Reaper's Sound

[nicholas]

Originally Posted by plush2

We win a free pie for this thread.

Or how about maybe this?
http://sonicfinger.com/DeadQuietenator.html

[PAPT]

Originally Posted by Jim Roseberry

If you do a dither "shoot-out", where you take say a snare drum hit running thru a reverb (to create a nice long decay), you can most definitely hear a diffence between different dither algorithms.

I did a shootout a good while back (Pow-r, UV22HR, IDR, etc).
Some algorithms are quieter than others.
Some algorithms produce smoother results than others.
Some algorithms produced dither noise that irritated my ears.
In short, there was actually a lot of difference between the algorithms.
Granted, I was zero'ing in on the above decay... where it was isolated and the gain was cranked... but you could definitely hear a marked difference between most algorithms.

In the end, I liked the results of using Pow-r 1 and UV22HR the best.
The dither was quiet... the resultant decay was really smooth... and no irritating frequencies caused by noise-shaping.



Jim Roseberry
www.studiocat.com
jim@studiocat.com

If you turn it up ridiculously loud enough to hear the very end of the tail as it dies out around -90 db. And then if the next track starts up you'll be permanently deaf.

[dub3000]

Originally Posted by plush2

Yes! We've hit the trifecta, daw summing, dither and analog vs. digital. We win a free pie for this thread.

we all get an extra pie EACH if someone brings up 192khz vs 44khz mixing, and free strawberries for dessert if someone who has never heard a SACD before starts demanding DSD editing capabilities for reaper.

...

i only ever hear the points made by the OP from people who are always mysteriously far too busy to do a controlled test between different DAWs.

there are differences in native dither and various other things (64bit fp numerical rounding direction was raised once?). but you know, if you could actually HEAR those differences at normal listening volumes (and presumably if you knew how to mix) you'd be able to change your mix to compensate for it. but since there are no actual serious differences, you just get people spreading misinformation about stuff they don't really understand and are too lazy to investigate.

full respect to anyone who's bothered A/Bing stuff though. and yes, i have been able to construct really weird tests where samplitude, pro tools and reaper behave slightly differently (e.g. by clipping the hell out of 16 bit files and rendering without dither, etc). but the residue was still down under -90dB. it's not why your mixes sound congested, really.

[Geoff Waddington]

Originally Posted by l0calh05t

You have a 12GHz oscilloscope? I would be quite surprised if you did, and even then that would only be enough for the first two harmonics of that square wave

LOL, full points for knowing enough to catch that.

As to your question, yeah I bought it right after I got my RADAR A/D/A converters

[chip mcdonald]

Is everyone referencing the same "playback sample mode"?

Is everyone using the same "resampling mix mode"?


Because NOT in theory, none of those should yield the same results, either on mixdown or during monitoring.

[dub3000]

Originally Posted by chip mcdonald

Is everyone using the same "resampling mix mode"?

Because NOT in theory, none of those should yield the same results, either on mixdown or during monitoring.

that stuff doesn't have any effect (as in, that code never runs) unless you've imported audio into your session that's at the wrong sample rate. you can tell if this is the case if there are little 'i' icons on your media items.

[Fabian]

Originally Posted by nicholas

Or how about maybe this?
http://sonicfinger.com/DeadQuietenator.html

Was used on this https://www.youtube.com/watch?v=hUJagb7hL0E
Great piece of... eeh... music...

[l0calh05t]

Originally Posted by Geoff Waddington

LOL, full points for knowing enough to catch that.

As to your question, yeah I bought it right after I got my RADAR A/D/A converters

I actually got to use a 12GHz Agilent oscilloscope a few times, so I should know that kind of thing. They are pretty expensive...

[DBMusic]

Including human perception in a test will introduce an uncontrolled variable that will skew results. I know I'm WAY over my head here, but I would be very curious of the mathematics involved. Every fader adjustment, every DSP process introduces mathematics. With every fader adjustment, every summing process, every DSP application, are the exact mathematics applied to all DAW's? Or are there variations in algorithms that would result in less exact calculations which hypothetically could be attributed to perceptive differences?

