Music for Tablets: UVItouch Ravenscroft 275 Sampled Piano and Beathawk
An amazing sampled piano and a pretty fully featured music production powerhouse mark UVIs entry into the iOS platform.
by Warren Burt, Sept. 2017
When I first heard that Parisian sample-and-software-meisters UVI was entering the iOS market, I was excited. I wondered what they would do, especially with making elements of their amazing and deep sample libraries available to the world of tablets. The answer is here now in two very different apps, both of which are marked by UVIs high sense of quality and value.
Ravenscroft 275 Sampled Grand Piano
Ravenscroft 275 is the first one. It’s a sampled piano, weighing in at a hefty 837 MB. It’s for iPad, or iPhone, and although they say you need at least an iPhone 5 or an iPad 4 to run it, it also runs quite fine on my ancient iPhone 4S as well. Despite the size, it loads pretty quickly. On my iPad4, it takes about 13 seconds to fully load and be ready for play. It’s massively polyphonic, of course, receives MIDI, has an equalizer, a reverb, and an adjustable velocity curve as well as the ability to have either single or split keyboards, and has a sustain button, which functions just like a sustain pedal on a piano would. It the Global preferences you can adjust the reference frequency of A from 420 to 460 Hz. If you have multiple copies of the app on several devices, that means that you can easily set up different sets of 12-note tuning so that you could get, for example 24 or 36 or 48-tone tuning across all the devices, depending on how you set your reference A. It has ten factory presets, which have different settings of the EQ, reverb and velocity controls, and you can save your own presets as well.
The two questions one wants to ask about any sampled piano, of course, is how does it sound, and how responsive is it to play. The answer to the first is simply GREAT. It’s got a lovely timbre – the original piano was obviously a very fine one, and the sampling is up to UVIs usual excellence and high standards. If you want a piano sound on your iOS devices, you can’t do better than this. How responsive is it to play? On my iPad4 and my iPhone6S, the answer was very, very good. With a MIDI keyboard going into it, and a very low latency set to around 5ms (it will go lower), I couldn’t detect any delay from either when I pressed the keys or touched the keys on the screen. On the iPhone4S, older device that it is, there was some slight latency heard, but actually, not much more than with a less powerful desktop computer (bear in mind that an iPhone5 is the minimum they say is required, so my use of the iPhone4S is simply a test).
So we have here a great sounding sampled piano, and one that is very responsive and easy to play. At only $35.99 US (your price will vary based on your country), it’s expensive for an app, but very cheap for a sampled piano of this quality. Audiobus, IAA, AudioUnits are all supported. In Audiobus3, in fact, you can have multiple instances of it, each controlled with a different MIDI controller. Top marks to UVI for making what will probably become the standard in iOS sampled pianos.
Beathawk – Portable Music Production Studio
This is a very complex app, which I think is in the early stages of its development, but which is already a very powerful recording/sampling/performing environment. The central control unit is a four by four of pads. These pads can either play a single sample, or, if you press the “Pitch” control when a pad is selected, a keyboard shows up, on which you can play patterns, which will be remembered and stored as a set of sixteen patterns. These patterns can be polyphonic – that is, if you have a different sample on each pad, and you play on the internal keyboard (or use an external source of MIDI) to play the samples on each pad, all these can all be stored in one pattern. These polyphonic patterns can then be sequenced in a “Song” page. Also on the Song page, you can specify which of the sixteen pads will actually play in any particular instance of the pattern. The song can then be exported either as a wav file, a MIDI file, or stems for use in a DAW, to AudioCopy, iCloud, Ableton Live, and/or YouTube. The possibilities for live performance in this app are numerous. You can play the pads like a drum machine. You can play the keys in the “Pitch” subpage, either on their own or while a pattern is playing. You can play on the Song page live, selecting patterns to play in real time. Etcetera. It’s a very versatile app, which, while it is a bit complex (but its clear manual demystifies things pretty quickly), is nonetheless pretty easy to learn, and I found I was making interesting pieces with it rather quickly.
This shows the basic Pad setup of the piece and the central control page.
This shows a pad with its keyboard controller opened.
This shows the Pattern page, where you can store up to 16 sixteen voiced patterns.
This shows the Step Editor page for one pad only with its pitch patterns shown.
This shows the Song page, where you can sequence your patterns.
This shows the Sample Browser page, where you can assign sample sets to individual pads.
