FM3 as audio interface?

The 30 sample offset is with the Axe FX III, and it equates to 0.625ms at 48k sample-rate with a buffer size of 64. I don't think that is a big deal at all.

The FM3 probably has a much higher value, and it'll be a bigger deal to some.

But again... if you're monitoring in real-time through the FM3 and not through the DAW.... it isn't a showstopper.

Sorry, I misspoke while reading your comment before; what I meant was that the FM3’s offset is a big deal imo, at least the numbers being thrown around earlier in this thread. Thanks, my mistake :)

And yes agreed it’s not a showstopper, USB Audio is usable, you can manually compensate, etc
 
I never called it unusable. I obviously don't think it's unusable (???). I think it's a good product, like I've said, and I think the USB Audio IS usable (as I said in the first reply you quoted). I think this one thing should be a higher priority because of how that aspect is marketed though, and I thought Fractal's response was shitty considering that. Repeating myself here, yes, but if it was marketed as just a bonus feature, then I think it's fine as-is as this sort of issue is relatively common with many lower-end interfaces.

Also, I didn't say anyone said I couldn't discuss it. I was referring to the people saying "don't use it as an interface" over and over and over again (please see what I was replying to with that comment) or claiming that people "just want to rake Fractal over the coals" (??). It is marketed as an interface, so I think resistance to discussing how it performs as such but replying with "get a proper interface" etc makes no sense. If someone doesn't want to use it that way, that's fine, but other people obviously are interested in that, hence the thread.

And if we are going to be all weird about this, 2 of those quotes aren't of me (the first of which I disagree with, as it certainly is a serious pro product, as the second to last comment of mine you quote specifically says, and a great product too, and the second of which I thought was silly and petty), and 2 of them are quotes of the same comment which aren't even referring to the latency issue at all (so, taken out of context).
Fractals response was they felt their resources were better spent elsewhere at this time. Why is that shitty? YOU want better interface capabilities. I prefer they update the product to stay in line with the rest of the product line in terms of amp and fx models. I’d never use it as an interface. Taking one look at it tells me it wouldn’t serve needs above and beyond jamming along to YouTube.
As I understand it, the chief complaint is its lack of proper reporting back to the DAW. Big deal! Manually compensate your tracks. If you want to be an audio engineer, then ENGINEER. I started on protools v6 20 years ago, it didn’t even have ADC. You had to manually compensate everything. It wasn’t the end of the world. I wish my Apollo X8 interfaces had a few features their competitors have. They don’t. I work around it or work within the limitations of the hardware.
 
In the Fractal thread - the reference to "impressive USB audio" and the marketing was referenced only once but the post creator. The rest of the thread was talking about measuring latency and alternative devices, all very courteous and then Fractal locked it.

Yes people will repeat themselves but most of the responses here are to just buy an audio interface. That isn't an option for some.

I'm talking about this thread being an example of why Fractal closed their thread. If you guys are just repeating yourselves ad nauseam here, the same thing is going to happen there. It's not like this is there first go-round with a group of users unhappy about something.

I know in the last several months since the III has been getting updated, I do not believe I have seen one request pop up here for better latency reporting, but there's a WHOOOOOOLE bunch of -

Still Waiting Office Tv GIF by The Office
Where Are You Waiting GIF by This GIF Is Haunted
Bored Cabin Fever GIF
Hd Reaction GIF by MOODMAN


- in regards to the content being updated.
 
I found this thread originally started out as very good food for thought.
Piling on with an agenda isn’t exactly helpful.
I’m just attempting to make sense of the situation and wrap my head around what was initially discussed.
My knowledge of this is extremely lacking .
 
I do think it was premature and a bit of an optics own goal to close the thread over there at the Fractal forum. But hey.... this is exactly why TGF was setup! So that nobody needs to (hopefully!) be beholden to overzealous moderation and company tactics.
 
Yeah this is what I don't know myself either. 12-13ms is acceptable if the buffer size is 512. It is not so acceptable if the buffer size is 64.

But nobody has actually said anything in detail when it comes to the FM3.

@GuitarJon what latency figures do you get for which buffer sizes? You (or anyone with an FM3) could measure them quite quickly using the Oblique RTL Utility here: https://oblique-audio.com/rtl-utility.php
I just measured the RTL of my FM3 using RTL Util. No blocks in chain other than input and output.

Out 2 L to in 1 L, FM3 USB buffer size at 48 (minimum) in IO settings, M3 Pro (un-binned) MacBook Pro w/16GB Ram.
Sample RateBufferRep RTLMeas RTLMeas RTL (ms)
480001633357311.938
480003236560512.604
480006442966913.938
4800012855779716.604
48000256813105321.938
480005121325156532.604

-Aaron
 
Last edited:
I know in the last several months since the III has been getting updated, I do not believe I have seen one request pop up here for better latency reporting,

I think a lot of people don't know about the problem. They just put up with sloppy recordings without knowing the problem is the FM3 and not their playing. You wouldn't expect to see a problem like that in an audio interface, especially one from Fractal, so it's natural to not think there's a problem in the FM3.
 