The perception of audio differences between DAW's that initially struck me enough to prompt this thread were more a result, I believe, of accumulated processes. The deeper I got into the mixing process (multiple fader adjustments, plugins and various adjustments therein, etc.) the less ability I had to get the mix I wanted out of Sonar. So while admitting I know nothing of how mathematics is applied to the summing of digital audio, I would question if those processes are absolutely constant across all DAW's.

Regards,

DB

[Geoff Waddington]

Originally Posted by DBMusic

Including human perception in a test will introduce an uncontrolled variable that will skew results

In a general sense you are spot on !

The trick is to isolate the phenomena through repeated double blind listening tests (truly repeatable human perception), and then try through measurement to discover what is indeed different from a technical viewpoint.

This is a really tough thing to accomplish, especially with subtle effects, and even more so with subtle cumulative ones.

[JHughes]

Indeed. It sounds like some people here think humans shouldn't be involved in the judgment at all, which is hogwash. Humans are the ones hearing differences, not machines.

Some others here seem to think it is incumbent upon those who hear a difference to then prove it somehow, which is also hogwash.

A theoretician does not have to prove anything, only come up with theories. Then you have other people who are good at designing experiments to prove the theory. Then you have others who are good at designing equipment to enable the experiments. You have still others skilled at building the equipment, and so it has always been. Einstein never proved his theory of relativity. That didn't make him lazy or stupid, or a believer in black magic or a victim of the placebo effect. And because YOU don't hear what I hear doesn't mean I'm wrong, and also doesn't mean I have an obligation to prove what I'm hearing to you. If you have a compulsion to figure it out, follow your muse but leave me out of it.

And to those dogmatists who swear a summing test disproves all of this...

Joe: My Corvette gets better mileage AND is faster than your Chevette.
Bill: No way, you can only get so much energy from a gallon of gas, the law of thermodynamics proves it. And both cars are doing the same thing.
Joe: Nevertheless, I'm telling you I get more mileage and go faster.
Bill: Look, both cars do the same thing. They ignite fuel in an enclosed space and send that energy to the wheels.
Joe: Still...
Bill: Well, then there must be some differences you need to fix before you make a valid comparison.
Joe: Huh?
Bill: Make sure both cars weigh the same, have the same rolling resistance of their wheels, have the same body profile to account for wind resistance, and have the same size motor. Then do your test, and you'll see that both cars get the same energy from a gallon of gas.
Joe: But I don't want to drive a Chevette!

As you can see, the guys above are arguing apples and oranges, let's be careful not to do so here. I hear a difference, you tell me to eliminate all differences and then I won't hear a difference any more, which defeats the purpose. I don't need to configure every last detail of Reaper to make it exactly like Sonar, so that I can prove they both use exactly the same summing equations, I'll take the programmers' words for it. And that's not the point anyway.

[Lokasenna]

Originally Posted by Diogenes

Step 3 is where the wheels fall off...

You have introduced the human element. If you have a hundred test subjects, you are liable to get a hundred different responses.

Not quite. Doing it as a double-blind study removes that human element, and boils it down to the simple question of whether or not people can reliably tell which is which - without looking. Nothing subjective there.

Originally Posted by JHughes

I don't need to configure every last detail of Reaper to make it exactly like Sonar, so that I can prove they both use exactly the same summing equations, I'll take the programmers' words for it. And that's not the point anyway.

Actually, it is. The claim behind this entire topic is that one DAW can sound better than another due to how they sum things, tiny rounding errors, etc. "Better" being a subjective term, we're left with looking to see if there are simply any difference between them. If the original hypothesis is true, a null test (or double blind listening, if you insist) will find it, but you can't perform such a test unless the DAWs are set up to be as identical as possible. The various settings about dithering, pan laws, sample rates, etc have to be matched or they'll immediately give you a false-positive.

To continue your analogy, of course a stock Corvette is going to perform differently than a stock Chevette. You can't say that either is simply "better", though, without specific criteria and eliminating anything else that might affect the results. One car weighs more? The other car has an open convertible top? Yeah, your gas mileage is going to be off when you compare the two.

[Jim Roseberry]

Originally Posted by PAPT

If you turn it up ridiculously loud enough to hear the very end of the tail as it dies out around -90 db. And then if the next track starts up you'll be permanently deaf.

