Plugin latency overview
- KVRian
- 912 posts since 1 Nov, 2012 from Berlin
Let's get rocking!
Check the documentation of your favourite plugin to see how many samples and msec latency it produces and post it here.
This might help us all to define which values make sense to compensate and which not.
I'll start with a quote from the great Fabfilter Pro Q:
Zero latency mode is the default. While it introduces phase changes, it is CPU-efficient and doesn't result in any latency, so it is the best mode for e.g. live usage. Also, it's quite possible you might like the coloration introduced by the phase changes when mixing, for example.
Linear Phase - Low Latency provides linear-phase processing with a minimal latency. Use only with low Q settings, or when only changing the mid-high part of the spectrum. With a sample rate of 44.1 kHz, it results in a total latency of 3072 samples (about 70 ms).
Linear Phase - Medium Latency is a good compromise between low-frequency resolution and latency and we recommend to use this in general for linear-phase processing. The total latency is 6144 samples at a sample rate of 44.1 kHz (about 139 ms).
Linear Phase - High Latency gives very good low-frequency resolution. If you need to use high Q settings when changing the low end of the spectrum, use this mode. The total latency is 12288 samples at a sample rate of 44.1 kHz (about 279 ms).
Linear Phase - Maximum Latency results in even better low-frequency resolution at the expense of latency and possible pre-echo problems. The total latency here is 24576 samples at a sample rate of 44.1 kHz (about 557 ms).
To conclude, Pro-Q lets you freely choose between zero-latency and linear-phase processing as you go. If you use high Q settings combined with low-frequency filtering, you need to use a higher latency; if you only work on the mid-high frequencies, you can get by with a lower latency.
----
Sorry, guys, deleted two posts. Please keep it on-topic, the result should be a list where we all can look up stuff. Thanks!
Check the documentation of your favourite plugin to see how many samples and msec latency it produces and post it here.
This might help us all to define which values make sense to compensate and which not.
I'll start with a quote from the great Fabfilter Pro Q:
Zero latency mode is the default. While it introduces phase changes, it is CPU-efficient and doesn't result in any latency, so it is the best mode for e.g. live usage. Also, it's quite possible you might like the coloration introduced by the phase changes when mixing, for example.
Linear Phase - Low Latency provides linear-phase processing with a minimal latency. Use only with low Q settings, or when only changing the mid-high part of the spectrum. With a sample rate of 44.1 kHz, it results in a total latency of 3072 samples (about 70 ms).
Linear Phase - Medium Latency is a good compromise between low-frequency resolution and latency and we recommend to use this in general for linear-phase processing. The total latency is 6144 samples at a sample rate of 44.1 kHz (about 139 ms).
Linear Phase - High Latency gives very good low-frequency resolution. If you need to use high Q settings when changing the low end of the spectrum, use this mode. The total latency is 12288 samples at a sample rate of 44.1 kHz (about 279 ms).
Linear Phase - Maximum Latency results in even better low-frequency resolution at the expense of latency and possible pre-echo problems. The total latency here is 24576 samples at a sample rate of 44.1 kHz (about 557 ms).
To conclude, Pro-Q lets you freely choose between zero-latency and linear-phase processing as you go. If you use high Q settings combined with low-frequency filtering, you need to use a higher latency; if you only work on the mid-high frequencies, you can get by with a lower latency.
----
Sorry, guys, deleted two posts. Please keep it on-topic, the result should be a list where we all can look up stuff. Thanks!
- KVRian
- Topic Starter
- 912 posts since 1 Nov, 2012 from Berlin
Uh, nobody wants to report his favourite latency introducing plugin but everyone seems to care for PDC? 
That's a bit strange! Help us out here!
That's a bit strange! Help us out here!
-
- KVRer
- 22 posts since 27 Mar, 2014
I've redone my testcases with Ozone5, introducing total of 680,4ms, automated mute with Bitwig "tool" before, behind and between Ozones. While most cases failed with 1.02, 1.05 shows no more problems: Automation is always in sync, all (!) Metering is in sync and Song Position is on track.
Wow, impressive, this was great work - a superfast fix. Congratulations!
So my 4year no1. Ableton issue is solved within Bitwig now - in just a week.
If you keep on this pace, I will change very very soon. Well done!
One last PDC thing that at least does look strange in my scenario is the automated "knob" in tool. It is visually "ahead" - looking a bit strange while all the rest is optically in sync. But this would not stop me from attesting working PDC.
...I'm continuing with other tests now...
Wow, impressive, this was great work - a superfast fix. Congratulations!
So my 4year no1. Ableton issue is solved within Bitwig now - in just a week.
One last PDC thing that at least does look strange in my scenario is the automated "knob" in tool. It is visually "ahead" - looking a bit strange while all the rest is optically in sync. But this would not stop me from attesting working PDC.
