FM9 Latency - Around 5ms?

unix-guy said:
Thanks for the details (mostly over my head). I suppose the option to choose lower latency vs better accuracy could be useful for those who want it but I defer to your judgement.

I think a key question raised in this thread that you didn't talk about is why the FM3 latency is more in line with the Axe Fx III than the FM9. Curious about the results there...
Click to expand...
The FM3 and FM9 are the same.
 
Randalljax said:
How does the second method you mentioned smaller kernel impact tone and performance
What are the negative impacts of stop band or transition to the overall sound
Click to expand...
Less stop-band rejection means more aliasing artifacts.

Wider transition band means more rolloff in the high end.

You want your transition band as narrow as possible but the narrower the band the longer the kernel or the less stop-band rejection.

Here's a good paper on the subject:
https://www.vyssotski.ch/BasicsOfInstrumentation/SpikeSorting/Design_of_FIR_Filters.pdf
 
unix-guy said:
Ok, interesting since apparently G66 measured all 3 and found the FM3 to be about 2.9ms:

https://forum.fractalaudio.com/threads/fm9-latency-around-5ms.200203/post-2496
Click to expand...
I have tested the FM3. But unlike the FM9, not with an oscilloscope but only with the DAW.
IN-AMP-BLOCK (TUBE PRE) - OUT . Nothing more. The whole thing was tested with a sine wave generator at 120 Hz. The result was read by two people. Unfortunately, this is not so easy to read with a DAW, as the time scale was not exactly made for this and you have to enlarge the image on the computer. And at the limit, it was more of an estimate.
If Cliff Chase says it's 3.3ms then I would definitely take this value as true and it also shows that my measurement wasn't that bad. Come on, 0.4 ms. That's nothing ;-)

Several tests with the FM9 ( IN- DRV- AMP-CAB- OUT) showed almost the same value as described by Mr Chase. I can't remember now but it was under! 5 ms, not 5 or more ms.

And I don't find this value problematic. You only have to read through discussions in QC forums. There are reports of up to 10 ms, depending on the prestes. They say the Kemper has an average latency of 3.6 ms. If this is the case, then I can only say that nobody in the world will hear a difference of 0.5 ms. But the quality of the sound will.

I see the problem much more with latency stacking. If users report that they feel latency when playing in the studio, then they should please check their setup. Because many monitoring monitors also have a DSP on board that adds up to 5 ms of latency. This was the case with my KS digital C88. That's why I replaced it. Don't get me wrong, the C88 from KS digital are incredible speakers but not designed for guitarists who want to play through a modeller. That's why I bought these speakers.

https://www.psiaudio.de/a21-m-2022-2-weg-studiomonitor/
100% analogue and no latency.
Why don't you check your studio monitors to see if they have a DSP and ask the manufacturer about the latency ;-)

Another point is the FX loop. With the number of high-quality effects that the Fractal devices offer, I wonder why you still have to integrate external pedals. Take the Eventide H90 for example. Eventide itself has confirmed that it has a latency of 5 ms. If this is now looped into the FM9, the latency increases by this value plus the DA-AD conversion time.
If you use the H90 with a tube amp, you have more latency than with an FM9 (serial loop).

And why use cheap transmitters like the SHURE DLX? Here is an extract from their website:
QUESTION:
What is the Latency of GLX-D+ Dual Band?
ANSWER
Latency varies slightly depending on the Frequency Band and Group:

Z3, Z4, Z5 (2.4 & 5.8GHz Bands):
  • Group 1: 4.53ms
  • Group 2: 8ms
  • Group 3: 4.53ms
Z2 (2.4GHz band only):
  • Group 1: 4.53ms
  • Group 2: 8ms
  • Group 3: 4.53ms
  • Group 4: 8ms
These are ridiculous values. You should definitely prefer cable or use a high-quality analogue transmitter.
https://www.thomann.de/gb/sennheiser_ew_500_g4_ci1_aw_band.htm

In conclusion, you can only say that it doesn't matter whether a modeller has 3.2 or 4.5 ms. But whatever peripherals are added can ultimately result in a clearly noticeable latency.

