Recording Shoot-out: I made a mistake!

[Opz]

Ah, I discovered what went wrong. I'm still learning, I can admit when I made a mistake.
So at 96 kHz/32 bit they're the same indeed. After conversion to 44.1 kHz they're not due to conversion quality difference?

Cheers, Opz

I would like to know if anyone else noticed it, but I'm interested to see your current (2012) views on this subject which has passed this forum more than once before.

I usually record my friends and clients audio in Ableton Live using a MOTU 828mkII (Firewire) at 44.1 kHz/24 bit (recently 32 bit) WAV.
When I recorded a beat which came out of an Akai MPC2000 two days ago I decided to try to record at higher sample rates etc, so I switched my interface to 96 kHz and recorded the beat into ProTools 10 (which I had never done before as I only recently acquired PT10). The beat was made up of 7 channels coming from the MPC2000.
When I was done recording the beat at 96 kHz/32 bit in PT10, I also recorded the beat in Ableton Live how I would normally do, at 44.1 kHz/32 bit. Just to have a 'back up' of the recording the way I would normally record.

Well, the difference between the recording in PT10 and Live is extremely audible!
But then I figured I had used a twice as high sample rate in PT10, which might explain the sonic differences between the recordings.
So I recorded the same beat a second time in Live, but this time at 96 kHz/32 bit.
Comparing the two 96 kHz/32 bit recordings of both PT10 and Live, there still is a very noticeable sonic difference.
The recordings of PT10 are far more open, balanced and bright as compared to Live's recordings which sound squeezed, slightly more dull and has some sort of boxiness to it.
I had other people do the same comparison between the two recordings and they all agreed that Live had the lesser recording of the two, so I wasn't just me.

Thanks for any comments, pointers etc.
Cheers, Opz

[luddy]

Hi, can you post some samples?
-Luddy

[nylarch]

When you play back in Live do you have "auto warp" and "auto fade" turned off?

[tigali]

I'm going to duplicate this experiment and see what happens. Cheers.

[tigali]

Ok, done. And on my very sweet Adam ANF-10's I could hear NO DIFFERENCE.

Either my ears aren't as sensitive as yours or there simply is not any difference. If I had heard any difference I'd be inclined to try out phase cancellation tests and all the rest but as this has been done before and as I can hear no difference I think it would be a waste of time.

You might want to check your preferences between both DAW's and see what might be set up differently. I set them up exactly the same and recorded the same drum machine (Rhythm Ace FR-2L) through the same interface at the same sample rates and exported with the same bit rates, etc. No audible difference.

[3dot...]

I don't have PT.. but used Reaper instead..
same !

[Donnie]

The recordings of PT10 are far more open, balanced and bright as compared to Live's recordings which sound squeezed, slightly more dull and has some sort of boxiness to it.
I had other people do the same comparison between the two recordings and they all agreed that Live had the lesser recording of the two, so I wasn't just me.

Thanks for any comments, pointers etc.
Cheers, Opz

Turn off Auto Warp, seems to be the problem every time the argument comes up. I'd be very surprised if that wasn't the issue.

[hacktheplanet]

Here's a test Opz:

-Render the same loop in Live and also in PT
-Load both loops into a different DAW, make sure they are identical and line them up.
-Flip the polarity of one of them.
-If the two clips cancel each other out, there is no difference in Live and PT.

[thunderkyss]

I remember back in the day, this issue came up on the Sonar forum. Then the question was Sonar vs Cubase vs Pro Tools.

It was actually a topic that stayed alive for several years & I guess the cooks at Cakewalk figured it wasn't going to go away.

So they did their own "laboratory" study & found that the "summing engines" of each DAW did in fact color the sound. The thing is that some people liked Cakewalks algorithm better & others liked other algorithms better, so Cakewalk added an option that allowed you to change how your mix would be summed up.

I personally thought it was a waste of time. We're working with digital audio, as long as what is recorded is accurate to what I'm hearing, nothing else should matter. If I want the mix to be more "open" then it's up to me as the engineer (I'm not really an engineer) to figure out how to do that.

If I can't get the sound I want out of my DAW, then it's time to move on.

It doesn't mean that Pro Tools (or Sonar for that matter) is better than Live, or Reaper, or Rose Garden... just different. If 9 out of 10 people prefer a Pro Tools mix, I'm fine with that. But chances are, it was the artist (read engineer) behind the console that made the difference.

[regretfullySaid]

Forgive me if I'm ignorant about this, but is there any relevance about recording in 96khz when the source material isn't even 96khz? Does that MPC even support 96khz samples IF the samples were 96khz? Then where does upsampling come in? Does this have anything to do with the recording/playback differences in the hosts? Unless that's all irrelevant, isn't this part being overlooked?