...I'm continuing with other tests now...
-
- KVRer
- 22 posts since 27 Mar, 2014
by the way, ozone introduces up to 250ms - at least I wasn't able to configure more complexity. I typically have a maximum of 2 in the whole chain - each delaying between 100 and 200 ms.
Just for the testcase i did use 3 of them (2 on the automated track, 1 in the master)
Then I use a lot of compression, often with prelisten (by the way i miss lookahead in yours...)
my song-maximum i remember from live was ~800ms. Typically i have ~400ms.
Just for the testcase i did use 3 of them (2 on the automated track, 1 in the master)
Then I use a lot of compression, often with prelisten (by the way i miss lookahead in yours...)
my song-maximum i remember from live was ~800ms. Typically i have ~400ms.
- KVRian
- Topic Starter
- 912 posts since 1 Nov, 2012 from Berlin
Thanks, but please lets only list plugins and their latency in sample in this thread, as explained above!
-
glitter&trauma glitter&trauma https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=326026
- KVRer
- 12 posts since 1 Apr, 2014
Waves Native Plugins @ 44.1kHz
PLUGIN NAME - # OF SAMPLES LATENCY
Bass Rider - 1852
C1 Compressor (SC) - 340
CLA Bass - 133
CLA Vocals - 193
DeBreath - 32384
Eddie Kramer Bass Channel - 493
Eddie Kramer Drum Channel - 410
Eddie Kramer Vocal Channel - 149
H-Comp Hybrid Compressor - 64
H-EQ Hybrid Equalizer - 65
InPhase - 882
J37 Tape - 101
JJP Bass - 409
JJP Cymbals & Percussions - 557
JJP Drums - 622
JJP Guitars - 622
JJP Strings & Keys - 532
JJP Vocals - 473
Kramer HLS Channel - 161
Kramer Master Tape - 59
L1 Ultramaximizer - 64
L2 Ultramaximizer - 64
L3 Multimaximizer - 3528
L3-16 Multimaximizer - 6207
Linear Phase EQ (Broadband) - 2679
Linear Phase EQ (Lowband) - 2047
Manny Marroquin Delay - 64
Manny Marroquin Distortion - 128
Manny Marroquin Tone Shaper - 2968
Manny Marroquin Reverb - 1
Maserati ACG - 64
Maserati B72 - 68
Maserati GRP - 1
Maserati GTi - 621
Maserati HMX - 1
Maserati VX1 - 408
MaxxVolume - 64
MetaFilter - 84
Morphoder - 639
MV2 - 64
MV360 - 64
NLS Non-Linear Summer - 4
Reel ADT - 1881
Renaissance Axx - 64
Renaissance Channel - 65
SoundShifter - 6946
UltraPitch - 8239
Trans-X - 64
Waves Tune - 3072
Waves Tune LT - 3072
X-Click - 2625
X-Crackle - 2625
X-Noise - 5120
Z-Noise - 34702
thanks dom & everyone at Bitwig HQ, i'm in love with this software so i thought id make an account here and try to help out, danke!
PLUGIN NAME - # OF SAMPLES LATENCY
Bass Rider - 1852
C1 Compressor (SC) - 340
CLA Bass - 133
CLA Vocals - 193
DeBreath - 32384
Eddie Kramer Bass Channel - 493
Eddie Kramer Drum Channel - 410
Eddie Kramer Vocal Channel - 149
H-Comp Hybrid Compressor - 64
H-EQ Hybrid Equalizer - 65
InPhase - 882
J37 Tape - 101
JJP Bass - 409
JJP Cymbals & Percussions - 557
JJP Drums - 622
JJP Guitars - 622
JJP Strings & Keys - 532
JJP Vocals - 473
Kramer HLS Channel - 161
Kramer Master Tape - 59
L1 Ultramaximizer - 64
L2 Ultramaximizer - 64
L3 Multimaximizer - 3528
L3-16 Multimaximizer - 6207
Linear Phase EQ (Broadband) - 2679
Linear Phase EQ (Lowband) - 2047
Manny Marroquin Delay - 64
Manny Marroquin Distortion - 128
Manny Marroquin Tone Shaper - 2968
Manny Marroquin Reverb - 1
Maserati ACG - 64
Maserati B72 - 68
Maserati GRP - 1
Maserati GTi - 621
Maserati HMX - 1
Maserati VX1 - 408
MaxxVolume - 64
MetaFilter - 84
Morphoder - 639
MV2 - 64
MV360 - 64
NLS Non-Linear Summer - 4
Reel ADT - 1881
Renaissance Axx - 64
Renaissance Channel - 65
SoundShifter - 6946
UltraPitch - 8239
Trans-X - 64
Waves Tune - 3072
Waves Tune LT - 3072
X-Click - 2625
X-Crackle - 2625
X-Noise - 5120
Z-Noise - 34702
thanks dom & everyone at Bitwig HQ, i'm in love with this software so i thought id make an account here and try to help out, danke!