The samples can be your own, or you can use some of the supplied factory content, or you can buy sample sets from UVI. These sample sets are inexpensive ($5-$20 each, on average), and are drawn from UVIs extensive sample libraries. A number of them, for example, are drawn from UVI’s wonderful World Suite. Each set usually has a multi-sampled keyboard instrument of samples, some loops, some phrases and some “elements” which can be used to set up custom drum (or other instrument) kits. The samples on these, although limited, are again of very high quality. Each pad has a number of controls which can change its sound significantly. There is an Edit box, with controls for Gain, Pan, Pitch (curiously, only in semitone steps, more on this later), HP and LP filters, Reverb and Delay. These settings only apply to the currently chosen pad. So you can have the same sample, or sample set, on different pads, treated differently. There’s also an ADSR box, to change the envelope for a particular pad. There’s also a Record box, which can record sounds coming into the Mic, or if you’re in Audiobus3, from another app (this needs work – again, more later). And finally, there is a Sample box, where you can edit the start and end points of your recorded sample, and also stretch or compress it from 50% to 200%. The time stretching is very good – there’s not a lot of degradation in the signal even with extreme (50%) time stretches. And again, if a sample is loaded onto different pads, you can have a different time stretch applied to each.
So, a very versatile app. Did I find any limitations? At this point, yes. The MOST glaring omission, it seems to me, is in the Edit box for each pad, there is a “Pitch” control which will adjust the pitch of the samples on it in semitones, but there is no “Fine Tune” control, where you could adjust pitch in, for example, Cents (1/100th of a semitone), say from -100 to +100 cents. Not having a fine tune control means that if you, for example, record an instrument that is slightly out of tune, there is no way to adjust the pitch to have it play in tune with the internal UVI samples. Nor is there any way to get “beating” effects between pads with similar samples on them. That strikes me as a very curious limitation, and it is one that I hope the designers will remedy as soon as they can. And of course, no fine tuning means no possibility of microtuning between the pads, which again, is a real limitation.
Also, at the moment, only the UVI samples can play in a looping fashion. Your own recorded samples can’t be looped. I’ve been informed by UVI that this will be remedied in a future update. I hope so, because without looping, BeatHawk is at a real disadvantage when compared with platforms such as Ableton Live, or even Loopy.
I’ve also noticed that when working with Audiobus3, and feeding an external app’s sound into BeatHawk for sampling, I’m also getting input from the iPad’s microphone – sometimes. I think this needs to be looked into.
The sample sets supplied, are, of course, rigidly in twelve-tone equal temperament. No microtuning or even pitch bend allowed here. I asked technical support at UVI if microtuning using, for example, Scala files, as in the UVI Falcon, would be possible, and they said theoretically yes, but they didn’t feel that the demand was there for such an “esoteric” feature. So microtonalists, if you like BeatHawk, feel free to start bombarding UVI with requests for this feature to be implemented.
Rather than give you more technical details about the app, at this point, I think I’ll depart from Soundbytes usual protocol, and describe for you the processes I used to make a piece with Dhalang MG, BeatHawk, and AudioShare late in July. It’s also a diary of where and how I was composing the piece – mostly in my spare time on my 107 km (67 miles) commute from Daylesford, Victoria, where I live, to Box Hill Institute, Melbourne, where I work, and back. This commute has me taking buses and trains, and when I’m in the City, trams as well. As you read this, you’ll see how I used BeatHawk to make a microtonal piece with these apps, one that I’m quite happy with. To hear the piece, go to http://www.warrenburt.com/storage/Dhalang%20melodies_Beathawk_Sequenced.mp3 and download it. To see the diary/article, see below.
In conclusion, would I recommend BeatHawk? Absolutely. At this stage it has some curious limitations – like the lack of fine tuning on sample pads, and the lack of looping on your own recorded samples – hopefully those will be fixed in a future update – but on the whole, it’s a very lovely system – well thought out, and with lots of possibilities for both recorded and live performance. As a sketchpad, it’s superb, and as a performance platform, it’s full of potential. AND it works well with other apps, being Audiobus3, IAA, and AudioUnits capable. It’s also really cheap at $9.99 US (where they hope to make their money, of course, is in the sample sets, and these, at least the ones I’ve used, have been of very high quality). A lot of fun, and you can get professional sounding results from it.
So welcome UVI, to the realms of tablet-based music making. Both of their debut apps are well worth having, and both offer lots of potential and great value for money.
How I’m making a microtonal poly-melodic piece with BeatHawk:
- On bus from Daylesford to Woodend, on 26/7/2017, conceive of a piece with 8 melodies as main material, each in a different ET: 20, 19, 18, 17, 16, 15, 14, and 13. Make these melodies with Dhalang MG and record them into AudioShare.
- On train from Woodend to Southern Cross, Melbourne, on 26/7/2017, make melodies with Dhalang MG: First, set up the tuning in the Tuning Editor Page. Got to the Tuning Page and set the scale to be all available pitches in the scale. Then go to the Additive Synthesis page. Make timbres of 2 harmonics each. Tune the two partials to very tiny deviations around 2.00 each. Ie 2.01 and 2.00; 1.97 and 1.99 etc. Use a different timbre for each melody – so 2 each of sinesum, triangle, saw, and square. Set filter settings and filter Q differently for each melody. Third, go to the Particle Page and set up a Particle system to generate melodies – one generator and two surfaces – each surface set to a different pitch range. Change position of the generator and surfaces, and the gravity rate for each melody. Also pitch ranges and overall tempo as appropriate. Fourth, record about 10-20 seconds of each melody with AudioShare. Once recorded, normalize and trim so that they all start with sound and are all a good level. Rename them all “Dhalang 20et” “Dhalang 19et” etc. Send to trash all the intermediate waveforms used in making these.
