Audiobus: Use your music apps together.
What is Audiobus? — Audiobus is an award-winning music app for iPhone and iPad which lets you use your other music apps together. Chain effects on your favourite synth, run the output of apps or Audio Units into an app like GarageBand or Loopy, or select a different audio interface output for each app. Route MIDI between apps — drive a synth from a MIDI sequencer, or add an arpeggiator to your MIDI keyboard — or sync with your external MIDI gear. And control your entire setup from a MIDI controller.
Download on the App StoreAudiobus is the app that makes the rest of your setup better.
Comments
One thing really isn't clear with me: up until a certain generation of devices you could choose any sample rate you wanted, even without an interface connected. For example on my 2017 iPad Pro there is no restriction, same goes for older phones.
From what I can tell, the restriction only applies to devices that don't have a headphone socket. The question that I just can't figure out is why?
Why have Apple locked those devices to 48k? Is there any technical reason or is it just to add insult to injury, by first taking away the headphone jack and then locking the sample rate to really piss us off?
I then created a 96k 32-bit float project in MultiTrack DAW, drag-and-dropped the audacity file in a track and mixed down / exported again - file attached.
It is still a 96k file and contains all frequencies!
So, Auria Pro and MultiTrack DAW may be the only DAWs that allow for this ...
I will try with nanostudio 2
Yes, this is what I was wondering about and wether the whole thing is even required at all.
Not true, my iPad Air 3 has a headphone jack and is locked at 48k
Just Apple?
Ah OK, so it's not directly related to the headphone jack, but it seems to have happened at around the same time as the jack was removed in the phones and later the iPads.
I created a nanostudio 2 project, added a Obsidian track, loaded the file as patch and mixed down the project to 96k and 32-bit float.
But I got "output file is silent" ... this is surely due to me doing something wrong.
Most probably, I need to play the sample somehow ...
The difference in the 24k frequency is most likely just down to the binning that happens because of the window size they are using to generate the spectra.
-90dB is pretty much zero volume, so unless there is some technical thing you are doing like looking for noise floor or something like that, it's easier to make your graphs assume a min of -90dB. A 32-bit float is effectively around the bit depth of a 21-bit or 24-bit int (or fixed point depending on how you are doing things). They have the ability to deal with a higher volume level but they don't really get more resolution for the minimum. They are also much nicer to work with algorithm wise.
Thanks for the explanation, @NeonSilicon
nanostudio 2 did render the whole thing as stereo file, not sure how to change that ... file attached.
I was quite helpless to create this in nanostudio 2
But anyways, it also contains all the spikes - even if they look totally different.
One thing to note about synths and effects using upsampled audio is that many don't and there isn't any reason that they should. You really only need to upsample if you are doing any non-linear processing that can add higher harmonics that would cause aliasing. For a synth, using a higher sample rate and downsampling is most likely not going to be the best path. You are almost always going to be better off using a band limited algorithm that eliminates any aliasing before it can happen.
I don't know the answer to why they have been locked down, but it could be something as simple as they went to a fixed sample rate DAC on the newer iPads for the internal speakers and so they locked the sample rate to the DAC so there wouldn't be any question of needing to do SRC on the output or the mics.
I finally tried Cubasis 3, it cannot change the project to 96k and instead offers to convert the file to 48k, which I declined.
Then I did create a mixdown - file attached.
Cubasis also created a stereo file, which was my mistake, I think - but anyways, it could only create a 48k file!
So, as expected, Cubasis cannot be used to handle content of a different sample rate properly - MultiTrack DAW and Auria Pro can, and nanostudio 2 too ... even if it is a different situation without audio tracks.
Yeah, but it also contains a spike it shouldn't contain -- the one at 8 kHz. There's your aliasing
I'd give it a try in Cubasis 3, but I don't have that one. I would be surprised though if they did any sort of SRC to mix down a file to a sample rate that was different from what the thread was actually running at. I'd expect that they would use the manual rendering thread I mentioned in a link above or they'd use whatever they are doing in their own internal threading to run at the sample rate the project was set to.
Ohhhhh
As I was so hasty in creating this, i did not even notice!
Many thanks.