Quantization noise sounds nasty (no matter how subtle).
Knowing that there are numerous dither options and having access to them... uhh... I'll take the dither that produces the better sounding result. ;-)
You can argue indefinitely about how subtle the final difference, but the bottom line is that with better dither, the issue is put to rest.
Similar to using 64Bit summing, where "rounding error" is put to rest.


Jim Roseberry
www.studiocat.com
jim@studiocat.com

[karbomusic]

Originally Posted by PAPT

If you turn it up ridiculously loud enough to hear the very end of the tail as it dies out around -90 db. And then if the next track starts up you'll be permanently deaf.

Of the few samples I have examined, ridiculously loud and permanently deaf were not a factor. I only needed two scenarios to occur... Loud but not crazy loud and a spot or two in the material with low dynamics or silent spots and it was easy to hear. My only point is that there sometimes seems to be the assumption that all music is at -.01 dB from start to finish thus no one will hear it and/or they just don't care. As I have stated before, just because there is a volume war and MP3s everywhere doesn't mean people don't listen and one doesn't have to be an audiophile to have good ears and pay attention.

Karbo

[JHughes]

Originally Posted by Lokasenna

The claim behind this entire topic is that one DAW can sound better than another

True. And that could be because the settings are different out of the box. I'm fine with that.

Originally Posted by Lokasenna

But you can't perform such a test unless the DAWs are set up to be as identical as possible.

Why do you keep trying to do that? This is about the DAWs sounding different, not whether they do math differently. No one has said that they do math differently so no one needs to prove it, and yet that is the experiment you keep trying to define. If it is a difference in settings, then make a list of the differences in settings and test those. Don't try to match the settings then prove there are no differences. That's a different subject that no one is arguing.

[Lokasenna]

Originally Posted by JHughes

Why do you keep trying to do that? This is about the DAWs sounding different, not whether they do math differently. No one has said that they do math differently so no one needs to prove it,

Yes, that IS the claim. There are supposedly differences in the math, beyond the settings we have access to, that end up making one DAW better than the next.

[pipelineaudio]

my problem is, if the results exactly null with the original, even after all this math is going on, then human perception is completely irrelevant

The only time perception is even a factor is when you can show there exists a difference.

[Geoff Waddington]

Originally Posted by Lokasenna

Yes, that IS the claim. There are supposedly differences in the math, beyond the settings we have access to, that end up making one DAW better than the next

I don't think that is the claim at all the way I read it.

From the original post:

Each time, without fail, I get frustrated with my control over the sound coming out of Sonar... but when I set up a mixing project in Reaper using these rendered Sonar files, they invariably sound more pristine, distinct, and I spend considerably less time in dialing in the mix that I want

I would define this claim more like:

"I get frustrated in Sonar and have an easier time getting a good mix in Reaper."

That's what I believe we're investigating here.

[Lokasenna]

I actually meant the general claim about DAWs sounding different, as discussed all over, where math differences and summing stuff ARE what the argument is about. In this particular thread, you're right, I'm arguing with a bit of a straw man.

[PAPT]

Originally Posted by Lokasenna

I actually meant the general claim about DAWs sounding different, as discussed all over, where math differences and summing stuff ARE what the argument is about. In this particular thread, you're right, I'm arguing with a bit of a straw man.

Then you are re-defining the discussion as posed in the original post.

I've tried to make this point more than once in this thread.

[PAPT]

Originally Posted by PAPT

Then you are re-defining the discussion as posed in the original post.

I've tried to make this point more than once in this thread.

That's the way these threads usually go.
Someone says they hear a difference and then there is a flood of posts about theory that ignores possible differences in settings that are causing a difference in sound, turning the discussion to math in the sound engine instead of the entire finished product that is represented by the DAW that a person actually uses.

[Lokasenna]

Did you forget to log onto another account before replying to yourself?

As I've said, repeatedly, when this topic comes up it's almost always about hypothetical differences and errors in how the various DAWs sum, render, etc. I've acknowledged that I misinterpreted the original post in this particular thread, but that doesn't somehow invalidate the argument I was making. It's just irrelevant to the discussion.

Let me try to start again...

Obviously each DAW will be set up a bit differently out of the box.

This might mean that a default-setting Reaper sounds better than a default-setting Sonar to you. In that case, great, use Reaper.

However, every one of those differences can be accounted for (as null-tests prove) by adjusting Sonar's settings to match.