- In Theobroma Chocolate Lounge in Southern Cross Station, while having a dark iced chocolate (no cream), copy all the eight melodies to AudioCopy. Erase the melodies from AudioShare. Go to Trash in AudioShare and get rid of these waveforms and all intermediate waveforms. Open BeatHawk and import the eight melody wav files from AudioShare. Then erase the eight waveforms from AudioShare. In BeatHawk assign melodies 1-8 (starting with 13et) to pads 1-8. Then assign melodies 1-8 to pads 9-16. For each pad adjust Release in ADSR to a long decay, adjust pitch for pads 9-12 to -5 semitones and pitch for pads 13-16 to +5 semitones. Adjust the panning for each pad from 70 Left for Pad 1 to 70 Right for Pad 16 in increments of about 10 steps for each.
- On train from Flinders St Station to Box Hill, play with the volume of each pad (make most about -6dB), and then notice that if you push the “Pitch” button, the keyboard comes up to allow you to transpose the samples to the pitches of 12ET. Playing one pad at a time, record individual tracks into one of the Patterns (set to a length of 16 bars with the tempo at 20 and the Quantize OFF), alternating just holding the pad which allows the whole melody to be played, and then to the Pitch keyboard, and play some short riffs with the melody as a sample transposed to the pitches of 12ET. Then back to the solo Pad, and then back to the melodies, etc. After each take, Mute all previous pads and record a similar process on the next Pad. On this train ride, I was able to record 2 tracks, with pads 1 and 2, of sequences using the 13et and 14et melody.
- At Box Hill at work, type up this description of what has happened so far.
- Future plan – record the other 14 tracks. At tempo 20 with 16 bars, the sequence is about 3:12 long. When you have a Pattern with 14 tracks in it, maybe you can have a number of playing of this pattern with different tracks being on and off for each playing of the pattern. Or you might want to make a couple of more patterns. In any case, when you finish the piece, and you’ve exported it as a wav file, then two-finger touch the wav file and export it to AudioCopy so it can be imported into AudioShare and from there, exported into the “real world.” (One idea is to do the microtonal piece from start to finish “in the box (iPad)” and only export the final result to the real world.
- On the train back into Melbourne from Box Hill, I recorded tracks 3-8, all of which used the same process as described above for tracks 1-2. This in the evening of 26/7/17.
- Then in the morning of 27/7/17, on the train from Melbourne to Box Hill, recorded tracks 9-16. For each of these, I decided to only play single instances of the samples, one at a time, on each track. For transpositions I only used the following D2 F2 G2 Bb2 C3 D3 F3 G3 Bb3 in a different order on each of the 8 tracks.
- In the afternoon of the 27th, at work, I prepared a “song” score of the sixteen-track texture 1 happening twice. Once with tracks 1-8 and the second time with tracks 9-16. I exported it, and then exported the exported wav file from Beathawk to AudioCopy and then from AudioCopy to AudioShare. I then listened to it. It was OK, but the shift from left-leaning textures to right-leaning textures, when you went from the first iteration of the sixteen-track texture (with tracks 1-8, then 9-16) was a bit jarring. I then added two more iterations of texture 1, each playing for only the first eight measures of the texture. The first used tracks 1-4 and 13-16; the second used tracks 5-12. I then exported this, and found that although at the end of the final playing of texture 1, there were some samples that decayed beyond the end of the measure, when you exported the track, the sound abruptly cut off at the end of the measure of that track.
- In the car, driving back to Daylesford (my wife brought in the car later in the day, and we drove home together), I thought that if I put a dummy blank texture at the end I could get the “overhang” of the decaying samples heard. Then it occurred to me that I could indeed make a “coda.” Using the “piano-roll” editor, I simply made a texture 4 measures long, and for the first two measures, just had a sustaining C3 in each of textures 1-8. This gave me all eight melodies playing simultaneously. This sounded good, so I glued that to the end of the timeline, and then exported the sequence, which I then exported (using two-finger touch on the sample) to AudioCopy, and thence from AudioCopy to AudioShare. The silence at the end of the Coda texture did indeed solve the overhang problem, and allowed all the waveforms to decay gracefully.
- On the morning of the 28th, at home, using iTunes, I transferred the final mix from AudioShare to my desktop PC, and looked at, and listened to the final mix using Audacity and some good loudspeakers. The result sounded fine. The piece was now finished. All that remains to do is to clean up all the intermediary waveforms in AudioShare and Beathawk.
- Again, if you want to hear the piece, download it, free, from here: http://www.warrenburt.com/storage/Dhalang%20melodies_Beathawk_Sequenced.mp3
28/7/16 Daylesford, VIC Warren Burt