Fruity vs. The World

[Guest bugbread]

Howdy.

 

Can anyone give me specifics of things that Cubase, Logic, etc. can do that Fruity can't? Not generalities like "Fruity is amateur", or "Fruity doesn't sound good", but things like "Preprogrammed gating".

 

Thanks.

[Guest Spindrift]

Fruity happens to be great at pre programmed gating, but the downside compared to logic (and to a certain extent cubase and sonar;-)) with most consumer software like fruity, orion or reason is...

 

Sound quality:

All DSP functions like mixing, sound synthesis and effects is generally of lower quality, mainly to preserve processor power to allow a broader userbase. This is what in the end turned me of using Orion, which I otherwise thought was excellent, and very quick and pleasant to work in.

 

Midi Editing:

The major old sequencer packages has a much more serious MIDI implementation, that allows for example customised swing, complex transform functions and advanced performance oriented arpeggiator-like functions like touch tracks or phrase synthesizer.

 

Advanced Functionality:

Logic for example can do things like extracting midi from audio, quantise audio files, and quantise midi according to the groove of an audio file.

 

Drivers:

A pro soundcard will probably now have very good drivers for consumer software.

If you use a soundblaster you be just fine, but i had problems running fruity using a pulsar card for example. Logic on the other hand will not work to well with many consumer soundcards.

 

Off course all the extra stuff mostly get's in the way a lot of the time, but when you get to know how to use a package like logic, there is a lot of stuff that really can help the creative process.

Althogh I can say I know Logic well, and can work very fast with it, all in all I would much rather use Orion, if the sound was good enough,the midi implementation better and there was drivers for my card.

[Guest Kits]

Hey

 

I use fruity as a midi sequencer. If you know in advance that you don't have a decent way of working with audio parts AND also not use the samples that come with fruity ( some of them aren't 16bit 44100 - for example the c_kick that many use ) AND use a decent soundcard (I use LunaII, a bit heavy but does the job) AND you know the workings of digital audio and midi you can do tons with fruity.

 

Just keep in mind that Cubase for example does have a decent sequencer for both audio and midi parts. Also in terms of stability 24bit applicatins(like cubase) are sometimes more stable than 32float apps... BUT - I do thing Fruity 3.56 is already far more developed than earlier versions.

 

I guess Logic would be the ultimate choice but I'd still take fruity over say Reason anyday.

 

Kits

 

P.S. For me the one most cool thing about Fruity is the midi implementaion.

If your music is based on external gear then Fruity is a great option for sequencing.

[Guest Andreas]

Just to nit-pick a little

 

MIDI implementation is the way a sequencer/program/synth acts on certain MIDI messages. It has nothing to do with the manipulation of MIDI data.

A sequencing program (Logic, Fruty, whatever) should implement MIDI in a way that allow the programmer to use EVERY single one of the controllers, program changes, whatever... Else it is not a decent app - I'm pretty sure Logic supports everything, and I would be surprissed if Fruity didn't also - it's easy to implement SUPPORT for these things - MIDI implementation is completely up to the producer of the 'ware, and can vary, without the one being better than the other... Boy, I'm in a mood today

 

tah for now...

[Guest bugbread]

Cool, great answers, I appreciate it!

[Guest Tim]

Sorry, I __HAVE__ to comment on this.

 

> Sound quality:

> All DSP functions like mixing, sound synthesis and effects is generally of lower quality,

> mainly to preserve processor power to allow a broader userbase.

 

This is utter and complete horseshit. Fruityloops realtime playing engine uses a lower-quality interpolation for re-pitched samples, and it does not set the 'High Quality' flag on the plugins (each plugin can use or ignore it as they see fit). This is ONLY for the realtime playing mode (in other words playing it in Fruity) and NOT the rendered output. It has been well established that FruityLoops CAN NOT POSSIBLY have 'worse sound' than Reason or Logic or whatever for the same input.

