Some experiment with mixing down in NUENDO and ProTools

I herd a lot about sounding in different DAW like almost everyone. And want to get sound as good as it possible like almost everyone. And I like NUENDO and want to help to get the good sound in this DAW.
I don’t understand deeply inner process of digital sound and I see that big mixing console mix sound better that DAW. And probably ProTools is one of the best in this among DAW family.At least I hear it like that.
Also I know about experiments with summing two samples one of which has reversed phase. So I decide to do the same with samples which was mixed down with volume curve on track (0 db on start of samples goes down to inf.)
Samples was moved from NUENDO to ProTools by AAF, volume curve was checked. Also I made the same (but without curve) with UAD Cambridge EQ with equal settings inserted in track in NUENDO and ProTools. And at the end I made two samples in NUENDO: using MIXING DOWN in REAL TIME and MIXING DOWN in BACKGROUND.Original sample was WIGHT NOISE -27,6 db mono. Nuendo was lunched as 64 bit app. I did up gain in two of gotten samples because difference was very small. “real time nuendo” test was up +40 db, “EQ” test +60 db. “Volume curve” test was’t gained because deference was obvious and not linear! My test was not precise as much as I wanted and I will appreciate if someone will repeat it or do similar or difference.
http://files.mail.ru/O1K2N5 - here is .zip of Nuendo trackarcheve. On 3-rd track you will see that ProTools sample is -1.6 db to make it closer to Nuendo sample. In “EQ” test samples moved relative each other because ProTools made kind of preroll.

Well, I am not at all sure white noise samples are going to tell you anything much that is useful.
It’s noise. Random noise too.
Much better to check on actual audio, or if youi mjust use test tones a sinewave.
Also make really, really sure that the Pan Law is set to the same, or there will always be differences.

You don;t say what version of alsihad you use, but PT9 and earlier use fixed point mixers.
PT10 uses 32-bit FP, same as Nuendo because apparently “it sounds better” (we could have told them that!)
So tjere should be zero difference as long as nothing random is used (like White Noise) and the pan laws are the same.

The other day I got paid to mix in nuendo, it sounded the best.

Then next day I got paid to mix in protools, it sounded the best too.

There is hardly much difference, at least between the two 32 bit float versions of Nuendo and Alsihad. If you insert FX plugs, you have to take into consideration that developers may have constructed the plugs differently (RTAS, VST …) which can result in different sound and levels. It is acceptable to check out the plugins, but may be problematic for your test of DAW sound.
As Neil stated, using noise is not ideal. I have done quite a few cancellation test and used a complex transparent piece of music, that gives you nice reverb and delay tails to judge the decays and its sound. There you recognise any shortcommings or differences.
The only time I had audible differences between real time and faster than realtime audio export was, when I had accidently set the latency of the RME card to a rather small buffer size > 1ms that messed up some plugin functions. With heavy projects, realtime export might be indeed the (feel-) better solution, though.

Big K

It’s very easy to test.

Just take any wav file.
Load it into PT.
Bounce
Load both files into Nuendo
Phase reverse.
Do they null?

Then do the same in reverse.

Do they null?

Then take both bounced files and test in both Nuendo and PT.

Do they null?

The answer in all cases should be yes, and if so, this will prove that there is no difference in the sound at all.

DG

As a music only user of Nuendo, for me it’s not important such tests, because: important is the musican, his feeling, his playstyle, his expressive strength … the Song overall, then good Mixing and Mastering …

I have done good Songs with my Amiga 500 for many years in a 8-Bit domain. All songs program in a list of HEX-code :mrgreen:

In spite of everything, as I listen my first record in Nuendo I thougth, WOW … very smoot and on that, it has not change anything to date.

Dose anyone download test project?

Agree


Answer NO. If you will up the gain of summed sample you will find out that it is not a silence. In “volume curve” test there is difference even without gaining. And I herd that it is not just linear noise but noise which is goes quieter than louder than quieter again.
In Nuendo sample it seems volume does not go smoothly down but has some angle.
I don’t claim to make any conclusions, I just want to understand how far I can go with Nuendo mixing.
I decided to make this test after my friend call. He was disappointed by comparison the same session (without any plug-ins and curves ) in Nuendo and Pro Tools. That was movie reel with hard war effects.
Also he found out that directing Nuendo channels on plenty of different busses makes situation better. I can not explain why.
Also I can not explain why changing oder of mixed tracks make different results. I mixed 6 tracks with Nuendo’s EQ:
1-st time it was 1,2,3,4,5,6
2-nd it was like 3,5,6,4,1,2
The results were not equal.After summing with Reversed Phase I had some noise -68 db. For me it is pretty wired but it should has reason.

