A word of warning re: very large buffers at mixdown - I have read this can play havoc with very tight automation. So there’s a happy medium to be struck on buffer size at mixdown (I can’t vouch for it, but have noticed less automation errors with buffers not maxed out).
(Real time - for example if your mixdown includes a hardware synth output.)
if you work fully ITB realtime is not needed.
Do you think this will work with the “hard drive” error that randomly pops up? I don’t remember the exact error message but I can duplicate it every time with one of my projects.
I really don’t think it is a hard drive error, I have an SSD drive with 80% free & a regular hard drive with 90% free, they both are operating very good. As I said before some songs do it every time, others will work after several attempts and others work every time.
BTW I’m using Windows 10 and Cubase 9 pro, both updated with with the latest. The only 3rd party plugin used is Melodyne all the rest are Cubase plugins.
Thanks, Raphie, now I get it. I’ve been working exclusively ITB, but need to put my rack synths back into the mix, and, I see how outboard processors might require real-time export as well. Take care.
Well for me, I have to disagree. Whenever I have done an offline export I feel I’m not hearing the same song I heard through my monitors when doing a playback in Cubase. Of course I’m listening to the export through the same monitors. When the export is done realtime (when it works) it sounds exactly the same as Cubase.
I admit I’m a rookie, only have been working with Cubase for about a year. Maybe there is something else influencing the offline export?
I don’t want to divert the conversation away from solving the error issue, just giving my ears opinion. 
Do you use the control room? Sounds like doubled outputs to me.
No I don’t use control room. Haven’t expanded my horizon that much 
. I’m not sure what you mean by " double outputs"?
The offline export and realtime export definitely sound different to me and others that listen to the progression of my projects.
I’m sorry I can’t really pinpoint what the difference is? Just that when I do a realtime export it sounds just like what I hear in Cubase doing a playback, offline sounds different. I’m not saying worse, just different. Obviously when I complete a mix and like what I’m hearing that is what I’d like my export to sound like.
Could be just me???
Quick way to tell - render both ways, then drag the two .wav audio files (not mp3, etc.) into a third project and do a quick blind test, toggling one then the other. (I think I remember you said you’ve been using Cubase for only a year or so, so maybe this little trick is new to you - mute one of the tracks, “S” solo the other, then select both and you can blindly toggle between them by hitting the M or S keys).
If you consistently find they do sound different to you … rest assured, they’re not supposed to and there will be a way to sort through that!
Wow thank you for the detailed description!!! Too often people in the know post what to do but not how to do it, so thank you again.
I will try this for sure.
Off to the studio!
Absolutely wrong, buffers also have an impact on offline mixdowns.
Easy to test, try setting the buffer as low as possible do a offline mixdown ( notice the time it takes to complete)
Do the same thing with buffers set to maximum possible. (Notice how fast the mixdown now completes)
no difference here? 64 or 1024, on a 7min arrangement doesn’t matter at all here?
(I7 6800, 32gb ddr4 3000 and R400 PCIe M2 SSD)
o___O
…Sorry, what?
That is strange, I just checked again if something was changed in C9, or if the ASIO-Guard could be the reason.
I used a small 4min 19 audio + 1 VSTi and a stopwatch
Rendertime 64 sample buffer = 1m 55s
Rendertime 4096 sample buffer = 0m 55s
Got the same numbers using ASIO-Guard at max.
On bigger projects the difference gets bigger.
There was a discussion about it in the C8 forum, I think there was even a feature request made allowing for some huge buffersize to be set.
My guess is that any rendering has to use the ASIO part of the driver to run the data through to keep everything sample accurate.
Thanks for posting your results.
My card is an RME MadiFX PCIe, not sure if that matters?
The Control Room is amazingly useful, and after a very flat learning curve is actually no harder to use than anything else in Cubase!
Here are some really nice references to get you started in the Control Room when you’re ready!
The most helpful functions in Control Room IMO:
The Listen Bus. It is like a solo button, but allows you to hear the track with the rest of the mix volumed down by as much as you like. Wow!
The Loudness Meter. This tells you exactly how loud your track is, using more meaningful measures than simply the “peak” loudness of the faders. It is equivalent to plugins that can cost more than $100 in other DAWs.
Stereo to Mono switch. This could be done without the Control Room, but it would use up a plug-in slot. I used this all the time until I got an interface worth a hardware mono button.
When you export, regardless of it being real time or not, how do you check the export? Do you bring the file into Cubase, do you listen to it with a different player? If all the instruments, audio tracks and processing is being done internally to the DAW, no external equipment in use, then there should be no difference between Real Time or Normal Export, as Raphie explained. If someone’s hearing a difference my guess is that they turned up the volume or somehow changed the listening conditions but didn’t do an actual A/B blind test, as someone suggested in the this thread. I’ve used both Real Time and normal exporting and don’t hear a difference in the final output file. I’ve not done nulling tests on this, but that would be a good first check.
Anyway, maybe someone will do some testing – or it’s already been done. I’ll have to look around the threads on this. Take care for now.
Raphie, I was thinking that the way to verify this would be with a standard null test. Yes? You know, we’d take the same project and export first in real time, then export ITB, bring both in to a new project, line up the tracks and invert the phase of one to see if they null. I’ve not tried such a test myself but it would be a good way to verify if there is or is not a difference. I suspect there’s likely no difference – the tracks would null – and so if one is working ITB there’s no reason to do real-time exporting of mixes. However, again, I’ve not run such a test. Your thoughts?
Seems valid, though the only thing realtime does is time crunching to the real time speed so external audio can be merged.
The ITB stuff doesn’t take another route, it’s just timed to line up with realtime audio channels if needed
I see. That’s good know. Maybe someone will find time to conduct a verifiable, duplicatable test of this.
FINALLY after few hours of frustration I FOUND THE SOLUTION!!!
put CTRL+ALT+DEL and close all tasks and all APPS and Background processes that suck your memoru,cpu and disk.
and That’s it guys!! enjoy