-
- KVRist
- 179 posts since 1 Apr, 2014
Ok, this is the latency reported in Bitwig:
Fabfilter Saturn(HQ mode) - 0.2ms
PSP VintageWarmer2 - 14.5ms
U-He Satin - 0.5ms
SPL DrumXchanger - 50.1ms
Brainworx XL v2 - 90.3ms
Fabfilter Pro DS - 26.6ms
Fabfilter Pro MB(linear phase, oversampling 4x, look ahead on) - 91.2ms
Equilibrium - 92.9ms
Ozone 5(everything on) - 380.3ms
U-He Uhbik plugins - 0.4ms each
Zynaptiq Pitchmap - 92.9ms
Lexicon Reverbs - 0.7ms
Waves GTR Amp - 0.8ms
Waves LinMB - 80ms
Fabfilter Saturn(HQ mode) - 0.2ms
PSP VintageWarmer2 - 14.5ms
U-He Satin - 0.5ms
SPL DrumXchanger - 50.1ms
Brainworx XL v2 - 90.3ms
Fabfilter Pro DS - 26.6ms
Fabfilter Pro MB(linear phase, oversampling 4x, look ahead on) - 91.2ms
Equilibrium - 92.9ms
Ozone 5(everything on) - 380.3ms
U-He Uhbik plugins - 0.4ms each
Zynaptiq Pitchmap - 92.9ms
Lexicon Reverbs - 0.7ms
Waves GTR Amp - 0.8ms
Waves LinMB - 80ms
-
Echoes in the Attic Echoes in the Attic https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=180417
- KVRAF
- 12044 posts since 12 May, 2008
^Not much good to list latency in ms as that depends on sample rate.
Here's a few:
Alchemy - 64 Samples
Addictive Keys - 64 Samples
2CAudio B2 - 64 Samples
Aalto - I forget, will check later. 64 samples maybe?
Here's a few:
Alchemy - 64 Samples
Addictive Keys - 64 Samples
2CAudio B2 - 64 Samples
Aalto - I forget, will check later. 64 samples maybe?
-
- KVRist
- 179 posts since 1 Apr, 2014
Ah, I was just reporting what Bitwig said in the mixer. I'm at 256 buffer size, 44.1khz.Echoes in the Attic wrote:^Not much good to list latency in ms as that depends on sample rate.
-
Echoes in the Attic Echoes in the Attic https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=180417
- KVRAF
- 12044 posts since 12 May, 2008
You can get an approximation like this:Simulant wrote:Ah, I was just reporting what Bitwig said in the mixer. I'm at 256 buffer size, 44.1khz.Echoes in the Attic wrote:^Not much good to list latency in ms as that depends on sample rate.
#Samples = Sample Rate x Time (ms)
So #Samples = 44.1 x (1.5) = 66.15. This isn't actually 66.15 because the 1.5 ms reported is rounded. Really it's 64 samples and the latency is around 1.45. But using the reported latency multiplied by the sample rate will get you close.
- KVRAF
- 1758 posts since 15 Mar, 2013 from Germany
Waves list their latencies here:
http://www.waves.com/support/tech-specs/plugin-latency
This list differs quite a lot from that posted a few posts above.
http://www.waves.com/support/tech-specs/plugin-latency
This list differs quite a lot from that posted a few posts above.
-
- KVRer
- 22 posts since 27 Mar, 2014
!
Just found that many plugins differ in samples depending on the sample rate. So time in ms could be a more valid approach...
Meaning, that BWS should adjust internal audio buffer / PDC delay sizes based on sample rate!
This makes overall sense: a compressor does have a lookahead in ms, not in samples. Same is meaningful for all maximizer and eq algorithms, otherwise they would sound different (better) @44khz compared to 96khz...
cheers
Kay
Just found that many plugins differ in samples depending on the sample rate. So time in ms could be a more valid approach...
Meaning, that BWS should adjust internal audio buffer / PDC delay sizes based on sample rate!
This makes overall sense: a compressor does have a lookahead in ms, not in samples. Same is meaningful for all maximizer and eq algorithms, otherwise they would sound different (better) @44khz compared to 96khz...
cheers
Kay
-
glitter&trauma glitter&trauma https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=326026
- KVRer
- 12 posts since 1 Apr, 2014
nope, its 100% accurate. theres 6x more info than needed there, were talking plugin latency at 44.1. Not 176 or 192. Not TDM. Use your eyesCableChannel wrote:Waves list their latencies here:
http://www.waves.com/support/tech-specs/plugin-latency
This list differs quite a lot from that posted a few posts above.
