maxman65
Member
Hi has anyone noticed a difference in performance with this ram upgrade . There's no potential to upgrade the CPU in this machine (i7,) many thanks
D
Deleted member 196982
Guest
I’d put my money on the fastest SSD tech and put all your stuff there on USB 3/C paths, unless you can quantify the VST demand and understand how that RAM would be used. (I assume you are in 64-bit MacOS or Windows?)
P.s. I am not a big VST or track count user, but IMO audio is all about storage speeds. My 16GB Mini2 (apple silicon and fast RAM) beats the pants off the Intel iMac with 80GB.
P.s. I am not a big VST or track count user, but IMO audio is all about storage speeds. My 16GB Mini2 (apple silicon and fast RAM) beats the pants off the Intel iMac with 80GB.
Last edited by a moderator:
maxman65
Member
Hi it's windows 10. Lenovo think centre i7 6700 . I record in a standalone multitrack so demands will only really be for a live performance of a given library rather than multiple track playback from a daw. For spitfire chamber strings (28gb) I drop the sample buffer or whatever to 256 or 512
DM60
Well-known member
A couple of things. Depending on the VST or VSTi, some are CPU intensive, some are RAM. I would imagine your Spitfire would be RAM intensive since you uses "sound" files. Reverb for example is CPU intensive since it processes the sound.
On SSDs, usually they are fast enough to fetch data pretty fast, very seldom if ever for audio is there a lag. Also, depending on your DAW, many have a nice little feature to "freeze" a track. That just means in renders the audio in a temp file and requires no processing. Still allows for volume and L/R function to mix.
You might get some extra mileage with more RAM, but if it is lag you are struggling with, that is CPU and intensive processing most likely the cause. Unless you are playing live or recording vocals live I would keep the buffer size 512 or higher. I only drop the buffer lower when I am recording and listening to my "processed" voice (VST) or playing the guitar through an AMP sim or some other processing VST.
Hope that gives you an idea.
On SSDs, usually they are fast enough to fetch data pretty fast, very seldom if ever for audio is there a lag. Also, depending on your DAW, many have a nice little feature to "freeze" a track. That just means in renders the audio in a temp file and requires no processing. Still allows for volume and L/R function to mix.
You might get some extra mileage with more RAM, but if it is lag you are struggling with, that is CPU and intensive processing most likely the cause. Unless you are playing live or recording vocals live I would keep the buffer size 512 or higher. I only drop the buffer lower when I am recording and listening to my "processed" voice (VST) or playing the guitar through an AMP sim or some other processing VST.
Hope that gives you an idea.
maxman65
Member
Im trying to remember what happens . If I play the ensemble patch sometime the whole sample set of sound doesn't come through when you strike keys . I ( think ) dropping the buffer solved that . Meanwhile the cpu usage is shown on the spitfire library seem to remember it going to say 40 percent . Otherwise if I play the legato patch it's kind of odd feels laggy or something . A usful thing I found is to take the amount of( polyphony ?) on the spitfire from 400 to 1200 notes .anything to deduce from all this ?
DM60
Well-known member
Dropping your buffer numbers will decrease lag, but usually creates a higher load on the CPU. If you can drop the buffer and not get noise and odd behavior then that is a good direction. If you start getting odd sounds or behavior, increase the buffer size.
If you're not already aware, if you are duplicating the VSTi, I would suggest if it possible, routing MIDI channels from different tracks to a single instance (where possible) of the VSTi. VSTi is on track 1. Track 2 MIDI will say is channel 1 MIDI out. Set the instrument on the VSTi you want track 2 to play to the same MIDI channel. Then just record each track and set the out channels to each instrument. That will reduce your load and actually a better way of doing your composing. Then you can name each channel the part is is playing and give you much more creative control than having everything in one MIDI track.
If you are doing this, then ignore, but this is the best way to score with strings and get a more authentic sound. But as a rule, I keep the VSTi track with no MIDI information.
On the polyphony, that is just how many sounds it will process. If you are increasing it and getting better sound, could be becuase you are requesting it to play more sounds (400->1200), but that means higher VSTi load (the way MIDI works is it will load certain parts of the instrument for attack and tail, then synthesize the middle) and more sound files to process.
MIDI can be hard to get it to sound more real, but I think if you have a good VSTi (and it looks like you do), then it is a matter of understanding how to best manipulate the computer to get it to do what you want. MIDI or not, this part of recording really is pure digital as is all VST's and VSTi's).
If you're not already aware, if you are duplicating the VSTi, I would suggest if it possible, routing MIDI channels from different tracks to a single instance (where possible) of the VSTi. VSTi is on track 1. Track 2 MIDI will say is channel 1 MIDI out. Set the instrument on the VSTi you want track 2 to play to the same MIDI channel. Then just record each track and set the out channels to each instrument. That will reduce your load and actually a better way of doing your composing. Then you can name each channel the part is is playing and give you much more creative control than having everything in one MIDI track.
If you are doing this, then ignore, but this is the best way to score with strings and get a more authentic sound. But as a rule, I keep the VSTi track with no MIDI information.
On the polyphony, that is just how many sounds it will process. If you are increasing it and getting better sound, could be becuase you are requesting it to play more sounds (400->1200), but that means higher VSTi load (the way MIDI works is it will load certain parts of the instrument for attack and tail, then synthesize the middle) and more sound files to process.