 

If you know anything about digital audio, you know why. Bits are bits. If you put the same sample or VST in Orion, Fruity, Reason (ok, no VST, but..) and apply NO FURTHER PROCESSING the output should be exactly the same. That is teh way digital audio works. Now it has also been proven (http://members.shaw.ca/stu.macQ/) that Reason is not sample-accurate and likes to muck with your input. I don't know about you, but I'd rather it didn't.

 

If you still think Fruity has inferior quality, please run this test:

 

1) Open Fruity

 

2) Load a VSTi (any will do)

 

3) Set the volume of the channel to 100%

 

4) Set piano-roll to play 1 bar of a note (middle C will do)

 

4) Render to 16 bit wav

 

5) Repeat for Cubase/Orion/Logic/whatever

 

6) Post the two rendered wav files

 

 

If you feel industrious, repeat this test for a sample and for a sample and a VSTi at the same time.

 

In other words, if you think this is true - prove it to me, I'd love to be proved wrong.

 

 

> Midi Editing:

> The major old sequencer packages has a much more serious MIDI implementation,

 

'serious' is a pretty dubious word, but I'll let that one go. On this most people (including the author of FruityLoops) agree. It was never meant to be a MIDI sequencer.

 

> Advanced Functionality:

> Logic for example can do things like extracting midi from audio, quantise audio files,

> and quantise midi according to the groove of an audio file.

 

midi from audio is interesting, but not really the business of a sequencer, IMHO. As for quantizing audio files, the Fruity Slicer does a good job of syncing rhythmic audio files to the master tempo or to whatever tempo you want.

 

> Drivers:

> A pro soundcard will probably now have very good drivers for consumer software.

> If you use a soundblaster you be just fine, but i had problems running fruity using a

> pulsar card for example. Logic on the other hand will not work to well with many

> consumer soundcards.

 

THAT is just weird. If the driver works and adheres to the driver API properly, there is NOTHING the app can do to work better or worse. FruityLoops ASIO support has matured a lot, and works very well. The MME support is as solid as any app in the market.

 

Now, I'm not just a Fruity cheerleeder (I just wanted to say that . I think Fruityloops has plenty of failings and drawbacks, not the least of which is their unwillingness to implement certain VERY KEY features like an Undo that works. However, your bit about DSP quality just raised my hackles because it is total BS. Never attribute to malice what you can attribute to stupidity. I'll go on the assumption you just don't know, or were misinformed by some other misguided person.

 

FruityLoops is a VERY GOOD program for quickly putting together pretty good music. It does not work in the absence of hard-work, determination, or talent. You need to provide those yourself. Despite what any old-school hardware people tell you you CAN get 'pro' quality (whatever that means) from it. But you need to spend more than a few minutes learning the ins and outs.

 

Just have to get the truth out.

 

Peace

Tim

[Guest Spindrift]

Hey Tim,

I'm not trying to tell ppl that they need to use pro equipment to make good tunes, just trying to explain my experience of the downsides of most consumer software.

When it comes to the sound quality, try a few diffent pieces of software whith mixing down a bunch of tracks to a stereo pair. When I have been trying, I compared logic, cubase and the yamaha 03d mixer. The difference was much greater than i could have imagined, i kinda thought bits was bits aswell before this experiment.

Sure, If you render a single track the result should be pretty much same in logic or fruity.

But when mixing signals, they need to be processed by a dsp algorithm, and due to the quality of this, I found it hard to be satisfied with the sound I could get out of orion.

 

I agree fruity is a good software, which has it's advantages compared to say logic, in it's user friendlyness, I was just trying to explain why ppl suffer with all the extra complexity.

 

I know there are several well know act's that use both reason and orion, and of course you can make great tracks with them, and the smoother workflow with simpler apps could matter more than small difference in sound quality.

I agree that all the advanced functionality go wasted in logic 99% of the time, and i could manage to get the same sound from orion as logic, that would prolly be my sequencer of choice.

[Guest bugbread]

Tim:

 

You may just be the person I'm looking for. You mention that while sound quality is in principle the same, that there are "plenty of failings and drawbacks", mentioning the UNDO (which bugs the hell out of me too). The other thing I notice is that it's hard to copy multiple patterns from the playlist. The end product produced by Fruity isn't made lower quality by these probs, but it IS more frustrating playing with sequencing because of it. Do you have any other specific issues where Reason/Cubase/Logic are superior?

