Repeat, especially with the variation parameter (use small values, though). Or the chop it using envelopes and other effects, to make them sound unique.

Maybe most of us agree that space is at premium when looking at files and clips inside Ableton Live. There’s not much screen estate. Having long names, even with meaning, is tough.

Here’s my current solution to encoding information to both files and clips.

The first part is the key and BPM, the second is artist, and the third is name.

Let’s look at each part, using d#25 Lucine-Slur.

d# means d sharp minor. If it would be major, I would use D, that way I don’t need to put in another word or character into the encoding. Thus, lower case is minor, and upper case major, and if it’s a sharp, I put a # after the key.

25 means 125 BPM, 90% of my tracks are in the 100-200 bpm range, so there’s no need to put in the lonely 1. If I have ballads, I just add a 0, as in c#090.

The next is the artist or group name, if it has more than one word, I use camel caps notation (any programmer would know this one, at least those doing Macintosh programming), hence MatthewDear.

After the dash I have the name of the track, again using camel caps notation if needed, to save the space, as in DrEiff.

All this is to save screen space, as you see in the screen shot I could see much more info in a small audio track view. Some might want to place the artist name first, for sorting purposes. The reason I like the key and bpm first is that I could quickly look at combinations in a listing.

Anyway, this is just a suggestion, you could build your own personal encoding systems — but try to stick to it, if you modify it along the way, you could easily get confused, and it’s tedious to recode hundreds of files/clips.

Recently I’ve been testing out Pandora with a strange new angle, using it to put together DJ like sessions. Now, it was not about mixing together tracks, rathering testing out if suggestions of tracks would make a cohesive listening experience using dance music recommendations.

As the DJs will get more and more tracks as part of their laptop archive, it will not be uncommon to get lost in let’s say 10,000 tracks, figuring out how each one sounds like, or what works together with each other.

This is where Pandora’s suggestion system could be a meta-level solution that might show up in future DJ browsing systems.

So far I’m both hopeful and not delighted. One problem with Pandora’s marking of similar tracks is that it’s not deeply granular with underground dance music. For example, for a channel where I tried to make it play Berlin style techno, Tiesto tracks show up from time to time.

But I’m hopeful about this kind of browsing technology, with better and better classification of tracks — if possible where the end user could adjust the tags for each track — this might work just fine. So if someone is interested in another music software project, here was the idea.

While driving yesterday in the car, and I had a compilation of concurrent Berlin techno music playing, my wife said that this kind of music was designed for partying and going out, but not for listening purposes.

It was just one opinion, but it might have some truth in it. A lot of electronic underground music produced today is really for the dance floor. Just the format of the intros an exit parts, with extended drum parts, are designed for DJs so they could easily mix in and out tracks.

Hmm. Actually I would prefer to compose for listening purposes, too, where dancing is one element. One way is to release multiple remixes of the same track, but in the long term that’s also a drag.

However, something that is more and more common with contemporary DJs is that they don’t really need the extended intros and exit parts. With CDJs, Ableton Live, Traktor, Torque et rest, it’s not so hard to set loop points and make your own intro and exit sections. It’s really for the vinyl and non-computer DJs where they need such parts. So in the long term there is really no need to make really extensive mixes.

So what I will actually do with my productions from now is to just make them more listening-centric, and not worry so much about the remixing live, as more and more DJs know how to do it, and I should not worry about that part. So I think this compromise from now will work out just fine with my production work.

Going through dance tracks, it’s sometimes very obvious if a DJ has produced a track or not. In one sentence: they make tracks that are easy to mix together with other tracks.

Going through details about this. Such tracks have a long intro with just plain drums or material with no strong melodies, making it easy to mix it in. Similarly, at the end there’s a long section where it’s easy to mix it out.

The format is very 16 or 8 bar centric, the breaks start at such parts, are multiples of 8 or 16 bars. Same with any other known sections in the track. It’s seldom the tracks have an odd thing thrown in, causing problems mixing other tracks in, or even for someone dancing getting back to the pace.

The track has long sections of repetitions, makes it possible for the DJ to use filter sweeps or other effects to put in their own touch to the track.

The track as no sudden surprises, such as abrupt volume changes, tempo changes, empty spaces with nothing suddenly happening, and so on.

The track is also of decent length, not just barely three minutes, rather at least five minutes or more. It has a smaller breakdown in the beginning, and a bigger breakdown around four minutes or so. The breakdowns are not mandatory, but it’s just one of those standard patterns most tracks have today.

As a nice bonus, such tracks have here and there a nice repetitive pattern, such as a singer pattern, that could be used as a acappellas somewhere else.

Now, producers could indeed do anything they want with their tracks. With Ableton Live and other similar tools the DJ could put together a DJ-friendly track, as well. If you really want both, make two versions, one is your artistic representation, and one is a remix that is suitable for DJ purposes. Also, if you never DJ:ed, but are producing dance music, it’s worth trying it out so you learn all this in a real-life scenario.