MB2 Petition

[Jouvert]

keep in mind MMM that i said it was a quick listen ( was a bit lazy to listen again in a more controlled environment).
I would not change my original answer, that would not be fair, so i am out of the running.
By the way were the v3 files normalized ( i know that's on by default), because bass perception may change.

Jouvert

[madmaxmiller]

keep in mind MMM that i said it was a quick listen ( was a bit lazy to listen again in a more controlled environment).
By the way were the v3 files normalized ( i know that's on by default), because bass perception would change.

Jouvert

My files were *not* normalised in both versions. 44.1/24 in - 44.1/24 out, no normalising. But Tassy indicated already that the normalised 16bit exports match, too... (and yes I know Fletcher-Munson...)

MMM

[Jouvert]

As you see MMM in an edit of my post which you quoted, that i would not change my answer in all fairness
I would just like to say that, even if we do not come to a common conclusion, this thread wasn't a complete waste of time because other discoveries where made that otherwise may not have (such as CLA's discovery).
Even though we may disagree, let's continue to keep things civil, because we still have to interact with each other on other posts.

jouvert

[Yriisem]

I made experiment mixed one and that at the simple song in two versions it is independent. Then the Mix from the second version ported in the third version with all settings. Interestingly that approach to mix in the second and in the third version differ. My sound in the second version turns out quicker and more comfortably for me. In the third at me that I do many doubts or not that

https://yadi.sk/d/ppc2q-KymYfZM

[Tassy]

Thank you very-very-very much MMM, for taking only seven words to proove wrong your own "approoved" work of sooo many days hard work...!
You just commited "suicide"...!!!

It's better to burn out than it is to fade...
My my... Hey hey... Rock'n'roll Is here to stay !!
My-my... Hey hey... Mb2 is here to stay !![/b]

LOL thanks for your professional comment, it's always a pleasure to work with people like you.

Prost,
MMM

---

I imported the MMM song as it is from the original Cambridge wave .zip into MB2 and V3.
did not do a thing only disengaged the master limiters.
Then exported
- MB2 to cd quality
-V3 to CD quality without Normalize
- V3 to CD quality with the default normalize on

Comparison result is very interesting:

listen: V3 without normalize has a lot more highs, cymbals, more bass all a bit fuller, louder

analyze: MB2 and V3 with normalize EXACTLY THE SAME spectrum
V3 without normalize a lot more bass, highs all different places

(do not think is is just a boosted but same-looking spectrum line so louder, NO, it is different a lot little places, frequencies more and less in a lot places)

What is more interesting that MB2 was on PC win7/64, V3 was on laptop win XP, nevertheless, they resulted the same spectrum and same perceived sound if normalize was on.
So my conclusion is that the sound difference between MB2 and V3 CAN BE really big but mostly DEPENDS on the chosen export options.

Will put a video later.
Tassy

...and that is a different test alltogether. Jouvert claimed the mixing engine has changed - which was proofed wrong. You proofed that MB2 export in CD quality must be automatically normalise (as it is written in the red book) while MB3 has the option to not normalise. If you normalise in 3.0 the result will match the one from 2.5.
Would be interesting to export 44.1/16 unnormalised from 2.5 and see what this is like.
Which dithering method did you use?

Also, have you done a null test between the exports? Can you provide the exported files somewhere for download? I would be particularly interested if I can null out the unnormalised tracks against the normalised ones using region gain...
2.5 unnormalised vs 2.5 normalised
3.0 unnormalised vs 3.0 normalised
...and then across versions...

Food for more experiments
MMM

Thanks for telling that in MB2 the normalize was there as dafault unseen, it explains a lot to me

Yes, I made some more tests leaving out human mistakes with V3
in the link
https://db.tt/vconTEqA

there are:
V3 export not normalized
v3 export normalized
The difference between the on/off normalize, (phase 0 test) in a neutral daw
The difference between the on/off normalize, (phase 0 test) in MB2.

Also saw that the exports from V3 sure are aligned but the export fom MB2 has a shorter start so cannot be aligned very sharp on my gear to make a proper 0 test between MB2 and V3 version.
that is the most align but cannot be called exact
http://screencast.com/t/F7e5bUMPEFi

and here is the sound of the difference,
https://db.tt/zjQTLKB9
but it is no proof because the align is not perfect to maths.

I did not use dithering because it is also something that can do different things
The samples were compared in wave but in the .zip i put mp3 to be shorter. the difference is so great that any opinion can be formed from the mp3 versions as well.

