Reason Sound Quality

[theshoemaker]

It's not a DAW issue, it's a case of PEBCAC.

I have to agree on that. Been a PEBCAC for quite some time before realising how to improve on different techniques to get the result I wanted. For the things I'm doing I love to put Audiomatics with VHS and TAPE on the mix channel and then compress it a bit (turn up the squash) with the pulveriser with the init patch without filter, and finally a little bit of reverb.

[normen]

If you read discussions on other forums there is a lot of discussion about how bad the sound quality in Ableton is compared to xyz. It's a constant topic of debate debate.

Reason was once commonly criticised for its sound quality. The criticism was levelled at the 14:2 mixer which apparently has a poor summing algorithm and lacks headroom. If you read Gearslutz back in the day common advice from mix engineers to was bypass the line mixers entirely and rewire the individual outs into Pro Tools. However the new SSL mixer is pretty much universally praised for its sound quality. With its near infinite headroom and floating point summing.

That something is a constant topic of debate doesn't mean its based in reality. The plural of anecdote isn't "data".

"Summing algorithm" is pretty fancy words for sampleA+sampleB+sampleC...=outputSample, which LITERALLY is how summing works in a digital system. (That is LITERALLY, not "instagram literally")

[miscend]

If you read discussions on other forums there is a lot of discussion about how bad the sound quality in Ableton is compared to xyz. It's a constant topic of debate debate.

Reason was once commonly criticised for its sound quality. The criticism was levelled at the 14:2 mixer which apparently has a poor summing algorithm and lacks headroom. If you read Gearslutz back in the day common advice from mix engineers to was bypass the line mixers entirely and rewire the individual outs into Pro Tools. However the new SSL mixer is pretty much universally praised for its sound quality. With its near infinite headroom and floating point summing.

That something is a constant topic of debate doesn't mean its based in reality. The plural of anecdote isn't "data".

"Summing algorithm" is pretty fancy words for sampleA+sampleB+sampleC...=outputSample, which LITERALLY is how summing works in a digital system. (That is LITERALLY, not "instagram literally")

You can get digital clipping when you sum audio. Because intersample peaks don't show up on mixer meters and you don't know how well the summing engine of any particular DAW would deal with getting clipped internally.
Certainly back in the early days with every DAW you had to be extra careful with gain staging or you would get worse sound. It's less of an issue now that most modern DAWs have updated their engines to give you lots of headroom. But I'm willing to bet that when people say DAW X sounds better than DAW Y, if it isn't the difference in pan laws, then it's a gain staging issue.

[jonheal]

This is my completely subjective take on the "Reason sound." Honestly, some Reason patches do sometimes sound "thinner" to me. I don't think Reason is lacking, but perhaps the tastes of those who make the patches favor a sound that somewhat emphasizes the midranges, or a sound that simply does not overly emphasize the highs and lows.

So much of what is considered a "fat" sound, I think is nothing more than the Bass and Treble turned up.

[miscend]

I have read countless discussions on the sound quality of Reason, and tonight as a result of weeks of frustration I was messing around on Ableton with some pads and synths and couldn't help but notice that the sound quality seemed so much more fuller and stronger than Reasons. What do you guys think?

I am currently debating whether or not to upgrade to Reason 9 within the next month or so, I'm having tremendous writer's block at the moment but I think it's about time to upgrade.

If you read discussions on other forums there is a lot of discussion about how bad the sound quality in Ableton is compared to xyz. It's a constant topic of debate debate.

Reason was once commonly criticised for its sound quality. The criticism was levelled at the 14:2 mixer which apparently has a poor summing algorithm and lacks headroom. If you read Gearslutz back in the day common advice from mix engineers to was bypass the line mixers entirely and rewire the individual outs into Pro Tools. However the new SSL mixer is pretty much universally praised for its sound quality. It has near infinite headroom.

14:2 mixer has none of the issues you mention, at least none I've ever been able to hear or measure. In almost 14 years of using Reason!