That's why I am happy with my FM9 wehen playing with a cable and a pair of RESOUND ELIS 8.
The difference between a true analog setup is, that you have to move the cabinets 1.2 meters closer to your ear. 😄
 
Keybi said:
In conclusion, you can only say that it doesn't matter whether a modeller has 3.2 or 4.5 ms
Click to expand...
Sorry, you cannot "only" say that, that's just an opinion. There is another opinion, but I will not use "my experience" again for this kind of argument, let's hear some experienced people discuss it:
 
Keybi said:
I have tested the FM3. But unlike the FM9, not with an oscilloscope but only with the DAW.
IN-AMP-BLOCK (TUBE PRE) - OUT . Nothing more. The whole thing was tested with a sine wave generator at 120 Hz. The result was read by two people. Unfortunately, this is not so easy to read with a DAW, as the time scale was not exactly made for this and you have to enlarge the image on the computer. And at the limit, it was more of an estimate.
If Cliff Chase says it's 3.3ms then I would definitely take this value as true and it also shows that my measurement wasn't that bad. Come on, 0.4 ms. That's nothing ;-)

Several tests with the FM9 ( IN- DRV- AMP-CAB- OUT) showed almost the same value as described by Mr Chase. I can't remember now but it was under! 5 ms, not 5 or more ms.

And I don't find this value problematic. You only have to read through discussions in QC forums. There are reports of up to 10 ms, depending on the prestes. They say the Kemper has an average latency of 3.6 ms. If this is the case, then I can only say that nobody in the world will hear a difference of 0.5 ms. But the quality of the sound will.

I see the problem much more with latency stacking. If users report that they feel latency when playing in the studio, then they should please check their setup. Because many monitoring monitors also have a DSP on board that adds up to 5 ms of latency. This was the case with my KS digital C88. That's why I replaced it. Don't get me wrong, the C88 from KS digital are incredible speakers but not designed for guitarists who want to play through a modeller. That's why I bought these speakers.

https://www.psiaudio.de/a21-m-2022-2-weg-studiomonitor/
100% analogue and no latency.
Why don't you check your studio monitors to see if they have a DSP and ask the manufacturer about the latency ;-)

Another point is the FX loop. With the number of high-quality effects that the Fractal devices offer, I wonder why you still have to integrate external pedals. Take the Eventide H90 for example. Eventide itself has confirmed that it has a latency of 5 ms. If this is now looped into the FM9, the latency increases by this value plus the DA-AD conversion time.
If you use the H90 with a tube amp, you have more latency than with an FM9 (serial loop).

And why use cheap transmitters like the SHURE DLX? Here is an extract from their website:
QUESTION:
What is the Latency of GLX-D+ Dual Band?
ANSWER
Latency varies slightly depending on the Frequency Band and Group:

Z3, Z4, Z5 (2.4 & 5.8GHz Bands):
  • Group 1: 4.53ms
  • Group 2: 8ms
  • Group 3: 4.53ms
Z2 (2.4GHz band only):
  • Group 1: 4.53ms
  • Group 2: 8ms
  • Group 3: 4.53ms
  • Group 4: 8ms
These are ridiculous values. You should definitely prefer cable or use a high-quality analogue transmitter.
https://www.thomann.de/gb/sennheiser_ew_500_g4_ci1_aw_band.htm

In conclusion, you can only say that it doesn't matter whether a modeller has 3.2 or 4.5 ms. But whatever peripherals are added can ultimately result in a clearly noticeable latency.

That's why I am happy with my FM9 wehen playing with a cable and a pair of RESOUND ELIS 8.
The difference between a true analog setup is, that you have to move the cabinets 1.2 meters closer to your ear. 😄
Click to expand...
Thanks for the details, and I'm definitely not complaining.

For me it's just trying to understand the numbers for comparison.

It sounds like a) the FM3 test wasn't as precise as the others and b) the FM3 didn't have the same signal chain (no Cab).
 
