Axe as audio interface... Flexible sample rate?

Wolfenstein98k

Wolfenstein98k

Power User
Would it be possible for the sample rate that the Axe puts out (48kHz) to be alterable by the user?

I am buying an audio interface for recording soon, and would like to run my projects at 96kHz - but dislike the idea of having to run an analog out to this interface, to then convert the signal for a third time just to get it into a project at the desired sample rate.

Is this an Axe III wish, or can it be altered in the existing Axe II?
 
Wolfenstein98k said:
I dislike the idea of having to run an analog out to this interface, to then convert the signal for a third time just to get it into a project at the desired sample rate.
Click to expand...

Unless you have unbelievably great (golden) ears, you won't hear a difference. I certainly can't. That said, I can guarantee you won't hear a difference between connecting the Axe via S/PDIF to your interface and using the Axe as an interface.
 
Axe II sample rate is fixed at 48 kHz 24 bit. Many DAWs can work in mixed rates and resample media on the fly. Real time resampling does cost CPU though, so its often better to convert it offline. In Reaper, you just right click the clip and select "Render item as new take" to create a new rendering of the clip in the project's set sample rate.
 
Ah, thanks guys.
mr_fender said:
In Reaper, you just right click the clip and select "Render item as new take" to create a new rendering of the clip in the project's set sample rate.
Click to expand...
Wouldn't that introduce quantisation noise? Or is that only a concern for downsampling?

Jason Scott said:
That said, I can guarantee you won't hear a difference between connecting the Axe via S/PDIF to your interface and using the Axe as an interface.
Click to expand...
Yeah, I was thinking of doing that, only I wanted to be able to buy a 2i2 which doesn't have S/PDIF unfortunately!

My concern here wasn't so much that I could hear the difference, but that constantly re-sampling seems like a bad way to do it.
 
Jason Scott said:
To me, it's only bad if it produces an undesirable result that's audible. I mean, if you can't hear it...
Click to expand...
Yeah, I just find avoiding introducing "errors" etc where possible streamlines trouble-shooting later. If I find I need ten times the dither, at the end of the project with downsampling, then... :p
 
Wolfenstein98k said:
Wouldn't that introduce quantisation noise? Or is that only a concern for downsampling?
Click to expand...

Yeah I think quantization noise is a bigger deal when down sampling because the lower bit depth will have a relatively higher noise floor.

Wolfenstein98k said:
My concern here wasn't so much that I could hear the difference, but that constantly re-sampling seems like a bad way to do it.
Click to expand...

You'll be resampling either way unless you use digital inputs and your project is 48 kHz 24 bit. When recording the analog output of the Axe II, The AD converter of your chosen interface will be resampling the analog signal so it likely will no longer be a sample accurate copy anyway. You're taking a digital source, converting it into an analog approximation and then resampling it back to digital. That process will likely introduce it's own errors and noise regardless of the new sample rate chosen.
 
Wolfenstein98k said:
Yeah, I just find avoiding introducing "errors" etc where possible streamlines trouble-shooting later. If I find I need ten times the dither, at the end of the project with downsampling, then... :p
Click to expand...

Since the Axe is locked at 48 kHz, why not record at 48 kHz rather than downsample from 96 kHz?
 
Jason Scott said:
Since the Axe is locked at 48 kHz, why not record at 48 kHz rather than downsample from 96 kHz?
Click to expand...
That's what I'll do, if I must (due to the Axe limitations) - I would just prefer to work at 96 if I could.

We could be working at 16-bit with the same arguments being made to defend the limitation... And then 44.1kHz, and then a lower bit depth, etc...

I guess I'm just wondering why this particular thing is limited when very little else about the Axe is! But of course it's a fine result anyway, I'm really just splitting digital hairs here :D
 
Wolfenstein98k said:
We could be working at 16-bit with the same arguments being made to defend the limitation... And then 44.1kHz, and then a lower bit depth, etc...
Click to expand...

Recording at 16-bit may be perfectly fine depending on the material, though I prefer recording at 24-bit due to the increased headroom and dynamic range. If you ask me, there comes a point when higher sampling rates can become an exercise in diminishing returns, however I personally consider the quality of the converters to be more important.
 
Wolfenstein98k said:
That's what I'll do, if I must (due to the Axe limitations) - I would just prefer to work at 96 if I could.

We could be working at 16-bit with the same arguments being made to defend the limitation... And then 44.1kHz, and then a lower bit depth, etc...