[ttilberg]

So they did their own "laboratory" study & found that the "summing engines" of each DAW did in fact color the sound. The thing is that some people liked Cakewalks algorithm better & others liked other algorithms better, so Cakewalk added an option that allowed you to change how your mix would be summed up.

I'd buy this. I'm betting that the actual individual channel clips sound exactly the same, but when processed together may sound a little different from one to the other...

This is an excellent point to bring up. I have a very hard time thinking that Live inherently mangles your audio recordings, but could certainly believe that the processing it does to give you a final product is going to be different than other tools.

[Opz]

To answer some questions and comments first:

@the_planet:
I'll try that polarity/phase cancellation test you describe as soon as I have the time. But hearing the difference of the rendered files below I can't imagine full cancellation taking place.

@shadx312:
The MPC2000 delivered 8 analog signals using unbalanced jack outputs.
It's irrelevant what happens inside the MPC on a digital level, it's outputs are analog.
I wanted to try and capture that analog output to the best degree my equipment can handle, which is at 96 kHz.
The only thing that takes place is A-D conversion inside the MOTU 828mkII.
So when recording in different hosts, the result comes down to the host because what I hear being recorded/played is taken from the same source, (the AD converter of) the 828mkII.
It's not even relevant what the quality of the AD converter is, as it's quality/result stays the same for all recordings using different hosts.

In Live's Preferences Auto Warp and Auto Fade were (and have always been) off during recording.
BUT, I discovered that even though those settings were off that in Clip View in the 'Sample/Clip box' Warp was activated on every file even though the file had no warp end marker.
It only had a warp marker at the beginning even though no warp mode was selected.
Why does that happen? Quite strange if one asks me. Maybe a new 8.3 bug, because I haven't noticed this behaviour before.
So for the render test described below I made two versions. One has a 'W' at the end of the name, the other 'NW'. 'W' means Warped and 'NW' means Not Warped.

Upon recording, all inputs were recorded with as much gain as possible before clipping the inputs of the 828mkII.
All needed gain was supplied by the MPC2000 and no gain was added in the hosts or the 828mkII interface.
During recording the only thing that matters is no clipping on the tracks being recorded. I start the host, set up the tracks for recording and that's it when it comes to recording.
During rendering my recording results for this post here, both hosts have all faders of the audio tracks and the master bus at unity, as was the case during recording.

I used the 96 kHz/32 Bit Float PT10 recording and rendered a small part of it as 44.1 kHz/32 Bit Float WAV.
Then I used the 96 kHz/32 Bit Live recording and rendered a small part of it as 44.1 kHz/32 Bit WAV. (Live does not indicate whether or not it is 'Float').
I also made a recording at 44.1 Khz of which a small part was also rendered as 44.1 kHz/32 Bit WAV.
The PT file is smaller in size due to a difference in length of the loop I quickly made.
So I ended with the following results upon bouncing in both PT10 and Live, the links are shortened Dropbox urls:

I don't know why but the Live recordings are far louder, more squeezed and boxed and less bright and open.
I monitor using Adam A7X's with balanced audio coming out of the 828mkII, going through a passive monitor controller (SM Pro Audio M-Patch 2) to the A7X's.
One might argue about rendering/bouncing differences when it comes to conversion, dithering etc, but the fact is that this audible difference is noticeable on my setup during playback after and during recording and before bouncing.
I can't explain the difference as the audio passed on from the inputs of the 828mkII does not change at all during use with both hosts.
Both hosts just seem to tap in and handle that audio differently. Both hosts have all settings the same as far as possible.

I just don't like the result Live gave me. When it comes to producing my own music, I will still use Live. I love that piece of software and it's workflow.
But from now on, when it comes to recording other peoples music and mixing those projects or my own I will use Pro Tools.
That's not even a shame when it comes to Live as it perfectly fills my need for my own music production using a MDP32.
I already bought PT to start doing my recordings and mixes in that DAW. This recording result just makes me move to it faster for those purposes.
I also bought Wavelab for specific editing reasons, project (album cd etc) buildup and a bit of mastering.
Each tool has its own strengths and weaknesses. One of the main features of PT10 was Clip Gain. But Live has had it for years already. So Toeachizown

[regretfullySaid]

Thx for clarifying.

[agent314]

You're doing something different between the two when exporting, because the PT is peaking at around -1.1db and the Live version is clipping throughout

[regretfullySaid]

Might check them out in a spectrogram later just out of curiosity.