So, our list of "capable" DAWs is reduced to Auria Pro and MultiTrack DAW.
I will use Auria Pro for mixing and mastering from now on
I should do everything in it ...
Please see above.
For some years, I did buy just everything ... so I have many Apps I rarely use. But I am in the process of stripping down ;-)
One thing to remember is that in iOS and to a large degree in macOS, the DAW doesn't have control of the audio thread. On iOS in particular the DAW doesn't really control the thread that the AU's are processing on. The plugins are totally separate applications and everything is being handed off to them using IPC. For the most part, the DAW is only asking the OS to do something and the OS gets to say yes or no and then it orchestrates the whole rendering chain. That's why Apple has the API in AVAudioSession and AVAudioEngine for doing the manual rendering.
Oops! I hadn't seen your Cubasis results yet. So, it looks like they downsample the input file and then run at the sample rate the audio thread is set to.
Cubasis asked wether to downsample the file to 48k when I imported it, but I declined the conversion!
So it may probably internally still be 96k ... but as you cannot export as 96k, it get's cut out at mixdown.
I'm assuming that they left the file at 96k but did SRC on the audio from the file at input and ran the project at 48k. I'd have to run a debug session in Cubasis to confirm that though.
Sorry, I should have quoted...
which was the reason for slightly extending the perspective, but just ignore it in this context.
On the other hand it is related to the „locking to samplerate“ by Apple, as they have to consider a general audience.
Afaik their audio codec chips are custom designed with no general specs available.
These chips vary with hardware and there maybe whatever consequences affecting some and some not.
Frankly said: Apple sacrificed a feature that only 1 out of 1000 customers even knows about, and only 1 out of 100 from this group considers it relevant.
Apple knows that almost every member of the latter group owns an external audio interface, so they just spare the effort.
They are neither stupid nor naive ... and only in it for the money anyway...
Now back to tech stuff .
I just wish 1 out of 1 devs would fix their plugin to work at both 44 and 48 (or maybe it is the hosts faults?) ... le sigh
I tested sample rates on my iPad Pro 12.9 2nd gen. and it seems that it can use either 44.1k or 48k, but no other rates - from what was possible in AUM and Cubasis 3.
So, it is not fully locked, but still restricted.
But at least, it can handle both CD and DVD / streaming resolutions - great!
All still without an interface attached.
Which plugin shows problems? And in which host?
Multitrack DAW is a really good example:
one of the first apps considered a „DAW“, supported from the first to the most recent iDevice on both phone and tablet platform.
When I checked the 96khz capability on iPad-1, sample rates above 48khz were marked red as „unsuggested“, yet still usable (leaving the decision up to the user).
Btw it’s mix engine is not standard 32bit float, but something more sophisticated.
(I don‘t remember the specs, but it once was a topic in their forum)
MTD was multichannel input from day one on, supporting almost any interface on the market.
I tried to use auGEN X within an Auria Pro project with 96k to create sound at 32k, but it refused and restricted the sound to 20.48k.... even as the session was shown as 96k:
Shouldn't that be possible, @auDSPr?
Any other plugin or App that can produce sound higher than 22.5kHz?
I seek out those mentioned above
This is esp. strange as the iPad is locked to 48k, which means that audio could be generated at up to 24k - if the apparent project rate of 96k somehow does not apply, @auDSPr
Have it do a sawtooth at the highest frequency it can and see what it produces.
I would expect that some synths don't produce any waveforms with content over about 20kHz, especially those that are using some form of band limited synthesis.
I'd expect most plugins work fine at any sample rate 44.1 and above.
BTW, I had issues in some of my plugins running at sample rates below 44.1kHz. They would run there, but they weren't stable. I didn't expect any DAW to be running anything down there, but discovered that GarageBand does have tracks, some samples and smart instruments that run at 22.05kHz on iOS.
Egoist in NS2 has problems. I should give it a more clinical try again in BM3 and Cubasis.
The VST has problems too with 44/48 so I just imagined it is a Sugar Bytes thing.
I have DrumComputer, Factory, Aparillo, Unique, Cyclop, Thesys and Turnado, but not Egoist :-D
May try later with one of them.