[Guest Tim]

> But when mixing signals, they need to be processed by a dsp algorithm, and due to the

> quality of this, I found it hard to be satisfied with the sound I could get out of orion.

 

See this is the problem: on this you are right, but more wrong than you know. The 'dsp algorithm' used to mix multiple digital audio streams is a complicated one called 'ADD'.

 

zero output stream

for each incoming stream

for i = 0 to num_samples

output = output + input

endfor

endfor

 

It's not a very highly optimizable algorithm, nor do Logic's '+' signs begave any differently from Fruity's or Orion's. If there is an audible difference (I'm not saying there can't be) it is because of SOMETHING ELSE and is not innate to the sequencer core.

 

Tim

[Guest Tim Hockin]

> The end product produced by Fruity isn't made lower quality by these probs, but it IS more

> frustrating playing with sequencing because of it. Do you have any other specific issues

> where Reason/Cubase/Logic are superior?

 

I have a laundry list of little bugs and mis-features that I'd fix in FruityLoops if they would let me hack on their code (I've offered to do it for free, and I _am_ good at it). I can't recall them all (they're at home, I'm at work) but a few:

 

* Unlimited FX Busses

* Send to any FX Bus

* A proper Undo or at LEAST checkpointing

* Display actual values and units for all knobs

* cut-n-paste that isn't brain-dead

* don't scrub delay output when you click the playlist (Song mode - kills live performance)

* better layer control of VSTi channels

* loads of minor enhancements

 

I always talk about doing a free, open-source sequencer that keeps all the goodness of Fruity but is more flexible. I just haven't made the time - too busy using Fruity

[Guest Spindrift]

Tim,

Good idea, I just been ranting about that someone need's to do that lately. I think the time is ripe for a nice open source sequencer.

 

Otherwise I'm sorry for my obviously very offensive posts earlier, sorry for shitting on any fruity fanatics doorstep ;-)

I just tried to help bugbread with his question, and since I have thought about the matter a lot, I wanted to tell why I went back to logic after a short flirt with orion.

I still don't see what would be incorrect with the differences i described, except for maybe the wording sometime.

I can't belive that Tim and Andreas manage to write pages trying to pick my argument to pieces without actually managing to dispute any of my main points.

 

What is fruity lacking?

Top sound quality

Good MIDI functionality (also for using the internal synths and fx automation)

Advanced functionality

Support from high end hardware

 

Maybe I should add:

What is fruity good at?

User friendly.

Consumes litlle resources.

909 style drum programming.

 

About the digital audio mixing, if now a simple adding of values will do (not taking into account pan and volume control), how come mixing consumes a fair bit of dsp on my pulsar card?

All i know is my experiences with both orion and fruity, compared with logic.

Like I said, I loved orion to start with, but grew to notice that the difference in depth and separation was very obvious to me.

If using the same vst instruments and the same plugins produce a different result, what else could affect the sound.

I have tried mixing multiple channels together using different software, and there will be an audible difference between the ones i tried.

Would maybe orion or fruity do a lot of bit rate conversion with dodgy dithering or something.

 

I thought that digital processing is getting quite advanced, and most filters for example now work on emulating the sound of the analogue designs, rather than making a strictly mathematical operation. Could it be that some software mixers actually contains algorithms that tries to recreate what a for example a Neve woluld do to a signal passing thru.

[Guest Tim]

> Good idea, I just been ranting about that someone need's to do that

> lately. I think the time is ripe for a nice open source sequencer.

 

you volunteering?

 

> I can't belive that Tim and Andreas manage to write pages trying to pick

> my argument to pieces without actually managing to dispute any of my

> main points.

 

How can I refute what you claim "sounds better". It is totally subjective. I asked for you to prove it and show us the proof. Nothing.

 

> What is fruity lacking?

> Top sound quality

 

Again no proof from you or anyone that it is inferior to anything, just 'experiences' which fly in the face of the facts of digital audio.