Sent from my iPad using Tapatalk

Propellerheads rewrote their audio engine around the time of Record 1.0 and Reason 5.

I will mix down a project using the old mixer in Reason 4 and the new mixer in the current version. And see if they both null each other.

[MitchClark89]

(That is LITERALLY, not "instagram literally")

i LOVE so much that you said this

[selig]

If you read discussions on other forums there is a lot of discussion about how bad the sound quality in Ableton is compared to xyz. It's a constant topic of debate debate.

Reason was once commonly criticised for its sound quality. The criticism was levelled at the 14:2 mixer which apparently has a poor summing algorithm and lacks headroom. If you read Gearslutz back in the day common advice from mix engineers to was bypass the line mixers entirely and rewire the individual outs into Pro Tools. However the new SSL mixer is pretty much universally praised for its sound quality. With its near infinite headroom and floating point summing.

That something is a constant topic of debate doesn't mean its based in reality. The plural of anecdote isn't "data".

"Summing algorithm" is pretty fancy words for sampleA+sampleB+sampleC...=outputSample, which LITERALLY is how summing works in a digital system. (That is LITERALLY, not "instagram literally")

You can get digital clipping when you sum audio. Because intersample peaks don't show up on mixer meters and you don't know how well the summing engine of any particular DAW would deal with getting clipped internally.
Certainly back in the early days with every DAW you had to be extra careful with gain staging or you would get worse sound. It's less of an issue now that most modern DAWs have updated their engines to give you lots of headroom. But I'm willing to bet that when people say DAW X sounds better than DAW Y, if it isn't the difference in pan laws, then it's a gain staging issue.

We're talking about Reason, so:
Have to totally disagree with the idea of clipping when summing audio - read on...
You don't get intersample peaks when summing because intersample peaks come only when you move from floating point to fixed point - and summing is done with 64 bit floating point precision, making clipping when summing "impossible" to happen!

Early DAWs, at least Pro Tools, used FIXED POINT internal audio paths so it WAS possible to clip internally. Just making sure we're comparing apples to apples here.

Guess I better order more wipes, we're not done pooping yet…

[theshoemaker]

If you read discussions on other forums there is a lot of discussion about how bad the sound quality in Ableton is compared to xyz. It's a constant topic of debate debate.

Reason was once commonly criticised for its sound quality. The criticism was levelled at the 14:2 mixer which apparently has a poor summing algorithm and lacks headroom. If you read Gearslutz back in the day common advice from mix engineers to was bypass the line mixers entirely and rewire the individual outs into Pro Tools. However the new SSL mixer is pretty much universally praised for its sound quality. With its near infinite headroom and floating point summing.

That something is a constant topic of debate doesn't mean its based in reality. The plural of anecdote isn't "data".

"Summing algorithm" is pretty fancy words for sampleA+sampleB+sampleC...=outputSample, which LITERALLY is how summing works in a digital system. (That is LITERALLY, not "instagram literally")

You can get digital clipping when you sum audio. Because intersample peaks don't show up on mixer meters and you don't know how well the summing engine of any particular DAW would deal with getting clipped internally.
Certainly back in the early days with every DAW you had to be extra careful with gain staging or you would get worse sound. It's less of an issue now that most modern DAWs have updated their engines to give you lots of headroom. But I'm willing to bet that when people say DAW X sounds better than DAW Y, if it isn't the difference in pan laws, then it's a gain staging issue.

We're talking about Reason, so:
Have to totally disagree with the idea of clipping when summing audio - read on...
You don't get intersample peaks when summing because intersample peaks come only when you move from floating point to fixed point - and summing is done with 64 bit floating point precision, making clipping when summing "impossible" to happen!

Early DAWs, at least Pro Tools, used FIXED POINT internal audio paths so it WAS possible to clip internally. Just making sure we're comparing apples to apples here.