Keybi said:
I have tested the FM3. But unlike the FM9, not with an oscilloscope but only with the DAW.
IN-AMP-BLOCK (TUBE PRE) - OUT . Nothing more. The whole thing was tested with a sine wave generator at 120 Hz. The result was read by two people. Unfortunately, this is not so easy to read with a DAW, as the time scale was not exactly made for this and you have to enlarge the image on the computer. And at the limit, it was more of an estimate.
If Cliff Chase says it's 3.3ms then I would definitely take this value as true and it also shows that my measurement wasn't that bad. Come on, 0.4 ms. That's nothing ;-)

Several tests with the FM9 ( IN- DRV- AMP-CAB- OUT) showed almost the same value as described by Mr Chase. I can't remember now but it was under! 5 ms, not 5 or more ms.

And I don't find this value problematic. You only have to read through discussions in QC forums. There are reports of up to 10 ms, depending on the prestes. They say the Kemper has an average latency of 3.6 ms. If this is the case, then I can only say that nobody in the world will hear a difference of 0.5 ms. But the quality of the sound will.

I see the problem much more with latency stacking. If users report that they feel latency when playing in the studio, then they should please check their setup. Because many monitoring monitors also have a DSP on board that adds up to 5 ms of latency. This was the case with my KS digital C88. That's why I replaced it. Don't get me wrong, the C88 from KS digital are incredible speakers but not designed for guitarists who want to play through a modeller. That's why I bought these speakers.

https://www.psiaudio.de/a21-m-2022-2-weg-studiomonitor/
100% analogue and no latency.
Why don't you check your studio monitors to see if they have a DSP and ask the manufacturer about the latency ;-)

Another point is the FX loop. With the number of high-quality effects that the Fractal devices offer, I wonder why you still have to integrate external pedals. Take the Eventide H90 for example. Eventide itself has confirmed that it has a latency of 5 ms. If this is now looped into the FM9, the latency increases by this value plus the DA-AD conversion time.
If you use the H90 with a tube amp, you have more latency than with an FM9 (serial loop).

And why use cheap transmitters like the SHURE DLX? Here is an extract from their website:
QUESTION:
What is the Latency of GLX-D+ Dual Band?
ANSWER
Latency varies slightly depending on the Frequency Band and Group:

Z3, Z4, Z5 (2.4 & 5.8GHz Bands):
  • Group 1: 4.53ms
  • Group 2: 8ms
  • Group 3: 4.53ms
Z2 (2.4GHz band only):
  • Group 1: 4.53ms
  • Group 2: 8ms
  • Group 3: 4.53ms
  • Group 4: 8ms
These are ridiculous values. You should definitely prefer cable or use a high-quality analogue transmitter.
https://www.thomann.de/gb/sennheiser_ew_500_g4_ci1_aw_band.htm

In conclusion, you can only say that it doesn't matter whether a modeller has 3.2 or 4.5 ms. But whatever peripherals are added can ultimately result in a clearly noticeable latency.

That's why I am happy with my FM9 wehen playing with a cable and a pair of RESOUND ELIS 8.
The difference between a true analog setup is, that you have to move the cabinets 1.2 meters closer to your ear. 😄
Click to expand...
I measured the delay, posted it here and it was “around” 5ms.
I am in agreement that adding additional digital delay is problematic. I measured the JBLs I use and they are “around” 5ms like the FM9 and all I can say is it feels better without the extra pa delay(Fender fm12 user).
 
200man said:
Afraid to follow thru on this. Thanks. :)
So, WRT the FM3 we have Clif saying 3.3, G66 saying 2.9 and Clif says the fm3 and fm9 are the same. So, I’m puzzled a bit too.
Click to expand...
Note this from my later post:
It sounds like a) the FM3 test wasn't as precise as the others and b) the FM3 didn't have the same signal chain (no Cab).
Click to expand...
 
Chromatizm said:
Sorry, you cannot "only" say that, that's just an opinion. There is another opinion, but I will not use "my experience" again for this kind of argument, let's hear some experienced people discuss it:
Click to expand...

