Issue link: https://digital.copcomm.com/i/1529890
38 W I N T E R 2 0 2 5 I C A S Q U A R T E R L Y with what sounds like only two mics, and the balances were awful. In addition, there is a section where the band sings one chorus and they are almost inaudible (and not very good). I had hoped to be able to make their voices sound better. Additionally, I wanted to get the rhythm section separate enough to be remixed, but this also failed. Unfortunately, with a band of 25 players, none of the apps could separate things well. My final attempt was the notoriously badly mixed and mastered Chicago track "Make Me Smile." The master always sounded to me like it was played from a cassette tape. Everything is muddy, especially the drums. Listen to that track back-to-back with any other Chicago track from their second album on and you can hear the immediate difference. The processing results were much better. I suspect part of it is that the stereo separation of instruments is quite good. Here are the different apps I tried, along with my reactions to them. Your results may vary significantly based on the songs you try. These reactions are based on the sources I chose. The separation images are for the song "Make Me Smile." See the separate table for specifications and features of each program. Lalal.AI This app came very highly recommended. It is a "pay as you go" app that you pay for by buying a bank of minutes. The minutes are not the length of the song, it is the length of all the splits that you do. This app does the processing online, so you need to be connected to the internet and upload your material (which may not sit well with studios). It can split up to nine stems, but you can only process one stem at a time. There is a functional iOS app as well. The envelopes were one of the biggest problems with the stems that it created. The horns have almost no attack on them. One split was synths, but with no synths, it put some of the vocals on the synth track. It correctly deduced that there are no strings. Lalal had a better attack on the bass than Acon did. The biggest issue for me was that the sum of the stems sounded nothing like the original mix. If your goal is a complete remix, that may be okay, but if it is for something relatively simple, I would expect too much time spent matching the original mix. I was let down given all the recommendations for this program. Acon Digital Remix Acon Digital Remix can be used differently than any of the other apps by applying it as an insert and creating sends and returns to give you splits in real time. It splits into five stems, although oddly, there is no guitar stem. Most of the guitar wound up on the piano stem. One of the nice things is that it allows you to adjust the sensitivity controls for each track. However, I found it to be very weak in terms of aliasing. The aliasing on the vocals alone made this unusable for me. One commonality on all of these is that the envelopes of individual instruments tend to disappear, as though using a badly set expander. In this case, the bass track had no attack at all and that made it sound completely unlike the bass in the mix. The track labeled "Other" should have included only the horns, but it had bleed from all tracks, including the vocals. iZotope RX 11 Music Rebalance Relative to some of the other programs, I found iZotope's RX Music Balance to be extremely slow at processing. It gives only four splits; Vox, Bass, Drums, and "Other," which includes