Guess I better order more wipes, we're not done pooping yet…

I've fixed a track of mine lately where I've been thinking it is inter sample clipping. It didn't show up clipping on the SSL, but I could clearly hear it. There have been a couple of MonoPoly's for pads running playing each chords in the range of C0 to C2 with 5 to 6 notes each. Maybe because by evolving to long and the reverb I put on. I ended up to turning down the note length of the chords. So is that inter sample clipping?

[ScuzzyEye]

Inter-sample clipping can't happen in the digital domain, because digital audio only deals with samples, and inter-sample by definition is: between the samples.

Inter-sample clipping happens in the DAC. DACs don't just stair-step between samples, they plot a sine segment that fits from one sample to the next (part of why the Nyquist limit is the sampling rate divided by 2). It's possible for a sine wave connecting two points to be higher than those two. That's when you get inter-sample peaking. If the DAC can't provide enough voltage to recreate this peak, it will be clipped.

I think in the most extreme, test signal case you can probably hit +4 dB from an inter-sample peak. But for normal music type signals, it's usually limited to around half a dB. If you limit digital signals to -0.5 dB you're very unlikely to cause the DAC to clip on playback.

[selig]

If you read discussions on other forums there is a lot of discussion about how bad the sound quality in Ableton is compared to xyz. It's a constant topic of debate debate.

Reason was once commonly criticised for its sound quality. The criticism was levelled at the 14:2 mixer which apparently has a poor summing algorithm and lacks headroom. If you read Gearslutz back in the day common advice from mix engineers to was bypass the line mixers entirely and rewire the individual outs into Pro Tools. However the new SSL mixer is pretty much universally praised for its sound quality. With its near infinite headroom and floating point summing.

That something is a constant topic of debate doesn't mean its based in reality. The plural of anecdote isn't "data".

"Summing algorithm" is pretty fancy words for sampleA+sampleB+sampleC...=outputSample, which LITERALLY is how summing works in a digital system. (That is LITERALLY, not "instagram literally")

You can get digital clipping when you sum audio. Because intersample peaks don't show up on mixer meters and you don't know how well the summing engine of any particular DAW would deal with getting clipped internally.
Certainly back in the early days with every DAW you had to be extra careful with gain staging or you would get worse sound. It's less of an issue now that most modern DAWs have updated their engines to give you lots of headroom. But I'm willing to bet that when people say DAW X sounds better than DAW Y, if it isn't the difference in pan laws, then it's a gain staging issue.

We're talking about Reason, so:
Have to totally disagree with the idea of clipping when summing audio - read on...
You don't get intersample peaks when summing because intersample peaks come only when you move from floating point to fixed point - and summing is done with 64 bit floating point precision, making clipping when summing "impossible" to happen!

Early DAWs, at least Pro Tools, used FIXED POINT internal audio paths so it WAS possible to clip internally. Just making sure we're comparing apples to apples here.

Guess I better order more wipes, we're not done pooping yet…

I've fixed a track of mine lately where I've been thinking it is inter sample clipping. It didn't show up clipping on the SSL, but I could clearly hear it. There have been a couple of MonoPoly's for pads running playing each chords in the range of C0 to C2 with 5 to 6 notes each. Maybe because by evolving to long and the reverb I put on. I ended up to turning down the note length of the chords. So is that inter sample clipping?

Depends on a lot of things. Depends on whether there's anything else in the signal path such as a compressor or limiter that could be sounding "distorted" due to a short release time etc., depends on what your peak meters were showing (if you were peaking close to 0 dBFS it COULD be intersample clipping), depends on if you were pushing your monitors into clipping! Could also be something in the room rattling via sympathetic vibration (it's happened to me in one case!).

It's all speculation unless I was sitting at your system and could work through the process of elimination to discover the cause of the problem.

[selig]

Inter-sample clipping can't happen in the digital domain, because digital audio only deals with samples, and inter-sample by definition is: between the samples.

Inter-sample clipping happens in the DAC. DACs don't just stair-step between samples, they plot a sine segment that fits from one sample to the next (part of why the Nyquist limit is the sampling rate divided by 2). It's possible for a sine wave connecting two points to be higher than those two. That's when you get inter-sample peaking. If the DAC can't provide enough voltage to recreate this peak, it will be clipped.