MIDI can be hard to get it to sound more real, but I think if you have a good VSTi (and it looks like you do), then it is a matter of understanding how to best manipulate the computer to get it to do what you want. MIDI or not, this part of recording really is pure digital as is all VST's and VSTi's).
maxman65
Member
Many thanks will take me a while to absorb what's being said there . As far as I know it's usb ( maybe 3) audio out from the computer to the yamaha modx ( controller keyboard) . Not sure I've ever played with midi channels ? Or would know how to . In terms of recording there's no load in the daw itself because I come audio out from the keyboard into a tascam standalone 8 track machine and record there
DM60
Well-known member
Is the orchestra plugin a stand alone? I assumed it was a VSTi.
maxman65
Member
Yes it's a vst instrument that i think sits in kontact player however I don't use daw software to record in. I also have cubase ai as a kind of host but in effect don't really use it
DM60
Well-known member
OK, so you are more or less using the computer as an engine to pass through sounds to your Tascam? In essence, you are using the VSTi as a type of MIDI sound module. I use that word because early MIDI originally used sound modules to expand the sound pallet. That is why they had channels and would send program signals (what sound do you want to use). You are using your computer pretty much like a sound module.
I would think, to get you further along, you should record the MIDI part for each part. This is where Cubase would help. Select a track, put the VSTi on a track (or like I said earlier) then when you are done, you can mix down to the Tascam (or record multi tracks one at a time).
The Tascam is really best use for Live recording with multiple people. Once you get into MIDI, I think it falls down and does not serve you well. It is still usable as an input for guitar, mics, etc. It has ASIO drivers for your DAW (Cubase) so that you can use it for recording Live instruments and Listening to your Mix through speakers (monitors).
I really think you should get your DAW setup and learn how to use it. Record your MIDI to tracks, then MIX and produce. Render/mix down in the DAW. Use your standalone for mobile recording. Like getting sound samples, recording vocals in a certain area, there are all kinds of uses for your standalone, but it is not the best with MIDI. It can be used, but there are much easier ways.
Hopefully this doesn't discourage you, but I think you are getting frustrated because you are really doing it the hard way.
I would think, to get you further along, you should record the MIDI part for each part. This is where Cubase would help. Select a track, put the VSTi on a track (or like I said earlier) then when you are done, you can mix down to the Tascam (or record multi tracks one at a time).
The Tascam is really best use for Live recording with multiple people. Once you get into MIDI, I think it falls down and does not serve you well. It is still usable as an input for guitar, mics, etc. It has ASIO drivers for your DAW (Cubase) so that you can use it for recording Live instruments and Listening to your Mix through speakers (monitors).
I really think you should get your DAW setup and learn how to use it. Record your MIDI to tracks, then MIX and produce. Render/mix down in the DAW. Use your standalone for mobile recording. Like getting sound samples, recording vocals in a certain area, there are all kinds of uses for your standalone, but it is not the best with MIDI. It can be used, but there are much easier ways.
Hopefully this doesn't discourage you, but I think you are getting frustrated because you are really doing it the hard way.
Steenamaroo
...
With memory and CPU usage you can just open Task Manager and look at the performance tab and know for sure if you'll benefit from an upgrade or not.
Fire up the heaviest workflow you've got then watch TaskManager as it's running.
If the memory in-use % is well below 100, then you won't benefit from buying more RAM.
CPU usage is a little more volatile and harder to interpret. It's not uncommon to see that briefly spike to 100% then recover to some lower figure.
Unless it's constantly pegged at, or very close to, 100 under load you're good.
Fire up the heaviest workflow you've got then watch TaskManager as it's running.
If the memory in-use % is well below 100, then you won't benefit from buying more RAM.
CPU usage is a little more volatile and harder to interpret. It's not uncommon to see that briefly spike to 100% then recover to some lower figure.
Unless it's constantly pegged at, or very close to, 100 under load you're good.
maxman65
Member
Yes I'll try that again . When I looked last the time the CPU was moving dynamically and the memory was fixed at 20 percent never moved . Not sure what that's about
DM60
Well-known member
CPU moving dynamic is normal, just monitor and see if it goes to 100% and stays for awhile. Memory at 20% tells me, you don't need more memory. But you need to check it while running it at full load. If you are not doing anything the test doesn't mean anything.
Steenamaroo
...
+1 to this.
If I had a £ for every time people just went off and upgraded their ram regardless here I'd be retired but what DM says is true.
If you run a heavy session and your ram doesn't get over 20% (or any % well clear of 100), more ram won't help.
It'd be upgrading your 10 litre container to a 20 litre container to hold your 5 litres of water.
CPU usage bouncing around is, as I said and DM said, expected.
As long as it doesn't appear pegged out at 100% for sustained periods, all is well.
DM60 mentioned SSDs. If your machine is running a spinning disk definitely consider upgrading to an SSD.
We're at the stage where upgrading from mechanical to solid state is likely to show improvement on more or less any workload or use-case.
If I had a £ for every time people just went off and upgraded their ram regardless here I'd be retired but what DM says is true.
If you run a heavy session and your ram doesn't get over 20% (or any % well clear of 100), more ram won't help.
It'd be upgrading your 10 litre container to a 20 litre container to hold your 5 litres of water.
CPU usage bouncing around is, as I said and DM said, expected.
As long as it doesn't appear pegged out at 100% for sustained periods, all is well.
DM60 mentioned SSDs. If your machine is running a spinning disk definitely consider upgrading to an SSD.
We're at the stage where upgrading from mechanical to solid state is likely to show improvement on more or less any workload or use-case.
maxman65
Member
Hi it's ssd .can't speak for it's age but I suspect maybe 2016 like the machine
DM60
Well-known member
If it is SSD, the gains have been better over the last few years, but not in a huge way. I think if it is a PCI SSD, then the limit is on the PCI bus. Reliability and longevity have been the greatest gains. That is most likely not your issue. You might want to look at the software and see if there is a way to preset the RAM buffer size so everything is already preloaded. 32 GB should get a lot of the library in memory. I don't think it is the computer but the software. I would research how to improve performance there.