Sorry, I don't know what you're trying to say ;-)
 
200man said:
Afraid to follow thru on this. Thanks. :)
So, WRT the FM3 we have Clif saying 3.3, G66 saying 2.9 and Clif says the fm3 and fm9 are the same. So, I’m puzzled a bit too.
Click to expand...
The FM3 test was not precise because it was not that important. In any case, I would trust Mr Chase's values.
The FM9 was tested in more detail because a customer contacted me and asked why the latency was higher than with his Kemper. His value seemed a bit too high to me, so I tested it. I measured a lower latency than he did.

Finally, I split a signal and sent it through an FM3 and the FM9. An identical preset was used on both devices.
The signal that ran through the FM3 was faster at the DAW than that of the FM9. See pic.

The test is easy for anyone to reproduce.

Bildschirmfoto 2023-12-17 um 19.59.23.png
 
Keybi said:
The FM3 test was not precise because it was not that important. In any case, I would trust Mr Chase's values.
The FM9 was tested in more detail because a customer contacted me and asked why the latency was higher than with his Kemper. His value seemed a bit too high to me, so I tested it. I measured a lower latency than he did.

Finally, I split a signal and sent it through an FM3 and the FM9. An identical preset was used on both devices.
The signal that ran through the FM3 was faster at the DAW than that of the FM9. See pic.

The test is easy for anyone to reproduce.
Click to expand...

It's easy to reproduce...if you have both an FM3 and an FM9 :). I can't follow what you're saying though. You measured the latency on both an FM3 and an FM9? What latency did you measure for each one? How did you measure it?

By the way, it's best to use a pulse for measuring latency, not a periodic signal like a sine wave. There's no way to tell from your picture which device is lagging the other.
 
Chromatizm said:
Some of us are very sensitive to latency.
It really matters to them.
Click to expand...
Yes sure. I am also very sensitive.
But what I wanted to say is that you don't notice a difference of 0.4 ms.
So if the Kemper has 3.6 and an FM9 has 4.2, nobody can perceive that.

Just to clarify the size ms. Light travels 300.000 km in one second. That is the distance Earth to Moon
In one ms, on the other hand, it travels just 300 km.
I think very sensitive players feel around 10ms.
 
GlennO said:
It's easy to reproduce...if you have both an FM3 and an FM9 :). I can't follow what you're saying though. You measured the latency on both an FM3 and an FM9? What latency did you measure for each one? How did you measure it?

By the way, it's best to use a pulse for measuring latency, not a periodic signal like a sine wave. There's no way to tell from your picture which device is lagging the other.
Click to expand...
Well, I started the recording and then started the sine wave generator at one point. And the sine wave of the FM9 lags behind the FM3. You can recognise this by which wave passes through the x-axis first. I recorded 10 seconds and created a section. But you can trust me. In my setup, the FM9 runs behind.
I then swapped the channels and the FM9 was slower here too.
 
The FM3 and FM9 should be the same. I don't have an FM9 here to test but I'll have the engineers look into it this week. Perhaps there's an extra buffer somewhere in the FM9 that shouldn't be there.

Both products use a dedicated core for amp processing and both products use the same frame and interpolation kernel sizes.
 
Keybi said:
But what I wanted to say is that you don't notice a difference of 0.4 ms.
Click to expand...
Or maybe you do under some circumstances. I don't believe you have a reliable source about this other than an opinion.
And if you can't measure it reliably, it doesn't mean you can't feel it. Sure, you can fool your brain into believing that if it cannot calculate something, it cannot feel it, but that would be a lie.
You brain recognizes distances, phase issues, directions, etc. It knows the speaker is not "10 meters away" from you. It knows if someone's talking 10 cm or 1m away from you. You can fool your "logical" part but not your animal brain.
It was a child game to feel the difference in delay between Helix and computer plugins when you sit in the same spot playing with the same volume, and I wasn't even familiar with the sound of any of it. Sure, not when playing chords or loud chugs, but try fast runs - obvious. Maybe there were some other things at play, so we also tested this by playing though the same PC plugin but changing buffer value to affect the latency.
Keybi said:
I think very sensitive players feel around 10ms.
Click to expand...
Incorrect. In a few blind tests by different people in different rooms I recognized 4ms vs 8ms 100% of time. There is simply no mistake. Probably a volume could fool me, but I'm not playing loud at home, and my fingers are quite fast for me to be very uncomfortable at 8ms close to speakers.
If it's not your experience, it doesn't mean it is the same for everyone. As Chris says: some are very sensitive for this. Not "10 ms" sensitive.
 