I think in the most extreme, test signal case you can probably hit +4 dB from an inter-sample peak. But for normal music type signals, it's usually limited to around half a dB. If you limit digital signals to -0.5 dB you're very unlikely to cause the DAC to clip on playback.

Agreed! For 'mission critical' music (especially acoustic tracks) I limit to -1.0 dBFS just to be sure. These types of music aren't a part of the "loudness wars" anyway, so it's not even something you'll notice IMO.

[theshoemaker]

I've fixed a track of mine lately where I've been thinking it is inter sample clipping. It didn't show up clipping on the SSL, but I could clearly hear it. There have been a couple of MonoPoly's for pads running playing each chords in the range of C0 to C2 with 5 to 6 notes each. Maybe because by evolving to long and the reverb I put on. I ended up to turning down the note length of the chords. So is that inter sample clipping?

Depends on a lot of things. Depends on whether there's anything else in the signal path such as a compressor or limiter that could be sounding "distorted" due to a short release time etc., depends on what your peak meters were showing (if you were peaking close to 0 dBFS it COULD be intersample clipping), depends on if you were pushing your monitors into clipping! Could also be something in the room rattling via sympathetic vibration (it's happened to me in one case!).

It's all speculation unless I was sitting at your system and could work through the process of elimination to discover the cause of the problem.

Thanks for that insight. Is there a book or video or any other reference which you recommend on Mixing, Mastering, Sounddesign, Loudness, Volume .. ? I didn't read a lot about this the last 4-5 years, when I started with synths. But now I feel like I want to understand more of this and dig deeper and actually learn more in order to improve on the quality part.

[ScuzzyEye]

DACs don't just stair-step between samples, they plot a sine segment that fits from one sample to the next...

Just an extra bit of info on this. If you've ever used a limiter with inter-sample limiting, and seen how much extra CPU power that feature requires, that's because sine-fitting is hard. Basically, given two points there is exactly one band-limited sine function that passes through both. Finding that curve is difficult, and then finding the maximum amplitude of that curve between the two points is even more work.

You can cheat, and get close, by oversampling and filtering, but that's still an expensive operation.

[selig]

DACs don't just stair-step between samples, they plot a sine segment that fits from one sample to the next...

Just an extra bit of info on this. If you've ever used a limiter with inter-sample limiting, and seen how much extra CPU power that feature requires, that's because sine-fitting is hard. Basically, given two points there is exactly one band-limited sine function that passes through both. Finding that curve is difficult, and then finding the maximum amplitude of that curve between the two points is even more work.

You can cheat, and get close, by oversampling and filtering, but that's still an expensive operation.

Wait, when you export, you add a reconstruction filter, right? No "sine fitting" goes on at export, no oversampling goes on at export - or does it?

What I know is that the more you limit and push the envelope (with regards to extreme loudness), the more samples are likely to clip after export. Then there's conversion to mp3 etc…

[selig]

I've fixed a track of mine lately where I've been thinking it is inter sample clipping. It didn't show up clipping on the SSL, but I could clearly hear it. There have been a couple of MonoPoly's for pads running playing each chords in the range of C0 to C2 with 5 to 6 notes each. Maybe because by evolving to long and the reverb I put on. I ended up to turning down the note length of the chords. So is that inter sample clipping?

Depends on a lot of things. Depends on whether there's anything else in the signal path such as a compressor or limiter that could be sounding "distorted" due to a short release time etc., depends on what your peak meters were showing (if you were peaking close to 0 dBFS it COULD be intersample clipping), depends on if you were pushing your monitors into clipping! Could also be something in the room rattling via sympathetic vibration (it's happened to me in one case!).

It's all speculation unless I was sitting at your system and could work through the process of elimination to discover the cause of the problem.

Thanks for that insight. Is there a book or video or any other reference which you recommend on Mixing, Mastering, Sounddesign, Loudness, Volume .. ? I didn't read a lot about this the last 4-5 years, when I started with synths. But now I feel like I want to understand more of this and dig deeper and actually learn more in order to improve on the quality part.