Last edited:
In regards to closing threads
My thoughts are still what would be gained by having it opened ?
Generally the purpose of posting in a bug thread is to get the attention of the manufacturer.
Once the company has made a formal statement like they did really what is left to discuss ?
They are just going to have to moderate the hell out of it as both sides are going to get ugly and the end result will be please refer back to post #18 by AdminM
 
I just measured the RTL of my FM3 using RTL Util. No blocks in chain other than input and output.

Out 2 L to in 1 L, FM3 USB buffer size at 48 (minimum) in IO settings, M3 Pro (un-binned) MacBook Pro w/16GB Ram.
Sample RateBufferRep RTLMeas RTLMeas RTL (ms)
480001633357311.938
480003236560512.604
480006442966913.938
4800012855779716.604
48000256813105321.938
480005121325156532.604

-Aaron
Great info, thank you.

I think those are horrible numbers. Since I like the idea of integrating a multieffects with plugins (NAM, basically), this would mean I'd not go for an FM3.

I've heard some time ago that Helix was more or less this latency heavy. It'd be great if some owner could make an RTL measurement too.
 
Great info, thank you.

I think those are horrible numbers. Since I like the idea of integrating a multieffects with plugins (NAM, basically), this would mean I'd not go for an FM3.

I've heard some time ago that Helix was more or less this latency heavy. It'd be great if some owner could make an RTL measurement too.
Yeah. I don't use the AI in it at all. Just the SPDIF or analog outs. My Presonus Quantum 2 gets 1.125ms RTL @48k/16 on the same machine. Reported latency differs only by 2 samples.

-Aaron

EDIT: IIRC my HX Stomp was a little better, but not much. I still never used the built in AI on that either. Frankly, I'd rather just use a dedicated reamp box and an external AI and have the guitar processor just do SPDIF or AES I/O really well. SPDIF in the FM3 is good enough, but having to use it as a clock source is kind of a bummer. Having SPDIF in would be really helpful too.

All this said, I love my FM3. It's literally ALL I use for my guitar ever as far as amps and effects go.
 
Last edited:
I just measured the RTL of my FM3 using RTL Util. No blocks in chain other than input and output.

Out 2 L to in 1 L, FM3 USB buffer size at 48 (minimum) in IO settings, M3 Pro (un-binned) MacBook Pro w/16GB Ram.
Sample RateBufferRep RTLMeas RTLMeas RTL (ms)
480001633357311.938
480003236560512.604
480006442966913.938
4800012855779716.604
48000256813105321.938
480005121325156532.604

-Aaron

Wow, yeah, this is unfortunately quite bad. This is actually worse than some numbers I had seen when I did a quick search to see if anyone else had measured. It would be great to get someone else to test too to see if there are consistent results.

Thank you for doing the test!
 
I'm talking about this thread being an example of why Fractal closed their thread. If you guys are just repeating yourselves ad nauseam here, the same thing is going to happen there. It's not like this is there first go-round with a group of users unhappy about something.

I know in the last several months since the III has been getting updated, I do not believe I have seen one request pop up here for better latency reporting, but there's a WHOOOOOOLE bunch of -

Still Waiting Office Tv GIF by The Office
Where Are You Waiting GIF by This GIF Is Haunted
Bored Cabin Fever GIF
Hd Reaction GIF by MOODMAN


- in regards to the content being updated.
I wouldn't necessarily expect specific issues to be raised here as they can be raised as a bug on the Fractal website and are more likely to be picked up by the Fractal team then here.

I think GuitarJon was just asking a general question which people will share their experiences.
 
Wow, yeah, this is unfortunately quite bad. This is actually worse than some numbers I had seen when I did a quick search to see if anyone else had measured. It would be great to get someone else to test too to see if there are consistent results.

Thank you for doing the test!
Yeah. It's entirely possible I did it wrong too. Lots of weird settings on the FM3 for what gets sent where. :wat

-Aaron

EDIT: Pretty sure I did it right. Basically just measuring at input 1 using direct monitoring setting from Out 2. So RTL Util sends tone out of Out 2 and measures it at input 1. I tried it with some blocks patched in sampling from the OUT 1 block and the various block types increased latency as expected. So my numbers above are best case scenario Out 2 DA to In 1 AD with no blocks on the grid.
 
Last edited:
I just measured the RTL of my FM3 using RTL Util. No blocks in chain other than input and output.

Out 2 L to in 1 L, FM3 USB buffer size at 48 (minimum) in IO settings, M3 Pro (un-binned) MacBook Pro w/16GB Ram.
Sample RateBufferRep RTLMeas RTLMeas RTL (ms)
480001633357311.938
480003236560512.604
480006442966913.938
4800012855779716.604
48000256813105321.938
480005121325156532.604

-Aaron
Yeah that is a bit cack for sure.
 
