GaryMedia
Max Output Level: -86 dBFS
- Total Posts : 217
- Joined: 2003/11/05 23:04:20
- Location: Cary, NC
- Status: offline
Mixdown performance options?
I've been going through a series of full project mixdowns lately; putting in the final tweaks on a 14-song CD. The waiting time for each mixdown has lead me to inquire about what the "gating" factors are for mixdown performance.
I can see that Sonar ensures proper plug-in processing by using a single engine of the Quad-Core CPU. That leads me to believe that raw clock speed of the CPU is a major factor. I can't tell if the hard drive particularly challenged during this time, my guess is that wouldn't work much harder than it does during playback, possibly in proportion to how much faster than real-time the rendering is able to process. By the way, I have UAD-1 DSP's in this machine.
So, folks, please share any performance tweaks or insights about what it would take to slice minutes off of my mixdown waiting time.
CbB Win10 | Mac Pro 12-core 3.33GHz/48GB | TCL 55" 4K UHD | 480GB SSD | 6TB HDD RAID-5 array| 1.5TB SSD RAID-0 array | Midas M32 | 2x Audient ASP800 | UAD-2 Duo PCIe | Adam A7X. http://www.tedlandstudio.com/articles
|
SteveD
Max Output Level: -47 dBFS
- Total Posts : 2831
- Joined: 2003/11/07 13:35:57
- Location: NJ
- Status: offline
RE: Mixdown performance options?
2009/04/06 21:55:48
(permalink)
ORIGINAL: GaryMedia
I've been going through a series of full project mixdowns lately; putting in the final tweaks on a 14-song CD. The waiting time for each mixdown has lead me to inquire about what the "gating" factors are for mixdown performance.
I can see that Sonar ensures proper plug-in processing by using a single engine of the Quad-Core CPU. That leads me to believe that raw clock speed of the CPU is a major factor. I can't tell if the hard drive particularly challenged during this time, my guess is that wouldn't work much harder than it does during playback, possibly in proportion to how much faster than real-time the rendering is able to process. By the way, I have UAD-1 DSP's in this machine.
So, folks, please share any performance tweaks or insights about what it would take to slice minutes off of my mixdown waiting time.
What sample rate and bit depth are you using?
|
Ron Vogel
Max Output Level: -69 dBFS
- Total Posts : 1074
- Joined: 2008/07/18 14:14:40
- Status: offline
RE: Mixdown performance options?
2009/04/07 01:34:49
(permalink)
The CPU is DEFINATELY a factor. Put a CPU temp meter on while mixdown is happening. I have a AMD 6400x2 (which is known to be hot running anyway). It only took one big mixdown for me to order a bunch of better case fans! CPU temps climbed as high as 68C.
|
GaryMedia
Max Output Level: -86 dBFS
- Total Posts : 217
- Joined: 2003/11/05 23:04:20
- Location: Cary, NC
- Status: offline
RE: Mixdown performance options?
2009/04/07 12:55:25
(permalink)
The sample rate is 44.1Khz, the bit depth is 24. I mentioned the UAD-1's because I suspect that their DSP contribution to mixdown processing is not going to change with the native CPU speed.
I'll do some formal timings later, but my impression is that there is a "floor" to the mixdown time that is independent of the track count and the number of plug-ins. I'm curious about the contributors to that 'floor'. The 47-track mixes with 29 active plug-ins that take longer than the real-time duration of the song are the ones that piqued my interest in this issue.
My new motherboard just arrived today, so I'll be working fervently to completely finishing everything before I tear apart my DAW to do the re-build. I'll at least get timings for the smallest mixes (acoustic guitar and lone singer) as well as the extravaganza with the four synths, 3 bkgrnd vocalists, six instruments and the cowbell.
CbB Win10 | Mac Pro 12-core 3.33GHz/48GB | TCL 55" 4K UHD | 480GB SSD | 6TB HDD RAID-5 array| 1.5TB SSD RAID-0 array | Midas M32 | 2x Audient ASP800 | UAD-2 Duo PCIe | Adam A7X. http://www.tedlandstudio.com/articles
|
SteveD
Max Output Level: -47 dBFS
- Total Posts : 2831
- Joined: 2003/11/07 13:35:57
- Location: NJ
- Status: offline
RE: Mixdown performance options?