Sounds like you're on your way - you can hear the clipping, you can eliminate it!

If you really want to dig deeper, as in learning digital theory and how it applies to our work, I found this book to be extremely helpful.
Digital Audio Explained (for the audio engineer):
https://www.amazon.com/Digital-Audio-Ex ... 141960001X

[ScuzzyEye]

Wait, when you export, you add a reconstruction filter, right? No "sine fitting" goes on at export, no oversampling goes on at export - or does it?

What I know is that the more you limit and push the envelope (with regards to extreme loudness), the more samples are likely to clip after export. Then there's conversion to mp3 etc…

Digital export from a DAW doesn't involve the reconstruction filter. That's part of the DAC. It's actually a kind of magical physical process. The filter does the sine-fitting automatically. But if you want to model it, it's difficult math.

MP3 encoding actually store cosine coefficients not samples. But because it's doing a (fast) Fourier transform to break the signal up into frequency bands it already knows the correct sine functions (FFT also isn't cheap). Because of the lossy storing values as cosines decoding an MP3 back to PCM can result in higher peaks than in the original data.

[theshoemaker]

Depends on a lot of things. Depends on whether there's anything else in the signal path such as a compressor or limiter that could be sounding "distorted" due to a short release time etc., depends on what your peak meters were showing (if you were peaking close to 0 dBFS it COULD be intersample clipping), depends on if you were pushing your monitors into clipping! Could also be something in the room rattling via sympathetic vibration (it's happened to me in one case!).

It's all speculation unless I was sitting at your system and could work through the process of elimination to discover the cause of the problem.

Thanks for that insight. Is there a book or video or any other reference which you recommend on Mixing, Mastering, Sounddesign, Loudness, Volume .. ? I didn't read a lot about this the last 4-5 years, when I started with synths. But now I feel like I want to understand more of this and dig deeper and actually learn more in order to improve on the quality part.

Sounds like you're on your way - you can hear the clipping, you can eliminate it!

If you really want to dig deeper, as in learning digital theory and how it applies to our work, I found this book to be extremely helpful.
Digital Audio Explained (for the audio engineer):
https://www.amazon.com/Digital-Audio-Ex ... 141960001X

Thanks for the pointer. Will definitely have a read. I'm pretty sure I'll get it ... being software developer for almost 20 years now will help

[normen]

You can get digital clipping when you sum audio. Because intersample peaks don't show up on mixer meters and you don't know how well the summing engine of any particular DAW would deal with getting clipped internally.

Again, what "summing engine"? You know EXACTLY how adding two or more double (i.e.64bit) or float (i.e.32bit) values behaves. And no, they don't clip at 1 (which represents 0dB), they don't clip at +100dB either. And inter sample peaks are not relevant while still in the digital domain as others pointed out.

Look at it this way: Would you expect one spreadsheet program to give different values than another when summing some values? Would you say one gives "warmer" results than the other?

If any DAW sounds different when summing two tracks its either because one of them has a bug - like ProTools TDM did for a while (in the 90s I think), which caused the "analog summing craze" until they found it (the last bit was swapped between left and right channel) or because they deliberately do some other things (like Harrys MixBus or whatsisname). But a bug like in ProTools is very unlikely in a DAW because if adding two floats in C++ would give false values then the problem in the compiler, operating system or CPU would occur in ALL applications, not just a DAW. ProTools was proprietary hardware so DigiDesign messed up there.

[ScuzzyEye]

Again, what "summing engine"? You know EXACTLY how adding two or more double (i.e.64bit) or float (i.e.32bit) values behaves. And no, they don't clip at 1 (which represents 0dB), they don't clip at +100dB either. And inter sample peaks are not relevant while still in the digital domain as others pointed out.

Look at it this way: Would you expect one spreadsheet program to give different values than another when summing some values? Would you say one gives "warmer" results than the other?