Remember that it is not a mix only the same raw tracks imported, nothing done, no fader no pan... nothing. Just a raw import
Tassy

[smallbutfine]

I have a few final notes from my side for the discussion.
There was an unanswered question:
How can i have an expectation bias if i listen to files from someone else, even in a blind test?
It is a very simple expectation bias. If you compare 2 files you think are different you expect them to sound different. The bias is against hearing that they are the same and for hearing a difference.
This happens all the time actually. People are often very sure they hear a difference, when they don't in replicated double blind tests. Sometimes they are even sure about tonal qualities and think they can pick the right one always. Even if the result is that they pick randomly.
This is the only reason i came up with a null test. I know how easy you can hear things that aren't there. E.g. you might know the "magic producer device". an EQ or compressor for the occasion a client wants "more" of something that is perfect. The magic device is adjusted, the client is happy he got his change, and nobody has to know the magic device is nowhere in a signal flow. Simply not connected ...
About normalization and different spectral quality:
Normalization is nothing else but clean gain to adjust the highest peak to a determined level without changing spectrum or transient behaviour. In the digital domain gain is a simple multiplication. If results in 2 normalization processes from 2 different DAW to the same level differ on any percievable level, then one of them has a very serious problem. I.e. programming error.

[Tassy]

For me only one question remained and ask Harrison to answer:

If it is a fact that the normalize or not is chosen makes a difference in the exported sound I wonder what can be the sound I am hearing while mixing?

At the start of Mixbus 2 some years ago I found the mixing sound and the exported always having some difference. Thought that dither or conversion can be the reason, now I know it should not be. Could not find it was my "feeling" or real. Then I did not care. Now just for my understanding:

What version is what I hear when mixing, playing tracks, and what when exported?

Thanks
Tassy

[madmaxmiller]

Normally, as others pointed out, normalising does just a clean gain, the ideal of a transparent amplification if you want so
So it should be possibe to phase flip the unnormalised version, the gain match it to the normalised version and the result should be zero, flat line.
When you hear the exported normalised version through the same set right after the normal mix it should sound differently because it's louder and psycho acoustics (yeah good old Fletcher-Munson) kick in...
MMM

---

As you see MMM in an edit of my post which you quoted, that i would not change my answer in all fairness
I would just like to say that, even if we do not come to a common conclusion, this thread wasn't a complete waste of time because other discoveries where made that otherwise may not have (such as CLA's discovery).
Even though we may disagree, let's continue to keep things civil, because we still have to interact with each other on other posts.

jouvert

Ummm Jouvert did I write anything you didn't like? Of course I will not loathe you because you have a different opinion
I'm happy to collaborate with everyone who wants to.

[Andy76]

Normalization is nothing else but clean gain ....

I thought the same until I read Bob Katz (mastering engineer) FAQ:

http://www.digido.com./audio-faq.html?option=com_fsf&Itemid=93&view=faq&catid=-2&search=normalize

Normalization Basics

Normalization is a DSP calculation, not a very nasty one, but it adds a minute amount of (probably imperceptible) distortion. But a little bit of imperceptible distortion accumulates the more processing that the track goes through. If you are going to be sending your material for mastering, do NOT normalize. To repeat: THERE IS NO NEED TO NORMALIZE IF THE MATERIAL IS GOING TO BE FURTHER PROCESSED. Let the next DSP step or analog processing step take care of two birds with one stone, be in the hands of the mastering engineer to avoid additional calculations, etc. In general, normalization should be avoided. In my book I cover this in more detail, but basically, once a track has already been recorded, you do not gain any quality by changing its gain, you only lose quality by requantizing it. If you are mixing it, you are going to be changing the gain once again anyway, so why do an extra quality-reducing DSP step prior to mixing?

[madmaxmiller]

Normalization is a DSP calculation, not a very nasty one, but it adds a minute amount of (probably imperceptible) distortion.

Yes, if you insist the highest peak to be exactly 0 dbFS or -1 dbFS you will have rounding errors underneath due to the digital nature of the thing... however, in 24 bit resolution they wouldn't be perceptible (as Bob wrote...).

But a little bit of imperceptible distortion accumulates the more processing that the track goes through.

Absolutely. But that's not what we have here, we are talking about the "final" normalising process and compare the results right after that without any further processing.

Thanks Andy76 for the input, I once used to normalise tracks before I actually started mixing... I stopped that for other reasons, so now here is one more to *not* doing it.

MMM

---

Yes, I made some more tests leaving out human mistakes with V3
in the link
https://db.tt/vconTEqA

...

and here is the sound of the difference,
https://db.tt/zjQTLKB9
but it is no proof because the align is not perfect to maths.

Thanks Tassy, I will do tests later today... and yeah your summing sounds like not aligned.

MMM

P.S.: nice and easy song, doesn't hurt too much listening without mixing, that's why I chose it...