In my test all files nulled. As soon as you move a fader the test means nothing.

DG

In my experience with the state of ALL the software available for mixing nowdays, the biggest limitation is the human sitting in the chair and the state of the ‘old technology’ of speakers and room acoustics. Those 3 things are going to have the biggest impact before any summing engine bollocks. Time and time again I have people tell me that Protools was better due to fixed point math, now PT10 is superior because it has 32bit Floating point math (like we have had for years in Nuendo).

Its all bollocks. Get your room right. Calibrate your speakers properly in that room. Use good AD/DA. Choose a DAW that works for your mindset and flow. Then mix and mix and mix and mix and mix and mix and mix and mix on that setup and then you will see how far you can go.

Volume curve and null tests have NOTHING to do with a good mix. If people spent all the time actually mixing than they spent doing worthless shootouts that are invalid (In the vol curve test you explain it would be easy to say there were variables such as automation resolution differences, dithering added in on PT side etc etc) - then maybe the internots would be a better place for good discussion on our side of the art

Exactly. All a null test proves is that there is nothing wrong with the DAW mixdown or summing engine. After that it’s all up to individual. If an engineer can only get a transparent mix in one DAW or another, then he is not a good engineer. All the engineers I know can get a good mix in any DAW. Some may take longer, because they don’t know the shortcuts as well, but in the end all are equally good. If the RTAS version of a plug sounds slightly different, then tweak it until you like it. Or use a different plug.

The thought of doing a null test on a mix is ridiculous. I am pretty confident that if I gave a series of wav files to any world class engineer an told him to mix them one day and then do an identical mix the next day on the same system, they would not sound exactly the same.

Just use your ears, and if someone gets a better mix than you can on another system, it’s because they are better than you are. What’s the expression? It’s the Wizard, not the wand. :wink:

DG

+1 well said

A good mix engineer has something special, and it’s not their gear. The gear only helps them get what they hear in their mind to come out of the speakers in a timely manner. A poor engineer on a state of the art system will still knock out shitty mixes.

I don’t try to to convince someone to choose PT instead of Nuendo. I work in Nuendo by myself! I just found out that all talking about equal summing in all DAW I cannot conform. All test I did said me opposite. Even test with 0 volume has result different from silence. Of coarse -138 db it’s almost nothing but it’s not a silence as was told. May be it’s mean nothing, may be Nuendo mix more precisely than PT. May be my test was not correct. I don’t know. I just ask someone to repeat tests and answer me what’s happened. Think of it like about curiosity.

It can’t.
24 bit = 144 dB of dynamic range
16 bit = 96 dB of dynamic range.

Floating point mixers produce distortion in the LSB or two LSB’s = -138dB or -132 dB.
In Fixed point mixers this distortion is “masked” by dither noise, so you will find noise in the LSB’s.

If your tests would have produced absolute silence, the something would be very wrong.
;o)

Fredo

Thank you Fredo. Finally there are some answers :slight_smile:

Well said.
It’s not what you use but the way you use it that counts. There are people who can mix so much better than I can it’s just not funny, and they can do it using pretty much anything that is lying around. It’s a constant learning curve and every day teaches me something I did not know that morning.
Life is good.

Just so you know, this has been talked about so many times many people don’t even bother with a serious reply. To you it might be new, but the topic pops up regularly on forums and has done so for years.

So if you find that people are dismissive or don’t take it seriously, just remember that it’s not personal, it’s just that we’ve been down this road so many times before and have better things to worry about.

Just in case you didn’t know…

During this tests and discussions I become to guess that all talking about all DAW mix equal seems to be myth because there are limitation for correct test. May be I am wrong but I close this question for myself and will trust my ears.

Good decision! After all, “it’s the ear, not the gear”.

Kind regards,

I had no luck in making mixes from different DAWs sounds different when testing correctly … Up to appr. -130 dB which doesn’t matter at all.