Given this ability, the statement "Reaper sounds better to me than Sonar" has no real meaning - that is, because it's just a matter of altering the settings, Reaper and Sonar have no functional difference in this area.

[PAPT]

I don't know what I did to end up quoting myself.

[DBMusic]

For you Sonar users, here's an interesting test:

In any audio app except Sonar, take any .wav file, run it though a limiter (I used Elephant), and boost the output until the limiter is actively reducing the peaks. Typically this is in essence no different that what is done in a final mastering phase. Render the audio file.

Take that file and load it into Reaper, WaveLab, Sound Forge, or any other audio app (except Sonar). Leave all applicable faders set to 0 db. Upon playback you will see the peak output being registered correctly per the metering. It should register no more than where you set your limiter to.

Now take that file and load it into Sonar. Again leave all faders to 0 db. What do you see?

Each time I run this test I see Sonar clipping the master output by +5-6 db. The output clipping is also verified in my installation of RME TotalMix.

It appears as if Sonar is applying an addition +6db of gain to any audio file run through it.

Personally, this explains a perception I've had about Sonar's audio that I had long before I started using other DAW's. Whenever I exported a .wav file and then ran that file through another audio app (I always mastered in WaveLab) it would always sound louder...which now makes sense. You can actually take any audio file, run it through Sonar, then run it through another audio app and you will see the peak output in Sonar will register about +6 db greater. Export it from Sonar (with no processing or fader changes) and the exported file will also be about +6 db hotter. Export that original limited file from Sonar and you have a clipped mess.

How this exactly relates to the original perception that prompted this thread is unclear. I could find no settings in Sonar that would explain this difference. But all things considered, Sonar's audio engine appears to be doing something different.

Regards,

DB

[off&on]

Originally Posted by DBMusic

... Sonar is applying an addition +6db of gain to any audio file run through it.

What's 5dB between friends?

[claudioaudio]

Originally Posted by NAS

Reaper just sounds sooooooo sooooooo soooooo analogue
I mean it sounds phatter than tape

NAS

LOL!!!!

[dub3000]

Originally Posted by DBMusic

For you Sonar users, here's an interesting test:

...

Each time I run this test I see Sonar clipping the master output by +5-6 db. The output clipping is also verified in my installation of RME TotalMix.

It appears as if Sonar is applying an addition +6db of gain to any audio file run through it.

weird!

i don't have sonar, but some thoughts:

* is it importing stereo files as mono? or is stuff being routed to master twice?
* is there a default track trimpot set somewhere?

[Tod]

Originally Posted by DBMusic

For you Sonar users, here's an interesting test:

In any audio app except Sonar, take any .wav file, run it though a limiter (I used Elephant), and boost the output until the limiter is actively reducing the peaks. Typically this is in essence no different that what is done in a final mastering phase. Render the audio file.

Take that file and load it into Reaper, WaveLab, Sound Forge, or any other audio app (except Sonar). Leave all applicable faders set to 0 db. Upon playback you will see the peak output being registered correctly per the metering. It should register no more than where you set your limiter to.

Now take that file and load it into Sonar. Again leave all faders to 0 db. What do you see?

Each time I run this test I see Sonar clipping the master output by +5-6 db. The output clipping is also verified in my installation of RME TotalMix.

It appears as if Sonar is applying an addition +6db of gain to any audio file run through it.

DB

Humm DB, I've used Sonar for many years and not seen anything like what your talking about unless there were certain setting(s) that may not have been set properly. One possibility could be the "panning laws" although I don't know why that should affect the situation your suggesting. I know I've run into many problems in Sonar, especially bounceing tracks to other track(s) because of the "panning laws". If not set right it will result in a +/-3db or +/-6db error in the rendered track.

However, I import audio files into Sonar all the time and I've never encountered the problem your talking about. When you say "master output", which output is this? Is it possible you might be getting it there twice?

I'm not saying you're all wet DB, on the contrary I appreciate people who will make tests like this. I myself have made many many tests over the years but I've learned not to jump to any conclusions based on the results because sooner or later I might just run into something I wasn't aware of that influenced those results. However heh heh, because of your post I will be watching closer to make sure things are working properly.