I guess I'm just wondering why this particular thing is limited when very little else about the Axe is! But of course it's a fine result anyway, I'm really just splitting digital hairs here :D
Click to expand...

Because that way they can design their output section to have less jitter; With a fixed clock you can use a high quality clock source (you don't need to have dividers and PLLs to drive your DACs. You can also place the clock closer physically to the DACs all to reduce jitter and noise.

Apart from appart from introducing jitter from a variable clock source, supporting different rates would mean either mean, the Axe-FX operates at its own internal rate (which would probably stay at 48KHz) and would have to resample in H/W to whatever rate you selected or all the effects/amp emulations/cab effects would have to operate at the user selected rate.

In the first case, the H/W resamplers are inferior to offline software based resamplers - so operating at 96KHz could actually be worse quality wise.

In the second case, the first effect of operating at a higher sampling rate would be higher CPU usage. Also IR packs would have to be captured at a variety of different sampling rates (and the axe-fx would have to store copies at different rates or preset switching time would have to increase maybe so the axe-fx would downsample IRs depending on rate, etc).
 
Unless your final mix is for a DVD-Audio or Blu-Ray Audio format, your project is going to be downsampled and most likely bit reduced at some point anyways.

I work in 96 KHz to future proof my stuff. I find ASIO4ALL works fine for on the fly conversion.

YMMV
 
Wolfenstein98k said:
That's what I'll do, if I must (due to the Axe limitations) - I would just prefer to work at 96 if I could.

We could be working at 16-bit with the same arguments being made to defend the limitation... And then 44.1kHz, and then a lower bit depth, etc...

I guess I'm just wondering why this particular thing is limited when very little else about the Axe is! But of course it's a fine result anyway, I'm really just splitting digital hairs here :D
Click to expand...

Hi Wolfenstein,

Don't beat me up....but honest, you have to believe me...or do some tests. There is no reason to record at 96 unless you are recording orchestral instruments that are not sonic. Don't buy into the hype...honest when I tell you, 24/48 for rock, blues, dance, R&B, country will work beautifully.

Here's the deal. If you have an interface that is not smoking hot, you WILL hear a difference at a higher sample rate. For example, on my stuff here....RME FF800, Echo Audio, Lynx, Midas M-32 etc....I can hear no difference from 24/48 all the way up to 24/96 on the interfaces that support that. The reason being? My converters are smoking. You shouldn't be able to tell a difference when using a good interface. That said, on like say, an M-Audio or a Creative Labs...something more consumer and affordable, you WILL hear a difference using the higher sample rate....but it's actually due to the the interface just not delivering the goods.

I did a mobile recording at a school one time. We recorded at 24/48 and they wanted to try this song they were just learning just to see what it would sound like. I decided to switch to 24/96 for that song. They played about 2 minutes of it and screwed up. This made me switch back to 24/48 and they played it again.

Comparing the versions when I got home, 24/96 was a little brighter if anything. I made sure to let my girl press the play button while trying to confuse me. Unfortunately, I could tell which was which due to performance differences, but there was a little more top end in the 24/96. The reason being? Mic's and nothing in the room was "sonic" like a distorted amp or electric bass.

Some VSTi companies claim rendering their plugs at 96 makes a difference. I say it's all hype and BS. I've done so many tests with this, it's nuts. I will never see the benefit or the justification of additional hard drive space and the taxing of my system for something that barely makes a difference.

We are digital at 24/48. Digital plays back what is fed into it. Now that said, I can tell the difference between 16/44 and 24/48....but again, it's due to interface processing. I did a rough project for a client one time just to send them a quickie idea. I did it on a Realtek soundcard using ASIO4ALL on one of my net computers. Something about that stupid interface and my guitar sound was magical. So much so, when I brought the project into my good system, I could NOT reproduce that guitar sound to save my life. LOL! The reason? The converters literally played a GOOD role in how that particular guitar sounded. Sort of like a tape machine coloring the sound. LOL!

As for the Axe Fx as an interface, I put it up against anything else out there. I can't tell a difference when using it. The only issue I have with it, is reamping when using it as an interface. Anything in your project, gets reamped too if you literally use it as an interface. This makes it difficult to choose guitar sounds when in the context of your mix. Now, there could be a work-around for this of course, but I'm not aware of one. So, for reamping, I use a different interface via SPDIF in/out to my Axe Fx and this way I can listen in real time without affecting my entire mix.

