Need for and technical posible with a "light weight mode" or a Mixbus light version?

[Ben@Harrison]

@dagh: I think you will find that your startup time improves noticeably if you remove ASIO4ALL from your system. Both Mixbus and ASIO4ALL will try to scan your soundcards, and in some cases this can cause a significantly longer delay. I've noticed this problem specifically with Focusrite usb devices. Mixbus can use MME devices directly, so there is no benefit in using ASIO4ALL to convert the MME devices to ASIO.

If you are seeing 100% DSP usage immediately on a small session, then this indicates your Windows system is not tuned for audio performance. (Macs work pretty well out-of-the box, but windows machines typically require tweaking). You should google for "windows audio performance" to try and determine why your system shows such high DSP% indication with a blank session.

Best,
-Ben

[Matt]

Aside from optimizing your system, which I would make sure is done first. You could also do all your tracking in ardour or waves live. They use the same session format as mixbus. And then once you're ready to mix, open the session in mixbus with your buffers set higher. Best of both worlds.

[dagh]

@Ben: Useful to know! Nevertheless I don't think ASIO4ALL made big differense on my two PCs. I had done most of the Windows tweeking long time ago on the mentioned PC, but useful to be reminded of it. Removed a JACK that I had not made working, and start up time was reduced with one minute on the 32-bits 32C version! A new added AV Linux install on the same PC did not give good result either, because of the two core processor I guess. So we let my PC be like this. My two core MacBook Pro managed it better, but would have been even better without the mixbusses.

@Matt: I did not know about the "Tracks-Live" from Waves, and found out there is also a "RADAR Session" from iZ.

Well, what shall I say to what I am been doing lately?
Putting files in a mix of to many programs on each of the old and not powerful 1-2 core computers, that then are not working properly:
Ardour, MB, MB32C-3 32-bits, MB32-4 with changing audio setups - oh, what a mess!
...and in AVLinux you must also change name of the song, when you go from tracking mode to mixing mode, to come into the audio/MIDI setup to change buffer size /latency.

I guess, the best had been to use one - 1 - program most of the time, the best program, Mixbus 32C (?)

It's like buying a car:
You choose to buy a truck (Harrison) for driving the big things, and do a swap between this and two seat sports car (the Reaper DAW) every time you are going for a fast and light trip. But you don't need more than ONE normal car (Ardour, Sonar...) and a convertible (Harrison with "convertible" mixbusses) would have done a splended job!

But, I think I have a clue now - from a sales view:
It may be smart to divide the production into two production lines: One line for a good and popular DAW (Ardour), and another for good console saturation sound and plugins, and a win-win salution for bouth who are popular and demanded, bouth benefit from each another. So I guess, no "convertibel" light weight Mixbus vertion from Harrison ;-)

And I... I am gonna use a digital mixer next time playing my dance music with my keyboard...

[Matt]

It definitely depends on your workflow. When I work, I tend to focus on tracking first. And don't really get too in depth mixing, until I'm sure I have all the tracks I need. Sure sometimes you get an idea for another thing or two. But you can work around that also. Anyway once I'm ready to mix, I change the buffers and such and start mixing.

If you do run into a case where you need another track. You can print a quick mix of where you are. Import that into another daw/session, and record the extra tracks. And then import the new track back into the mixing session. It's not that much extra work. Especially once you've done it a couple of times.

That said, I'm not sure your situation. But you might be on the lookout for a used workstation pc. I've been looking at Dell T3600 and T5600 along with the HP z620's. Most of the inexpensive ones have quad core E5 Xeons. And you can pick up 6-12 core CPUs for decent prices. Would be quite a bit more powerful than the system you have now. Fwiw I searched for your CPU on passmark. It has a score of 958. The E5-1620 that in a lot of the systems I've seen on eBay, has a score of 9,090. So roughly 10 times as much power. Just some food for thought.

[johne53]

If I open only a new file, whithout any new channels/tracks, Mixbus32C-4 (64-bit) and Mixbus32-3 32bit uses 50% and 60% of the DSP, after a while when it is stable. This is on my cheap Windows computer.

Do you know what make of CPU is in your computer? Intel CPUs have always been good at calculating the DSP load but for a long time (even back as far as Mixbus-2) AMD processors used to give errors in the calculation. In fact they still do AFAIK. Early in Mixbus-3, Harrison added some code to help minimize the errors but maybe there are still some edge cases where the corrections don't work.

OTOH if you're already using an Intel CPU, the most likely explanation is that your audio buffer size is too small (as others have already mentioned). In a previous post you said you're using a size of 256. On a cheap computer (not optimised for audio work) you probably need to use 1024 or even 2048.

[dagh]

Matt and also now johne53,
thank you for your good intentions.

Can I say it this way?

I want to achive product improvement for customers - to avoid walkarounds
You want to achive walkarounds to me - that avoids product improvement

I want to achive A - to avoid B. You want to achive B - that avoids A. Understand?

Although I lack lots of mixing experience, I know about and have used some walkarounds already, at home.

For the live performance situations, walkaround with PA in tracking mode and then playback in mixing mode does not fit in.
And just grab a non valuable PC for the scene, that I don't have to care about, would feel good after one of my PCs has already been destroyed by someone.

johne53, I have Intel processors

[johne53]

Open Window->Preferences->General and experiment with the topmost optimization (DSP CPU Utilization). For my machine, "all available processors" works best but maybe it'll be different for you. My machine uses a very ordinary CPU (a dual-core Intel G620). For a small (8-track) session I see a DSP reading of 38% with no plugins and a buffer size of 256. That drops to 21% if I increase my buffer size to 1024 (or 15% at 2048).

There's an old saying:- "you can't make a silk purse out of a sow's ear". The bottom line here is that if you want good performance you need a good PC. (or just stick with Ardour...)

[Matt]

If I open only a new file, whithout any new channels/tracks, Mixbus32C-4 (64-bit) and Mixbus32-3 32bit uses 50% and 60% of the DSP, after a while when it is stable. This is on my cheap Windows computer.

Do you know what make of CPU is in your computer? Intel CPUs have always been good at calculating the DSP load but for a long time (even back as far as Mixbus-2) AMD processors used to give errors in the calculation. In fact they still do AFAIK. Early in Mixbus-3, Harrison added some code to help minimize the errors but maybe there are still some edge cases where the corrections don't work.

OTOH if you're already using an Intel CPU, the most likely explanation is that your audio buffer size is too small (as others have already mentioned). In a previous post you said you're using a size of 256. On a cheap computer (not optimised for audio work) you probably need to use 1024 or even 2048.

According an earlier post he has a celeron n2830. Which appears to be a dual core at 2.16 GHz. And according to the benchmarks I've seen is probably good for general use. But imho is a bit lacking for audio work. Probably quite similar to the CPU in my first MacBook Pro. It would run MB 2, but also was approaching its max temps. Just looked it up. I had a 2009 MacBook Pro with a Core2Duo T7800, which scores around 1500.

And I can appreciate lack of funding. But I really do think a computer just for audio would be his best bet.