The only way summing could produce "warmer" results is if it did the math incorrectly. Like if adding 0.25 + 0.25 produced 0.46, and 0.50 + 0.50 is 0.79, and 1 + 1 = 1.1. That is to say, soft clipping, which adds harmonic distortion. Some people swear by out of the box, analog summing. But that's just because when summing too hot signals they'll start to saturate the circuits.

[normen]

The only way summing could produce "warmer" results is if it did the math incorrectly. Like if adding 0.25 + 0.25 produced 0.46, and 0.50 + 0.50 is 0.79, and 1 + 1 = 1.1. That is to say, soft clipping, which adds harmonic distortion. Some people swear by out of the box, analog summing. But that's just because when summing too hot signals they'll start to saturate the circuits.

The expensive boxes are hard to saturate with 0dB digital signals... I guess the main failure is in the A/B tests they set up, 0.1dB more always sounds "better". As I said, the whole myth came up when ProTools had that bug. Somebody (kudos to him!) heard that the stereo base took some hit when mixing in ProTools compared to his analog desk (not "warmth", "fatness" or some other BS) but not being an engineer in the technical sense his conclusion was that analog summing is better - causing an avalanche of esoterics.

To be fair though, our hearing is SO VERY influenced by our brains that its not exactly a fallacy to think something sounds better/different than something else even if theres absolutely no difference. Any audio engineer who says he never tweaked an EQ that wasn't even engaged and thought the results were what he meant to do is lying

[chimp_spanner]

The only way summing could produce "warmer" results is if it did the math incorrectly. Like if adding 0.25 + 0.25 produced 0.46, and 0.50 + 0.50 is 0.79, and 1 + 1 = 1.1. That is to say, soft clipping, which adds harmonic distortion. Some people swear by out of the box, analog summing. But that's just because when summing too hot signals they'll start to saturate the circuits.

The expensive boxes are hard to saturate with 0dB digital signals... I guess the main failure is in the A/B tests they set up, 0.1dB more always sounds "better". As I said, the whole myth came up when ProTools had that bug. Somebody (kudos to him!) heard that the stereo base took some hit when mixing in ProTools compared to his analog desk (not "warmth", "fatness" or some other BS) but not being an engineer in the technical sense his conclusion was that analog summing is better - causing an avalanche of esoterics.

To be fair though, our hearing is SO VERY influenced by our brains that its not exactly a fallacy to think something sounds better/different than something else even if theres absolutely no difference. Any audio engineer who says he never tweaked an EQ that wasn't even engaged and thought the results were what he meant to do is lying

Very true. Everything from the look and feel of a piece of software through to the brand name or even celebrity endorsements can influence how we feel about it. Like, I can probably make the same sorts of sounds with Thor as I do in PolySix (I watched a great comparison video of this actually). But the name/Korg brand carries weight for me, and I'm convinced there's some hidden mojo in their emulation. It's funny because in life, I'm not generally influenced by advertising. But I see words like "virtual analogue", "analogue warmth", "modelled circuitry" and I'm like, oh, well it MUST be good

[selig]

Wait, when you export, you add a reconstruction filter, right? No "sine fitting" goes on at export, no oversampling goes on at export - or does it?

What I know is that the more you limit and push the envelope (with regards to extreme loudness), the more samples are likely to clip after export. Then there's conversion to mp3 etc…

Digital export from a DAW doesn't involve the reconstruction filter. That's part of the DAC. It's actually a kind of magical physical process. The filter does the sine-fitting automatically. But if you want to model it, it's difficult math.

MP3 encoding actually store cosine coefficients not samples. But because it's doing a (fast) Fourier transform to break the signal up into frequency bands it already knows the correct sine functions (FFT also isn't cheap). Because of the lossy storing values as cosines decoding an MP3 back to PCM can result in higher peaks than in the original data.

Indeed, I missed big time there! So why is a reconstruction filter "difficult math"? Sorry for the mundane questions, it's an area I only understand at a basic level (clearly!).