Chromatizm said:
Or maybe you do under some circumstances. I don't believe you have a reliable source about this other than an opinion.
And if you can't measure it reliably, it doesn't mean you can't feel it. Sure, you can fool your brain into believing that if it cannot calculate something, it cannot feel it, but that would be a lie.
You brain recognizes distances, phase issues, directions, etc. It knows the speaker is not "10 meters away" from you. It knows if someone's talking 10 cm or 1m away from you. You can fool your "logical" part but not your animal brain.
It was a child game to feel the difference in delay between Helix and computer plugins when you sit in the same spot playing with the same volume, and I wasn't even familiar with the sound of any of it. Sure, not when playing chords or loud chugs, but try fast runs - obvious. Maybe there were some other things at play, so we also tested this by playing though the same PC plugin but changing buffer value to affect the latency.

Incorrect. In a few blind tests by different people in different rooms I recognized 4ms vs 8ms 100% of time. There is simply no mistake. Probably a volume could fool me, but I'm not playing loud at home, and my fingers are quite fast for me to be very uncomfortable at 8ms close to speakers.
If it's not your experience, it doesn't mean it is the same for everyone. As Chris says: some are very sensitive for this. Not "10 ms" sensitive.
Click to expand...
Well then let's agree to disagree. ☺️
When I was studying biology, I asked my professor at the time what latencies the human ear perceives in a tactile control loop.
This is a tricky subject and there will always be different opinions. He told me at the time that the limit was around 10 ms. And there are other factors as well. It is clear that the human ear is able to localise sounds spatially and that the ears are only about 10 cm apart. But this only works with sounds above 1.6 kHz. Below that, people can no longer localise sounds spatially. A fact that is utilised in subwoofers. To conclude from this that you can perceive 4 ms latency is very daring. Incidentally, reflexes are not controlled by the brain. The motor end plate transmits the electrical excitation from the nerve fibre to the muscle. This results in a reaction, the contraction of the muscle. This process happens so quickly and automatically that we don't even realise it. The speed of reflexes is in the range of 25 - 50 ms.

In the case of the guitar player, there is also the duration of the muscle movement, control impulses from the brain, electrical transmission in the axons and chemical transmission in the synapses. This only happens through chemical messengers through the synaptic cleft. And more slowly than through the axion potential. This is because these impulses do not move at the speed of the electric current but only at a maximum of 90 metres per second. The whole thing is a control loop with many points of influence.

But as I have already said. If you are of this opinion, I don't want to argue with you. I don't want to deny that.

If you've ever heard Beethoven's Ninth live, for example, you won't ask yourself how the gigantic orchestra, some of whose musicians sit 20 metres apart, can perform work that thrills the audience. These are latencies in the range of 30 - 40 ms. And that's why they need a conductor. And that's the beauty of music. It's not a high-performance sport.



And if we test all modellers, then a standard latency will emerge, and the differences between the individual products will again be in the 1 or 2 millisecond range.
And that was ultimately my point. If you're happy with 3.6, you won't hear any difference with 4.2.
 
Chromatizm said:
and my fingers are quite fast for me to be very uncomfortable at 8ms close to speakers.
If it's not your experience, it doesn't mean it is the same for everyone. As Chris says: some are very sensitive for this. Not "10 ms" sensitive.
Click to expand...
I only started playing about 7 years ago, when I was in my 40s, and so some combination of age/reduced plasticity or ability to learn/simply not having the time to devote that I would have had as a kid has kept me from developing any real speed - the tempo and unit of measure at which I start to lose the ability to be musical is 16th notes at about 120 bpm, with bursts of 16th note sextuplets at the same tempo. What range are you playing at where you start to feel disconnected to what you're hearing?
 
You must log in or register to reply here.
Back
Top Bottom