 

> Good MIDI functionality (also for using the internal synths and fx

> automation)

 

Huh? 'Good' is subjective. It is NOT a MIDI sequencer, so this point is essentially moot. Are you further stating that automation is weak? can you expound?

 

> Advanced functionality

 

THERE is a vague bullet point. Are you, by chance, in sales or marketing?

 

> Support from high end hardware

 

Again, bollocks. Name one Audio card that supports MME or ASIO that does not work in Fruity, or works worse than any other app? Now can you support that? I'd settle for a half-dozen other people with teh same complaint as proof enough.

 

> About the digital audio mixing, if now a simple adding of values will do

> (not taking into account pan and volume control), how come mixing

> consumes a fair bit of dsp on my pulsar card?

 

When adding multiple streams, you have to do a lot of work. At 44.1kHz, you have to do 44,100 additions per second for EACH STREAM. Woops, you want stereo? Make that 88,200. Combine that with having to wait for IO (memory is SLOOOW). Do you do any processing? Adjust gain? There is a multiply for every add. Let's not talk about reverb, delay, etc processing. Just because mixing waveforms is EASY doesn't mean it is free. It still takes a lot of CPU and bandwidth to do.

 

If you are refuting my statement that mixing digital waves is an addition, I beg you to PLEASE read a bit on digital audio.

 

> All i know is my experiences with both orion and fruity, compared with

> logic.

 

As long as we're clear that your statements are experiential, I can accept your opinions. I just don't want anyone spouting crap about how bad the mixing algorithm is without understanding what it is.

 

If using the same vst instruments and the same plugins produce a different result, what else could affect the sound.

 

Processing - any reverb? How about minor volume adjustments. Psychoacoustics show that if you lower the volume of a sound it sounds more muddy and less sharp.

 

> I have tried mixing multiple channels together using different software,

> and there will be an audible difference between the ones i tried.

 

Can you put the two results on the web somewhere, with details about what you did? I keep asking people to do this, and no one ever does.

 

> Would maybe orion or fruity do a lot of bit rate conversion with dodgy

> dithering or something.

 

I dunno about Orion, but I suspect it is no different in this respect. Once the Fruity core takes data it is 32bit floating point data. It does not get dithered or converted EVER until it gets sent to your soundcard or rendered to final output. (one exception: any VST can do whatever it likes to a stream, including downsample, dither, or truncate - but most don't).

 

> rather than making a strictly mathematical operation. Could it be that

> some software mixers actually contains algorithms that tries to recreate

> what a for example a Neve woluld do to a signal passing thru.

 

Any app that tries to do that would be PURELY BROKEN in anyone's opinion. If you want to emulate a sound you make it an effect in your chain. A digital mixer should do NOTHING but mix. As far as I know, no app does that.

 

I'm not trying to be a bitch or say you're an idiot, I just want to see some proof. Convince me that I should buy Logic, and I will. Impress me with the sounds.

[Guest coriolis]

dare i interject in such a heated thread... i just want to mention the issue of wordlength (16bit, 32bit, etc), which has big consequences when you talk about digital mixing. i don't know what fruityloops internal mixing wordlength is, but that will definately affect the way it mixes signals. see, with digital signals, a "+" is not a "+" across the board - the difference lies in what happens to the extra precision that is lost when you add two signals, divide them by two, and then chop off the remaining bits that exceed the precision of the wordlength. the lower the wordlength, the lower the number of possible values a single sample can be, and the more signals you add in a low-wordlength algorithm, the more precision you lose on each signal. the result of this is a reduced sense of spatiality and clarity, and this happens all the time in every DSP operation that doesn't just pass the digital signal directly from in to out.

 

spindrift - this is why the pulsar card takes up so much dsp in mixing - because it is doing it at high wordlength on all the signal paths.

 

logic uses 32bit internal mixing, and even then i hear the effects of mixing (dampening the sound, reducing clarity and spatiality). i don't know about fruity or cubase, but this is an important factor in the quality of the mixing. if fruity is 32bit, then i would expect its mixing to sound as good as logic. if its less, than it'll sound worse, especially with quieter signals.