[selig]

You can get digital clipping when you sum audio. Because intersample peaks don't show up on mixer meters and you don't know how well the summing engine of any particular DAW would deal with getting clipped internally.

Again, what "summing engine"? You know EXACTLY how adding two or more double (i.e.64bit) or float (i.e.32bit) values behaves. And no, they don't clip at 1 (which represents 0dB), they don't clip at +100dB either. And inter sample peaks are not relevant while still in the digital domain as others pointed out.

Look at it this way: Would you expect one spreadsheet program to give different values than another when summing some values? Would you say one gives "warmer" results than the other?

If any DAW sounds different when summing two tracks its either because one of them has a bug - like ProTools TDM did for a while (in the 90s I think), which caused the "analog summing craze" until they found it (the last bit was swapped between left and right channel) or because they deliberately do some other things (like Harrys MixBus or whatsisname). But a bug like in ProTools is very unlikely in a DAW because if adding two floats in C++ would give false values then the problem in the compiler, operating system or CPU would occur in ALL applications, not just a DAW. ProTools was proprietary hardware so DigiDesign messed up there.

Taking it further:
The "bug" in Pro Tools only occurred when you exceeded a certain number of channels, somewhere in the 40s IIRC. This was because of how the original engine was designed (and the limits of the hardware used at the time - remember PT was hardware based) and how they had to expand it to allow more channels, again IIRC. So many folks would never see the problem, and others would see it regularly, depending on how many channels they were summing and whether they spanned two cards or not. It was the way the data from the two cards was combined that caused the issue, where there was truncation that was in part because of the fixed point audio paths used by the hardware (PCI cards) used by PT.

Anyway, all DAWs use floating point audio now, so these issues are in the past!

[deepndark]

A Joke: Do you have a low quality engined DAW?
Answer: Yes, Reason.
Joker: How?
Answer: I use way too many Pelle Jubel's sounds.
Joker: WHAT??

[selig]

The only way summing could produce "warmer" results is if it did the math incorrectly. Like if adding 0.25 + 0.25 produced 0.46, and 0.50 + 0.50 is 0.79, and 1 + 1 = 1.1. That is to say, soft clipping, which adds harmonic distortion. Some people swear by out of the box, analog summing. But that's just because when summing too hot signals they'll start to saturate the circuits.

The expensive boxes are hard to saturate with 0dB digital signals... I guess the main failure is in the A/B tests they set up, 0.1dB more always sounds "better". As I said, the whole myth came up when ProTools had that bug. Somebody (kudos to him!) heard that the stereo base took some hit when mixing in ProTools compared to his analog desk (not "warmth", "fatness" or some other BS) but not being an engineer in the technical sense his conclusion was that analog summing is better - causing an avalanche of esoterics.

To be fair though, our hearing is SO VERY influenced by our brains that its not exactly a fallacy to think something sounds better/different than something else even if theres absolutely no difference. Any audio engineer who says he never tweaked an EQ that wasn't even engaged and thought the results were what he meant to do is lying

Indeed - the only engineers I don't trust are the ones that claim this never happens to them. Some of the best lessons I learned early on was in watching seasoned engineers do just that, and then laughing at themselves and saying words along this lines of "keeps me humble".

One thing that also happens is that we test our gear, then work for years without re-testing. Things change, and things are "fixed" but we don't always re-test every time something changes. So we assume the original problems are still there, but we only know once we test.

Reminds me of a story about the great-grandmother that always cut the turkey in half before roasting, which got passed down from generation to generation. Finally the young great-granddaughter asked the all important "why", to which the great-grandmother replied "because my oven was too small'!. Seems we sometimes do things out of habit rather than necessity, passing down the "how" but not the ALL IMPORTANT WHY!!!

This is one thing that really bugs me about tutorials that show "how" to do something but never explain "why" you would use the technique being demonstrated in the first place. So folks just end up doing "it" because it was shown, not because they were taught WHY (and therefore WHEN or IF) to do something.

Sorry for the rant/tangent, back to pooping…