I wouldn't necessarily expect specific issues to be raised here as they can be raised as a bug on the Fractal website and are more likely to be picked up by the Fractal team then here.

I think GuitarJon was just asking a general question which people will share their experiences.

No one is saying not to share them, but as someone who has worked in service-based careers my entire working life, when you get people who aren't accepting the current answer and are just repeating themselves over and over, there's nothing you can do to help them out. You can even say, "You're right, I agree with you. That shouldn't be the way it is. But it is what it is right now".....and then......."Well now I'm going to repeat myself again because I am not satisfied"

I did it with people over $15 steaks at Chili's and I now I do it with multi-million dollar leases, it's the exact same mentality and I handle them the exact same way.

If Fractal had a history of not taking care of things, I'd be joining in with ya. I have a big flag I like to wave around that says "Don't market shit your ass can't cash" And by all means, the number reporting can't be anything less than helpful for all involved, I'll move them to the front page if you guys want to track them better. At the very least it helps people adjust their units in the meantime and gives Fractal a good view of numbers/situations to look at.

And while you're right, reporting things at the Fractal forum is always going to be the best way to reach them, Cliff is quite active here and there's been a good amount of discussion here that's made it's way into some recent updates.
 
FWIW I just re-"measured" my FM3T (following this post's instruction) with my main PC and got a consistent 298 samples offset latency (~6.2ms).
Same offset worked within all ASIO buffer sizes I tried, only varied with the Fractal global setup USB buffer size parameter (set to 48 here).
Latency was also the same whether the preset was a pure In->Out loopback or 84.9% CPU usage one, numbers also consistent within reboots, cold boot and longer (2 hours) session.

It's definitely better than the last time I checked (~12ms offset) so not sure what to think of it.
My testing was with a click track in my DAW as RTL Utility crashes on startup for me. :unsure:
 
Last edited:
No one is saying not to share them, but as someone who has worked in service-based careers my entire working life, when you get people who aren't accepting the current answer and are just repeating themselves over and over, there's nothing you can do to help them out. You can even say, "You're right, I agree with you. That shouldn't be the way it is. But it is what it is right now".....and then......."Well now I'm going to repeat myself again because I am not satisfied"

I did it with people over $15 steaks at Chili's and I now I do it with multi-million dollar leases, it's the exact same mentality and I handle them the exact same way.

If Fractal had a history of not taking care of things, I'd be joining in with ya. I have a big flag I like to wave around that says "Don't market shit your ass can't cash" And by all means, the number reporting can't be anything less than helpful for all involved, I'll move them to the front page if you guys want to track them better. At the very least it helps people adjust their units in the meantime and gives Fractal a good view of numbers/situations to look at.

And while you're right, reporting things at the Fractal forum is always going to be the best way to reach them, Cliff is quite active here and there's been a good amount of discussion here that's made it's way into some recent updates.

I don't think anyone is too interested in the back and forth anymore, but rather just the info re: latency and latency reporting, but to be clear: for me the bad part about their response was that they didn't actually say they will fix it in the FM3, but instead said that addressing it "would need to take effort away from other areas".

So I don't think it's a matter of them lying or something like that, not at all, but just of it being too difficult of a problem for them to guarantee a fix in the FM3 at all, even though they'd like to fix it (of course).

Fractal seems like an honest company to me and they have a great track record of addressing issues, of course, and again the FM3 is a high quality pro product, no doubts there. So I hope a fix comes, but the response seemed to imply that it may not or at least is not a priority.

Here's the quote again for convenience:
"We are of course aware of the USB latency reporting complaint. We understand and empathize that this issue and required workarounds have made things more difficult for some of you. However, the FM3 team has been dedicated to keeping up with Axe-Fx updates for as far back as I can remember. It requires a lot more work for the "little brother" to inherit features from the powerful Axe-Fx III, and addressing the USB fix is would need to take effort away from these other areas. The vast majority of customers would prefer that we continue to add new amp models, new effects, new sound features, and more -- and even clamor on a regular basis for these to be done already. And this isn't like "dusting crops": the FM3 is a serious professional product which needs to be programmed and maintained with the utmost care. We will keep you updated on any developments in this area and maintain our commitment to high quality products"

I think everyone here just wants to know the situation since many people, including the OP, do use it as an audio interface or would be interested in using it in that way in at least some contexts, and that everyone here, whether they own it or not, thinks the FM3 is a great product and that Fractal is a good company.

It's great that we have some numbers now that confirmed suspicions both about latency reporting and actual RTL and, through the thread Orvillain linked to, can directly compare them to other audio interfaces too. Hopefully we get more to see if they're relatively consistent across setups.

I know modding can be rough, so while we disagree here, thanks for your work, I appreciate it.
 
Back
Top