Tod

[thebearingedge]

Originally Posted by l0calh05t

You have a 12GHz oscilloscope? I would be quite surprised if you did, and even then that would only be enough for the first two harmonics of that square wave.

sick burn

[DBMusic]

Originally Posted by Tod

Humm DB, I've used Sonar for many years and not seen anything like what your talking about unless there were certain setting(s) that may not have been set properly. One possibility could be the "panning laws" although I don't know why that should affect the situation your suggesting. I know I've run into many problems in Sonar, especially bounceing tracks to other track(s) because of the "panning laws". If not set right it will result in a +/-3db or +/-6db error in the rendered track.

However, I import audio files into Sonar all the time and I've never encountered the problem your talking about. When you say "master output", which output is this? Is it possible you might be getting it there twice?

I'm not saying you're all wet DB, on the contrary I appreciate people who will make tests like this. I myself have made many many tests over the years but I've learned not to jump to any conclusions based on the results because sooner or later I might just run into something I wasn't aware of that influenced those results. However heh heh, because of your post I will be watching closer to make sure things are working properly.

Tod

Well, I'm definitely all wet about this one

When I did this test in Sonar, each time I just created a quick test project using the Normal template. What I didn't notice is that there was an active Send created by default on the track I was using to playback the audio. Apparently it was creating a 6db boost to the Master because it was sending the same audio signal to the Master twice (via the Track output and the Send). Just tried this a few more times and sure enough there are active Sends created on each of the 2 audio tracks created by the Normal template. Go figure...

Boneheaded ideas about what should be "normal" behavior of an audio app are another reason I don't use Sonar much these days.

Anywho...my bad. Sorry for the wild goose chase...

Regards,

DB

[nicholas]

Originally Posted by DBMusic

For you Sonar users, here's an interesting test:

In any audio app except Sonar, take any .wav file, run it though a limiter (I used Elephant), and boost the output until the limiter is actively reducing the peaks. Typically this is in essence no different that what is done in a final mastering phase. Render the audio file.

Take that file and load it into Reaper, WaveLab, Sound Forge, or any other audio app (except Sonar). Leave all applicable faders set to 0 db. Upon playback you will see the peak output being registered correctly per the metering. It should register no more than where you set your limiter to.

Now take that file and load it into Sonar. Again leave all faders to 0 db. What do you see?

Each time I run this test I see Sonar clipping the master output by +5-6 db. The output clipping is also verified in my installation of RME TotalMix.

It appears as if Sonar is applying an addition +6db of gain to any audio file run through it.

Regards,

DB

Hey DB,

Out of interest I just gave this a try (Sonar 6) and didn't get the outcome that you did. The peaking in Sonar was the same as in REAPER and Adobe Audition. Sorry! I'd be really intrigued to discover what's causing this on your system.

[DBMusic]

Originally Posted by nicholas

Hey DB,

Out of interest I just gave this a try (Sonar 6) and didn't get the outcome that you did. The peaking in Sonar was the same as in REAPER and Adobe Audition. Sorry! I'd be really intrigued to discover what's causing this on your system.

nicholas,

Did you read my last post just before yours? Sneaky Sonar rascals...

Regards,

DB

[nicholas]

Originally Posted by DBMusic

nicholas,

Did you read my last post just before yours? Sneaky Sonar rascals...

Regards,

DB

Ah! You must have posted that while I was typing! Letters crossed in the post, that sort of thing! Mystery solved ...

[chip mcdonald]

Originally Posted by dub3000

that stuff doesn't have any effect (as in, that code never runs) unless you've imported audio into your session that's at the wrong sample rate. you can tell if this is the case if there are little 'i' icons on your media items.

No, it does have an effect if DAW user 1 says "it sounds different" and DAW user 2 says "it sounds the same", if they're not talking about the mixdown using different playback quality settings.

[GreyBeard]

That fact that people hear a difference doesn't surprise me. Do all DAW programs crunch the numbers exactly the same way using the same alga-rhythm? I really don't know but I'm guessing they don't. Music is about what people hear and if they hear a difference then there is a difference. Just because it can't be explained mathematically doesn't mean that it doesn't exist. I can hear a difference in projects in Reaper and PTLE. I say my ears are more sensitive than the mathematics used to describe what it going on.

