IMHO Cloud processing would be a hindrance, have you tried to use web based AI audio creating or separation tools? Slow AF, you know as well as I DJs have zero patience, let alone wanting STEMS in real time while performing. Look at Google’s AI cloud based processing, video or audio is done in hours.
Interesting concept. I thought the industry was pushing AI processing to the edge due to heavy cost of infrastructure, energy & cooling. This kinda goes against that.
As with any technology there are advantages and disadvantages.
Let’s start here by listing the main advantages that cloud computing could bring
The main interest lies in the fact that complex task calculations that consume CPU/GPU resources (typically stem processing) are done remotely in powerful cloud servers that can be associated with AI processing technology.
As all equipment using EngineOS has a WiFi connection, this would allow these complex tasks to be carried out remotely without having to place additional demands on hardware resources. So any device, even if it is not very powerful, can benefit from the computing power of remote cloud servers.
This could also benefit other equipment from the InMusic group based on the same type of hardware, typically for the calculation of stems on recent AKAI MPCs in standalone mode which are also equipped with a wifi connection. It would therefore be interesting for Inmusic to pool cloud servers for its different brands.
However, the disadvantages are numerous.
The cost of renting the servers would mean that InMusic would have to pass these costs on to the end user in the form of a monthly subscription similar to what you get when you subscribe to the Geforce Now service or Xbox Game pass.
So you’d end up with a new subscription to pay each month on top of your Tidal subscription and its DJ extension for those who want to use the streaming service.
Another drawback is the dependence on the internet connection. As with streaming services, cloud computing by definition cannot be used offline. You can use your mobile’s connection sharing, but if you’re in a place where the network is poor or non-existent, you can’t use it.
Obviously this does not apply to fixed installations in clubs or at home that are constantly connected to the wifi.
The speed of the connection should therefore be an important factor, because you need to consider the time it takes to send the track to the server + the time it takes to receive the separate stems + the time it takes for the server to process them.
Obviously this would require in-depth market research for InMusic but I’m not sure that many users would be willing to pay between $10 and $15 just to be able to access cloud computing with the possibility of not being able to use it everywhere and in all situations.
In conclusion, it seems more likely to me that InMusic will not take this path for only a few years of operation because the next generation will probably have much more efficient integrated chips capable of calculating the stems in hardware.
I would not be surprised if this new generation is already in the study or prototype phase in the inMusic premises because the current hardware is over its 7 years of existence if we start from the principle that the sc5000 , the first product in the prime range, was unveiled in January 2017.
On a strictly hardware level, the RK3288 has now been around for 10 years, which is a very long time.
Truth be told, it could be more or less around the prices of current models because Inmusic wouldn’t have as much research and development as when they created the prime line from scratch.
For example they can use the base of the SC6000 chassis + buttons + screens which are of good quality. The layout is already close to perfection. They could simply update the onboard hardware with a more efficient chipset, for example a Mediatek Dimensity 9300+. The only thing they would have to redesign is the motherboard, to accommodate the new CPU and larger RAM.
Not to mention that modern CPUs are engraved much thinner (5nm) than CPUs from 10 years ago. So at the production level this means that you can produce more CPUs on a single silicon wafer which also considerably reduces production costs.
They could take the Rane Performer platter for the M version and simply adapt it.
Call me old fashioned, but I do not and cannot rely on stable internet connections while DJing. Some of my gigs are in dark concrete basements, some are somewhere in the fields.
I would feel more for preparing stems at home and exporting them to your USB. But then again, you will lose the “spur of the moment” feel which those Zplane algorithms offer. I was never a prepare-at-home kind of guy…
Meh, still sounds like a lot of hassle for something I would seldomly use. Or, never say never, maybe I would use it if it’s available on the fly without wifi connection…
Better integrated chipsets for stem processing can’t beat the real computer/laptop CPU. Most modern AMD and Intel cpus especially for gaming and updated DJ softwares are powerful nowadays in multi-task operations of x64 bit OS programs.
High-end chips would be too expensive and not good at performance, we need at least specific cpu architecture for EngineDJ hardwares to run smoothly at high quality and performance the stems with other staff as well. It may be more overpriced but it worth if people don’t want spending money monthly or yearly for Cloud stems as another paid subscription.
I don’t believe that specific chipsets are powerful enough to handle these features, integrated on board 8-16 cores CPU should do that with fast cooling inside.
But the problem is who replace thermal paste on it without loosing product guarantee? Lol