 

disclaimer: i don't claim to be an expert. this is my understanding that i have developed through reading technical information on the subject. if i am wrong, correct me (for myself and for the sake of others), but don't flame me for this post!

[Guest coriolis]

i just realized that one important point may not be clear - digital mixing is three operations: addition, multiplication/division, and rounding/truncating.

 

say you have two signals at full volume, 16bit. if you added these two together by simple addition, you're resulting signal would require 17bits of precision, because it would be twice the first two signals in amplitude. you then have to divide the signal by two to get it fit within 16bit precision.

 

getting mathematical: 65536 (16bit max value) + 65536 = 131072 (cannot be represent with 16bit number, requires 17bits, and in fact 131072 is the largest number than can be represented with 16bits).

 

so, if your output signal is 16bits, guess what - you have to divide your mixed signal by 2 to get it to fit within the bounds of your output.

 

now, lets say your first two samples were 65536 and 65535, your mixed signal would be 131071. divide by two, and you have 65535.5. What do you do with the 0.5? throw it away or round up or down! and along with that 0.5 goes some of your signal quality, especially the small, subtle stuff that is important to the ears to hear things clearly! and thus you have it: the cons of digital mixing - it necessarily reduces signal quality.

 

the more bits you have to work with, the less you need to worry about this degradation because the signal has enough bits to handle the loss. but with low precision numbers, the loss is quite audible.

[Guest Tim]

Coriolis is pretty correct. To expound a bit for the non-technical folk (bear with me, please):

 

Audio waves are represented as a long stream of samples, each sample representing the amplitude of teh wave at some point in time.

 

Samples are generally represnted in one of a few ways:

1) an integer (whole number)

1a) 16 bits (values -32768 to 32768)

1b) 24 bits (values -8388608 to 8388607)

2) a floating point number (decimal)

2a) normalized (values -1.0 to 1.0, but can hold really big numbers)

2b) non-normal (really big numbers)

 

For the sake of argument , trust me when I say that MOST apps use normalized floating point or integers. Let's go with 16 bit int to demonstrate the issue.

 

As has been discussed - mixing two waves is a simple addition. If you have two samples: one of value 123 and one of value -6246, when you add them together you get -6123. Repeat this for every sample.

 

The problem arises when you have two samples that are too large together. Anyone who has taken a physics class knows what happens when waves resonate. If you have two samples: one of value 16000 (about half of max) and one of value 20000 (a bit bigger than half), when you add them up you get 36000. The problem is that a 16 bit integer can only hold numbers up to 32767. You end up with what is known as an overflow. The counter wraps around to 0 and you get something crazy, like 3233 in this case. Clearly, a few of these and your accurate digital signal becomes crap.

 

This is why most apps mix in a larger wordsize than the data. In 32 bit floating point, mixing two maximum value signals (1.0) get's you 2.0. You can retain the accurate value until the last possible moment. This gives you a chance to (for example) compress that value or gain-reduce it later.

 

Let's look again at the int example. If we could keep the 36000 value, and later apply a gain of -6 dB (50%) we'd have 18000 - well within range. But if we truncate too early, we get 1616. after gain reduction.

 

Thanks for humoring me. I hope it's clear.

 

Tim

 

For the record: Fruity is 32bit normalized floating point, as are most apps on the market.

[Guest Spindrift]

>you volunteering?

Sure, i'll make a special neve emulated mixing algorithm for it :-)

seriously, donno any dsp coding (as you might noted), would like to get the time to get into it though...

 

>THERE is a vague bullet point. Are you, by chance, in sales or marketing?

Read my earlier post.... you don't seem to care much for audio to MIDI and audio quantsize functions yourself, but maybe realtime transform would come in handy for your setup, maybe you even need sound diver for keeping your sounds for your module with the song. There is a lot of stuff, that is left out in software like fruity, because very few users actually bother to use it.

 

>How can I refute what you claim "sounds better". It is totally subjective.

>I asked for you to prove it and show us the proof. Nothing.

Perception of what sounds good is subjective, yes.

