normally I just mixdown offline because its quicker. However at the moment I am finishing of mixing an album and the projects are quite large, stretching Cubase to its limits. I was going to mix down using realtime to ensure all the plugins did what they were supposed to but unfortunately I keep getting cpu overloads
I hear people say offline mixdown compared to realtime mixdown is no difference. I have done some tests, and my ears are telling me different…but I could have gone completely crazy also.
I run all the mixes to a group channel and then to the master fader so could I set up a stereo audio track and send the master group to that and then just record the mix to an audio file? This would technically be the same as mixdown wouldnt it?
or am I just going mad and in fact offline export is absoloutely fine!
Yes, you can route all tracks into one Group, which will be your Master, in fact. Then, you can create new Audio track, and set this group as input. And you can record your whole mix into this new Audio track.
In my opinion, and experience, for my ears, offline mixdown is exactly the same, as realtime mixdown.
I have experienced subtile sound differences at high cpu load (with overloads) running a “live-mix” compared to ofline/rendered mixes, and that is exactly what you were talking about I think? When you render offline, the audio is probably different/cleaner because there are no cpu overloads in the stream as the computer just takes its time to calculate the stream, obviously this is impossible when re-recording your output 'live" while running out of cpu -power. Maybe that could explain the differences you hear?
If you can’t run your mix without cpu spikes at the highest possible buffersetting you will therefor never get adequate results, as your mix wil contain cpu-spikes (live) or sound a bit different (rendered).
freezing some tracks to free your cpu might be the solution here…
For me there is no difference in re-recording a live mix, or exporting real time, compared to ofline rendering as long as the cpu meter stays in the safe zone.
And by the way, if you route everything to a group, your effect returns are not included in that mix(!), as they can only be routed to the main mix (correct me if I am wrong here). You should always export or re-record the main mix to include your effect returns…
A trick can be to use group channels as effect returns (simply insert an effect in an insert-slot of an “(effect)-group-channel” and use the corresponding send in all other channels to “send” to it). Now it is possible to route your effect returns at 0db to another group…(your “mix” group) and thus include them in the output of your “mix-group”.
Way back I used to have an SPDIf-loop cable in my system (connect SPDIf in and out with a cable, be sure to use an optical connection or cut the ground on one side of the cable, or you will get phase problems or ground loops).
Send your main mix out to spdif-out, and re-record the spdif-in on an audiochannel (mute it!). Also some soundcards have options to create a loop, that is what I do now (RME card) when I want to re-record an output.