Is there a sound problem on mixdown with FL Studio or not?

[Burillo]

even the best analog equipment can only reproduce 18-20 bits anyway...

[dreamkeeper]

But really, the difference is very small .. in most cases absolutely inaudible even after further processing. It's not like a 16bit recording would immediatly sound bad .. in most cases it will sound exactly like as it would in 24bit.

The fact that many soundcards (esp. older ones) aren't much better than 16 bits anyway might have something to do with it.

I'm not saying that there is a noticeable difference, but looking at the numbers suggests that such a difference would be smaller than one might think anyway. For instance the well-known M-Audio Audiophile 2496 has a SNR of 100.4 dB-A on the AD side - that's hardly 17 bits worth, with about 7 bits of noise in a 24bit sample.

Even the best converters only give you about 120dB dynamic range, equating to about 20 bits of useful signal. That's only 4 bits more than 16 - best case. Some manufacturers don't even mention the AD and DA specs separately, and the DA side is almost always better - with the AD side conveniently omitted. Just something to keep in mind when making comparisons.

werner

[STML]

I'm with jupiter8 on the bit-topic.
I'd never record in 16bit, because there's no reason and I like to be on the safe side.

But really, the difference is very small .. in most cases absolutely inaudible even after further processing. It's not like a 16bit recording would immediatly sound bad .. in most cases it will sound exactly like as it would in 24bit.

No, I think subtle context is what really needs putting straight here. Its not that 16 bit recording sound bad. Not at all, as mentioned, there are a lot of early digital tracks that sound superb. However, those engineers probably had to work pretty hard to overcome headroom issues associated with getting clean sounding stuff above the 16 bit noise floor. If you use 24bit properly (i.e. not redlining stuff all the time), you have a lot more headroom to play with, without losing resolution or getting an audible noise floor and consequently, have, for the most part, a much easier time producing a good dynamic mix. YMMV ofc. Its not really about what is audible to the end listener, its about how much easier it is for the mix engineer to work with the content.

[crimsonwarlock]

For instance the well-known M-Audio Audiophile 2496 has a SNR of 100.4 dB-A on the AD side - that's hardly 17 bits worth, with about 7 bits of noise in a 24bit sample.

So now you have 17 bits worth of information and you have the choice of cramming that into a 16 bit audio file or giving it some room by recording to 24 bits. Hmmmm.... difficult choice.... NOT

[dreamkeeper]

For instance the well-known M-Audio Audiophile 2496 has a SNR of 100.4 dB-A on the AD side - that's hardly 17 bits worth, with about 7 bits of noise in a 24bit sample.

So now you have 17 bits worth of information and you have the choice of cramming that into a 16 bit audio file or giving it some room by recording to 24 bits. Hmmmm.... difficult choice.... NOT

Of course. But that wasn't my point. This actually tiny difference might be one reason why people can't notice a difference between 16bit and 24bit recording - IF there is an audible difference. Because in reality even the best case difference is much smaller than the numbers 16 and 24 suggest. I was trying to put things into perspective.

werner

[nucleon]

As I mentioned earlier heres a 'midi' project replicated in FL, MuLab, Reaper and Live.

http://www.mediafire.com/file/dca8lvexl ... 20Test.zip

Uses multiple instances of the free Kontakt Player and its free Factory Soundbank. http://www.native-instruments.com/#/en/ ... kt-player/

Other free VST's I tried were buggy.

It includes the DAW project files, renders at 16bit, the Kontakt Player presets, and midi files for each part.

The rendered results sound exactly the same. The only mix difference is setting the FL channel volume knobs to -3dB (an increase of 2.2 dB). (Or you can normalize all the renders later).

A good way to test is to load all the renders in the host of your choice, as audio tracks/clips all playing at the same time. Play back while randomly soloing the different renders.

Note they dont phase cancel completly when using inversion to test, as the way midi timing is handled to VST's in different apps can vary by a sample or more.

As you can hear if you try, the canceling is intermittent, not a constant difference.

But if you render the stems from any of the apps, then mix them in any other of the apps as audio tracks/clips, the mix renders do then cancel.