2009/04/07 13:52:23
(permalink)
ORIGINAL: GaryMedia
The sample rate is 44.1Khz, the bit depth is 24. I mentioned the UAD-1's because I suspect that their DSP contribution to mixdown processing is not going to change with the native CPU speed.
I'll do some formal timings later, but my impression is that there is a "floor" to the mixdown time that is independent of the track count and the number of plug-ins. I'm curious about the contributors to that 'floor'. The 47-track mixes with 29 active plug-ins that take longer than the real-time duration of the song are the ones that piqued my interest in this issue.
My new motherboard just arrived today, so I'll be working fervently to completely finishing everything before I tear apart my DAW to do the re-build. I'll at least get timings for the smallest mixes (acoustic guitar and lone singer) as well as the extravaganza with the four synths, 3 bkgrnd vocalists, six instruments and the cowbell.
I've got 4 UAD-1 cards in a magma chassis, and one UAD-2 PCIe card in the DAW, along with the Waves Diamond Bundle, and various Voxengo plugins... and of course the Sonitus Suite. I'm using SONAR 8PE on a Q6600 processor like you, with a Gigabyte motherboard. Here in my studio... it never takes longer to do a fast bounce than it does to play the song or perform a real time bounce... especially at 44/24. I do my bounces at 64bit with no dither, and that does add some time, but it's still faster than playing the song. My track counts are usually greater than 40 tracks with 50 or so active plugins. What plugins are you using?
|
GaryMedia
Max Output Level: -86 dBFS
- Total Posts : 217
- Joined: 2003/11/05 23:04:20
- Location: Cary, NC
- Status: offline
RE: Mixdown performance options?
2009/04/07 19:32:13
(permalink)
While churning though my latest set of revisions, I've been timing the mixdown elapsed time. Essentially I've learned that the old adage "A watched pot never boils" seems to explain my situation rather well.
In every case, the mixdown time was less than the real-time playback time. In the very worst case, it was just a few seconds short of real-time. In most cases it was about half of the real-time playback.
It would be quite tedious to describe all of the plug-ins that are active. In general I use a 14-buss layout wherein each buss has one or two plug-ins. There's a hall, room and plate buss, a vocal buss that has two plug-ins, a bass buss with two, and so on. Most are UAD's but a few are Voxengo and Sonar-packaged TL64's. Tracks are routed through some busses and/or and have sends to them.
Overall, I'm still curious about whether there is some way to make mixdowns a matter of 20 seconds rather than 3-to-7 minutes, but at least now I know that I'm consistently bouncing faster than real-time playback.
CbB Win10 | Mac Pro 12-core 3.33GHz/48GB | TCL 55" 4K UHD | 480GB SSD | 6TB HDD RAID-5 array| 1.5TB SSD RAID-0 array | Midas M32 | 2x Audient ASP800 | UAD-2 Duo PCIe | Adam A7X. http://www.tedlandstudio.com/articles
|
SteveD
Max Output Level: -47 dBFS
- Total Posts : 2831
- Joined: 2003/11/07 13:35:57
- Location: NJ
- Status: offline
RE: Mixdown performance options?
2009/04/07 19:36:59
(permalink)
ORIGINAL: GaryMedia
While churning though my latest set of revisions, I've been timing the mixdown elapsed time. Essentially I've learned that the old adage "A watched pot never boils" seems to explain my situation rather well.
In every case, the mixdown time was less than the real-time playback time. In the very worst case, it was just a few seconds short of real-time. In most cases it was about half of the real-time playback.
It would be quite tedious to describe all of the plug-ins that are active. In general I use a 14-buss layout wherein each buss has one or two plug-ins. There's a hall, room and plate buss, a vocal buss that has two plug-ins, a bass buss with two, and so on. Most are UAD's but a few are Voxengo and Sonar-packaged TL64's. Tracks are routed through some busses and/or and have sends to them.
Overall, I'm still curious about whether there is some way to make mixdowns a matter of 20 seconds rather than 3-to-7 minutes, but at least now I know that I'm consistently bouncing faster than real-time playback.
I use lots of busses too... very similar to what you're doing. 3-7 minutes is about right... although since External Inserts is now sample accurate and totally fixed with SONAR 8.3.1, it doesn't really matter much to me anymore. I'm usually doing real time bounces now.
|