The reason I started to question the concept of bit's being bit's was by a coincidence.

I was working in a studio running a Mac with cubase and korg1212 card going digital via a stereo spdif cable to a yama 03D mixer.

When installing a ADAT card to get eight individual digital out's to the mixer, we happened to have a project running that was eight tracks of audio coming straight from the hd, with no fx or processing applied.

As soon as we changed the output to go thru the adat connection, the sound opened up. Both me and the other producer started to question how this could be.

After some experimentation we concluded:

Signals was going from 16bit audio files to the 32bit environment of cubase and either

a) mixed in the application itself and converted down to 16bits to be passed on to the 20bit environment of the 03D

converted to individual 20bit signals, going straight in to the mixer via ADAT.

Thinking that it did seem like the difference of mixing it all down to 16bit, and send that to the mixer, rather than having the signals go individual as 20 bit. This would indeed be logical.

But to try it we took the same audio files in logic, and did the same test.

To our surprise, the difference between the signals going thru adat and spdif was no longer audible to us.

The only thing I had left to conclude was that somehow, when mixing a number of tracks in cubase, the audio would sound less separated and clear, compared to doing the same mixing with either the yama or logic.

I also found the same when working with both orion and fruity, using the same tools, the same way i do in Logic, but to my ears with somewhat different results.

 

I haven't tried the latest versions of any other sequencer than logic for a while, but if for example fruity haven't improved, and you want proof, go ahead, try mixing together some 8-16 well recorded racks with good stereo depth and send them to me, and I (and maybe you as well) spot the one made with logic.

 

>As long as we're clear that your statements are experiential, I can accept your >opinions. I just don't want anyone spouting crap about how bad the mixing algorithm >is without understanding what it is.

No one cares for the theory in the end, it's experiencing that the software is made for, so thats the best way to compare and evaluate.

Do you mean that information based on book or magazines is the only information that count's or whats your point.

 

>Processing - any reverb? How about minor volume adjustments. Psychoacoustics >show that if you lower the volume of a sound it sounds more muddy and less sharp.

Duh...ahh yeah....had loads of reverb on all the tracks, thats why they didn't sound the same....comeon....

I know about how phasing and masking of sound can do funny things when you change the volume on individual tracks, but I'm talking about making a lot of tunes with the software, as well as testing it as described above with flat levels.

 

Everything that has to do with an audio software app is subjective, that's probably why you haven't provided any 'proof' to counter my claims. The proof is in the pudding in this case.

How well the gui works and how good it sounds can only be determined by actually using the stuff and forming an oppinion....so actually instead of you asking me for proof you should answer the following:

Have u ever worked with logic extensivly?

What's your experience with audio engineering?

What music do you have to show for you experince with various packages?

 

If you want some sound from me:

Search for spindrift on dc and you should find breakpoint (orion) and latitude(logic)

Steptime-Timeloop(orion), Steptime-Drama(orion), exotonic-creative alternative(analog roland mixer)

If you find, leo&lumen 'in the air' (cubase), 'anyway you arrive' (cubase) or

Lumen-'Overboard'(cubase)

[Guest ms]

Listen all of you

 

There IS a difference when mixing in different sequencers.

 

In fact even the Cubaseprogrammers claim that Cubase VST 5.0 sounds "messed up" compared to SX.

I have worked with Yamaha 03d + hardware only before and it sounded much better than "pure digital" soft's.

The main issue is the mixing: It's much easier to make it sound good on a hardware mixer. I dont really know WHY, but ask any musician that worked with both soft and hardware and you'll see.. or try yourself!

 

I also did a test loading the SAME song with the SAME VSTi's in both Cubase VST 5.0 and SX.

I was surprised how much difference I could hear! SX has a much more distinct and clear sound... Don't ask me why and how. It's just my experience..

[Guest Spindrift]

word up ms :-)

[Guest Kits]

Hey

 

Isn't buzz open source?

[Guest coriolis]

ms: can you record your experiment with cubase vst and sx? i'd be interested in hearing the difference as well. if you do record it, try to do it at high sample and bit rates.