Anyway, don't put too much stock into that 96 thing man...honest when I tell you, unless you have a bad interface, you shouldn't be able to hear a blatant difference unless you are recording non-sonic instruments. Even there, the difference is so subtle, it's not worth it. But, that's just my opinion and experience over the years.

If I noticed a difference to where it left me saying "oh baby, I so gotta use this!" because it sounds like the difference between 1 inch tape at 15ips and 2 inch tape at 30ips, you better believe I'd use it. :) I'd be willing to bet Cliff would have added a 96 or above option as well as you know he never cuts corners. ;)

At any rate, if it does work for you, then hey, disregard what I'm saying here. I'm just sharing in hopes you may at least consider abandoning the hype. I've been at this recording stuff since dinosaurs walked the earth...lol...I don't know everything, but in all my years, I've never heard an obvious difference. I've written columns, teach recording classes, help others on forums....we've tested this so many times, it just never makes the difference people think it does.

-Danny
 
Great posts guys, thank you so much. @AlbertA made it clear why it's limited in terms of hardware, which makes sense.

To @Danny Danzi great post - I more or less agree with you, barring once exception:
I've read really good arguments for keeping all VSTs etc and processing within a project at a higher sample rate, simply to ensure an audio is rendered at the maximum resolution.

That way if you trim the whole project to 16/44.1 at the end, it's only one downsample + dither, with minimal digital "crud" added in the production process. That's why I asked my OP question originally.

Good thread all, and thanks again!

EDIT: here's a good analogy I found on another forum regarding why you'd bother recording and producing at a higher resolution than 16/44.1:

"Why film a movie on film or in HD if it's just going to get converted to 480i when broadcast on your TV?
...
The answer, if you are stumped, is because the film editors, visual effects artists and colour correction can make the hi res version look much better than if the film was shot at 480i. So, after they greatly enhance the hi-res version, when they convert to a lower resolution you get a much better picture than if you filmed at 480i, and then tried to add effects and colour correction to that low res image."
 
Last edited:
Yep totally with you. Just something to consider.....

Math to me has no place in recording other than counting within a song. Especially when the incorrect math cannot be audibly heard.

Some of the best recordings on earth were done analog. What do you think the bit and sample rate would be on a.Doobie Bros album....or Floyd, or early Van Halen? probably less than 16/44 with math problems everywhere, if there was a way to calculate vinyl.

Please keep in mind, I'm not trying to argue with you or sway you to my way of thinking. I'm just sharing my years of experience with you and how using your ears will help you from buying into hype.

We got people buying interfaces that go to 192 because they think it's needed. Try things yourself and see what your ears tell you.

Record something for 30 seconds to a minute at 24/48. Play something simple and play as accurately as you can.

Then change to 24/96 and rerecord the entire project again....the same 30-60 seconds.

Export the audio in each, name one audio a, the other audio b and let someone else press the play button on each so you have no clue which is which.

Take a piece of paper and let whoever is pushing play take a piece of paper so they can secretly write down what they played. As you listen, you write down which you feel is which.

If you can hear the differences without performance giving it away, by all means use 96 if it makes the difference better for you.

My girl was slick when she did it...she played the same file twice and really screwed with me.

Anyway, use what works. Nothing but respect brother. Just a guy with a busy studio sharing what happens with most people I deal with in the industry. In all my years doing projects, I've had 3 clients total use 24/88 and 24/96. None of them had mind blowing audio based on the sample rate. But if it works....God Bless. :)
 
There are several controlled tests that seem to confirm 44.1kHz/16-bit audio is indistinguishable from DVD-A/SACD content during playback. One such test (the paper for which was published on the Audio Engineering Society's website) was performed by the Boston Audio Society. To summarize, in 554 trials, the listeners, which consisted of professional recording engineers, students in a university recording program, and dedicated audiophiles, chose correctly 49.8% of the time. In other words, they were guessing.

That said, for recording, 24-bits can be important for keeping quantization noise, associated with effects processing and other operations, in check.
 
Last edited:
If you don't have a cooperative girlfriend, you can actually do real blind testing on your own.

The blind test plugin here https://hofa-plugins.de/en/plugins/4u/ will anonymise and randomise tracks and let you do several rounds of assessments before showing you your statistical results. Quite cool actually.
 
Wow. I just totally disagree. I can hear the difference between 96k and 48 easily. I always record at 96. I record the Axe Fx analog into my DAW. Easy.
 
You must log in or register to reply here.
Back
Top Bottom