(Note you dont need to alter the level of FL audio clips, as demonstrated in my earlier audio track/clip based test).

[sorohanro]

Nucleon, thanks for the test and explanations.
I guess now, that everybody can download the tests (yours and mine) the subject is pretty much over, so whenever appear (and probably will appear many times) same subject, we (or Tony/Gol) can give a link to this thread.

You deserve a prize

[nucleon]

Nucleon, thanks for the test and explanations.

No worries

[crimsonwarlock]

Of course. But that wasn't my point.

That was clear to me ... but I couldn't resist the result of your argumentation

[stonestreet]

I'm with jupiter8 on the bit-topic.
I'd never record in 16bit, because there's no reason and I like to be on the safe side.

But really, the difference is very small .. in most cases absolutely inaudible even after further processing. It's not like a 16bit recording would immediatly sound bad .. in most cases it will sound exactly like as it would in 24bit.

No, I think subtle context is what really needs putting straight here. Its not that 16 bit recording sound bad. Not at all, as mentioned, there are a lot of early digital tracks that sound superb. However, those engineers probably had to work pretty hard to overcome headroom issues associated with getting clean sounding stuff above the 16 bit noise floor. If you use 24bit properly (i.e. not redlining stuff all the time), you have a lot more headroom to play with, without losing resolution or getting an audible noise floor and consequently, have, for the most part, a much easier time producing a good dynamic mix. YMMV ofc. Its not really about what is audible to the end listener, its about how much easier it is for the mix engineer to work with the content.

STML i like that as an explanation, especially after watching that informative video.

Mark

[mauseoleum]

Nucleon, thanks for the test and explanations.
... blah ... blah ...
You deserve a prize

+1
(as I recently <..err..> "invested" in Signature upgrade from years of "fruity edition", your explanation is very much appreciated)

[Dreamw]

...Actually, there is a problem with sound quality in FL...

did you read / understand tony tony choppers post ?

Yep. Read and what he say it's very logical and figured myself years ago.
Tried myself several settings, same result

Again, I say. I'm a fan of FL. FL is probably the easiest program to work with midi, edit everything, all the tools are exactly where you think they should be.
Yet, there are some problems with sound

I use Max/msp/cubase/nuendo/acid/flstudio.

Fruity loops has a unique sound cant get on anything else things sound better played in edison than exported. I wonder if theyve changed the sound engine a lot as I notice the sound has changed in newer versions (i started on first version).

[Dreamw]

I'm with jupiter8 on the bit-topic.
I'd never record in 16bit, because there's no reason and I like to be on the safe side.

But really, the difference is very small .. in most cases absolutely inaudible even after further processing. It's not like a 16bit recording would immediatly sound bad .. in most cases it will sound exactly like as it would in 24bit.

No, I think subtle context is what really needs putting straight here. Its not that 16 bit recording sound bad. Not at all, as mentioned, there are a lot of early digital tracks that sound superb. However, those engineers probably had to work pretty hard to overcome headroom issues associated with getting clean sounding stuff above the 16 bit noise floor. If you use 24bit properly (i.e. not redlining stuff all the time), you have a lot more headroom to play with, without losing resolution or getting an audible noise floor and consequently, have, for the most part, a much easier time producing a good dynamic mix. YMMV ofc. Its not really about what is audible to the end listener, its about how much easier it is for the mix engineer to work with the content.

I was one of those early engineers. Noisefloor was not a problem in 16 bit - it was an issue with analogue.
However it is noticeably better to my ears to record/mix in 24/32 bit even if you mixdown to 16bit.

[rifftrax]

Let's put it this way: if you send in your mix to a mastering house (I'd say those guys DO know what they are doing) and propose to deliver your mix in 16-bit I'm pretty sure they'll advice against it

... although some will gladly take your money

Does nobody remember the 90's when mastering houses actually DID do everything in 16bit? You forget, that was the norm. Hit album after hit album in the 90's was 16bit all the way.

[UltraJv]

Maybe its time for someone to post these "night and day" differences in 16 vs 24 bit so we can all here them