Arturia V Collection 6

[ghettosynth]

I have synths that, technically, I'm sure you would say are "inferior" synths that I enjoy using. Because for me, they sound "good enough". The older Arturia synths are a perfect example of that. I like the 2600. I know it's not an "accurate" emulation but I don't care. I can still use it to make music and I'm sure the layman out in the street listening on his iPhone wouldn't know the difference anyway.

But I do understand the desire and enjoyment of tech talk.

I think that people don't get the parameters of weak filter modeling, aka, our friend who doesn't seem to understand what objective means. Not being modeled well is not a recipe for universally bad sounding patches. It means that some kinds of patches will sound much worse than on a better model and on some patches the differences will be less noticeable.

The 2600 is fun and, as I've commented before, I really like that it includes the sequencer model. However, when you rapidly modulate the filter and/or increase resonance, the weaknesses definitely make themselves known.

The 2600 is kind of the metric for Arturia's commitment to their old models for me. Had they upgraded all of the old models to more contemporary specs, I probably would have forked over the $199 for this release.

[ghettosynth]

So you used Reaktor for a while and learned something about circuits and math...
And then, how do you objectively prove the accurateness of Diva or the Legend using math?
Btw. going to read this bullshit tomorrow... gotta sleep

well, even if the used had access to the source code of both synths, it couldn't be proved that it's a more or less exact emulation, but only if the filter is a "good" ZDF filter or not.

Which, of course, I never claimed. I said that we can use mathematics to show that some models are more accurate than others which is a true statement.

[waltercruz]

You are probably talking about presets that take advantage of more modern features, which indeed can take some CPU. But presets that use just the FM synthesis as in the DX7 don't take that much (although they take more than DeXed and FM8, nevertheless)

Not really. I'm a simple man: I loaded that cliche Electric Piano (I'm a gigging musician, and eventually I need this patch for some gig.) Following the tips from Ingo, turning off the unnecessary filters, I got it almost usable, but I still have some CPU problems. My computer is not that good (Macbook i7 2.4GHz), but even so I think that this CPU usage is high. But it's the first release of DX7 S, so, I think that some optimization will eventually appear.

[waltercruz]

I found the CPU usage very high.

It eats quite a bit of RAM too.

Observe:

DX7 V, circa 245MB per instance. A gig of RAM gives you ~ 4 instances.

DX7V-RAM.gif

To compare: FM7, circa 30MB per instance. A gig of RAM gives you ~ 34 instances.

FM7-RAM.gif

PS: This is on a 32bit system. On 64bit its probably a smidgen more.

I think that the high RAM usage is the images of the GUI. It's something that happens on all >V5 Arturia synths.

[ghettosynth]

User powerdx7 from youtube has done a comparison between DX7 and Dexed. The first 4 minutes of the video he compares 3 factory patches and show the very little differences between DX7 and Dexed. Maybe this video can be used as a reference for comparison (at least for those 3 patches).



I have a TX-7 packed here, someday I will turn it on to make the comparison myself.

I don't know what's going on there but there are gaping differences in the bass response.

[waltercruz]

well, even if the used had access to the source code of both synths, it couldn't be proved that it's a more or less exact emulation, but only if the filter is a "good" ZDF filter or not.

Which, of course, I never claimed. I said that we can use mathematics to show that some models are more accurate than others which is a true statement.

agreed.

[chk071]

Which, of course, I never claimed. I said that we can use mathematics to show that some models are more accurate than others which is a true statement.

How?

[Ingonator]

After watching a tutorial video for DX7 V i found that at the envelopes page you could also switch the standard DX7 envelopes to a DADSR (= ADSR with adjustable delay) or a MSEG mode (in the envelope page it is a drop-down menu at the upper left of the envelope display).
In MSEG mode you get lots of options to make the envelopes as complex as you like including adding new envelope segments/points, editing the slopes and setting a loop point for the envelope. Also a Sync mode seems to be available.

Besides that each operator besides it own filter could also have it's own feedback. In the real DX7 there was only one feedback in each algorithm.

You also get 2 additional mod envelopes that could be routed in the mod matrix (e.g for the filters).

The MOD page besides a mod matrix also offers 2 LFOs and a step sequencer (besides the Arpeggiator found at the main GUI).

As alraedy mentioned overall there are 25 waveforms to go beyond the usual Sine.

The overall quality could be improved using the "Modern" DAC mode and setting the velocity mode to "Full" which allows using the full velocity range of 127 instaeed of 100 like in the first DX7.

The Unison (up to 4 Unison voices and voice detune) could rseut in really nice soudns too (but uses morre CPU of course).