[DBMusic]

Originally Posted by GreyBeard

That fact that people hear a difference doesn't surprise me. Do all DAW programs crunch the numbers exactly the same way using the same alga-rhythm? I really don't know but I'm guessing they don't. Music is about what people hear and if they hear a difference then there is a difference. Just because it can't be explained mathematically doesn't mean that it doesn't exist. I can hear a difference in projects in Reaper and PTLE. I say my ears are more sensitive than the mathematics used to describe what it going on.

As we see, this subject can be endlessly debated. The final analysis is all very esoteric. Bottom line, if your perceptions get you to sonic nirvana...it's all good.

Regards,

DB

[Kundalinguist]

I find that no matter what DAW I use, they all have this high pitched ringing sound coming from the right side of the stereo field. Wait! Unless that's my ear???

[Smurf]

Originally Posted by JHughes

True. And that could be because the settings are different out of the box. I'm fine with that.

And this has been my stand all along. If you have to adjust each DAW in the test to the EXACT same settings, then you are eliminating what is heard, and rendering the test useless IMHO...

[Lokasenna]

No, you're eliminating variables that could confuse a test of whether the two DAWs are actually any different.

I think we're all agreed, at this point, that the debate is [Argument 1]: "I think DAW X sounds better than DAW Y." First, it's a matter of opinion, blah blah blah. If you like one or the other, great, have fun, but don't bother telling anyone else about it.

Why? Because the statement literally means nothing. Seriously. You might as well have not said anything for the amount that it contributed to whatever discussion you were having.

If an investigation were to show that there IS, in fact, some magic going on in X's summing code, then Argument 1 would be a meaningful claim. Subjective, yes, but it would show that DAWs are not interchangeable and therefore one's DAW choice should give some thought to which sound you prefer.

So how do we figure out if X is doing something freaky with your audio? Easy - compare the math it does with Y and see if they get the same answer. But what if the difference is something as simple as the panning law, or some weird resampling? Well then, we'd better go through and make sure both DAWs are set up identically, no? If we don't, our test is invalidate and we're no closer to establishing the existence of any musical sorcery.

Once we strip off the different settings and determine that, at their core, both DAWs are doing their math/etc the exact same way*, we're left with [Argument 2]: "I think DAW X's default settings sound better than DAW Y's default settings." If Argument 2 is a meaningful statement, it follows that Argument 1 probably is too - Argument 2 is really just a more specific version of Argument 1.

Argument 2 is likewise an entirely subjective claim, so everybody's version of it is correct, but since the default settings ARE different it's definitely a meaningful thing to say - you can discuss the differences, and try to figure out why X's pan law is so much cooler than Y's.

There's one big reason, though, why we can't apply this conclusion back up to Argument 1 as well: in setting the DAWs identically to compare their defaults, so they now cancel each other in a null-test, etc, we've inadvertently proven that both DAWs have the capacity to sound like each other.

As a result, Argument 2 (and consequently Argument 1) simply has no real value - if I prefer X's default sound, but think Y has a better layout and workflow, then it's trivial to make Y sound like X and be done with it.

"This Lamborghini looks pretty cool, but my Camaro just feels better on the road" wouldn't be a meaningful statement if the Lamborghini had a couple of switches under the hood that could mirror the Camaro's performance.

A good software example would be Photoshop and Gimp. Gimp is an open-source graphics program, offering (more or less) all of the same features that Photoshop has. The interface, on the other hand, is entirely different, and a lot of people refused to give Gimp a chance because it was too much of a change from Photoshop.

Solution: Someone wrote an addon that rearranges Gimp's layout, menu names, etc to mirror Photoshop as much as possible.

Consequently, interface differences are no longer a valid reason to prefer Photoshop over Gimp.

Also:

Originally Posted by GreyBeard

That fact that people hear a difference doesn't surprise me. Do all DAW programs crunch the numbers exactly the same way using the same alga-rhythm? I really don't know but I'm guessing they don't.

Yes, they really do. That's the whole point.

* Hundreds of tests have confirmed this. A handful haven't, all of which turned out to be a result of settings that weren't, in fact, identical.

[heloego]

"Only Stephen Colbert can "feel" the difference and the truth of this in his gut. "

Look at Stephen carefully.
It's because he has two vastly different ears.