While DX7 V could sound and work like a real DX7 (including Sysex import) it could go so far beyond the original synth that due to that a comparison to a real DX7 gets almost pointless.
Overall the amount of advanced features even seems to go beyond that that in FM8.

[Ingonator]

Like many others i am a fan of the classical DX7 E-Piano factory preset (original name "11 E.Piano 1"). Based on importing the orignal Sysex file i made some changes within DX7 V (similar changes were also done to some other patches):

- Changing the DAC RES quality in the global settings from "Vintage" to "Modern"
- Changing the Velocity mode (bottom of the GUI) from "DX7" to "Full" to get the full range of 127
- switching off the filters in all 6 operators as they are not needed
- adding Unison (2 voices) and a little Unison detune (around 8 %)
- adding a smaller amount of built-in Reverb (around 15 %) at the FX page

That patch is also the default patch loaded with DX7 V when using a new instance but this does not include the changes mentioned above. I started this by importing the original Sysex file.

I attached the edited single preset file for DX7 V to this post.

[Tj Shredder]

It is the opposite, only by measuring (of more than one original machine; which is of course possible) you can really prove superiority of an emulation over another in an objective way.

As a scientifically trained person I would totally agree to that statement. But with instruments it seems, that objectivity is not of that much value. I also think it is not so much about difference in sound, that might be an illusion. In the past I had an Arp 2600 which was a very impressive experience. I can't complain about the sound of Arturias emulation, but the taktile and spatial presence is a very crucial factor. I never had that almost spiritual experience with the emulation!

Also it is not important if there is a difference in sound, normally you would go for the better sound... The original noise of a DX 7 is not of much musical value, even if you think it is, just run it through a bad mixer...

Most of the analog icons of the past had to fight technical limitations which are no issue today. Getting rid of those would free the creative possibilities. Thats why I am not at all interested in another Prophet or Mini Moog clone. There are enough of those around...

I am interested though to find out what makes the difference. The only valuable test would be a double blind test with real knobs, one controlling analog hardware, the other the simulation. The probands would only have to answer two questions, what they believe is the analog, and which one sounds better.

[Tj Shredder]

I don't know what's going on there but there are gaping differences in the bass response.

Youtube will certainly compress and change the sound. Its more an instruction how to do a comparison.
Its also not quite clear, if he used the same DX 7 as MIDI controller, and just recorded the same hit he was playing (but I guess so). The next test would be to normalize levels and subtract a phase reversed version from each other in an audio editor. Then listen to the residium...
The difference in beating frequencies shows a clear error in interpreting the parameters correctly. There must be something fundamentally wrong. Its probably hard to get the exact envelope curves, but that would not explain differences in beating...

[Ingonator]

Based on the modified DX7 E-Piano 1 patch mentioned in my previous post i created a pad sound that i called "EP-ic DX7 Pad 1".
I attached the single preset file for DX7 V to this post.

[aMUSEd]

Thanks Ingo

[Stefken]

Regarding DX7 presets. Anyone would not happen to know some good 'Electronic body music' syx resources?

Front 242 was a Belgian Body music band that used the DX7 a lot.
Some typical sounds are "bright basses" like in the included preset "Growling bas".

The song "Headhunter" demonstrates the typical sound of Front 242.

[jme-audio]

It can be mathematically shown that some emulations are more accurate than others

Aha, by what?

LOL! Cmon man, you're clearly out of your depth and you know it. You're just trying to save face at this point and it's not going to work. Read "The Art of VA Filter Design" and educate yourself. From there you can take a look at recent models in Reaktor, no reverse engineering necessary, and also older models in Reaktor, again, no reverse engineering necessary. Spend some time there and you'll see for yourself how models and circuits relate and how we can show that some models are more accurate than others via mathematics.

This all started because you didn't understand what the word objective means. You aren't going to convince anyone now that you know anything about circuits or mathematics. Anyone that does wouldn't have made that mistake in the first place.

So you used Reaktor for a while and learned something about circuits and math...
And then, how do you objectively prove the accurateness of Diva or the Legend using math?
Btw. going to read this bullshit tomorrow... gotta sleep

I said nothing about me, I gave you a path to understand how math and circuits relate. With every post yo make you give away that you have no command of the language of science. I'd let this go if I were you, you've been off topic for pages.

I am not interested in any recommendations about programs and books that you have never read. Also there is no need for a meta discussion about science theory or objectivity (let the philosophs do that...).

As you didn't provide any answer to the question (no suprise...), we're surely done with this.