Arturia DX7 V vs Yamaha DX7 done RIGHT! 

Exactly hardware snobs are always trying to justify their overpriced purchases to people who can't afford expensive synths this is harmful behaviour to beginners because they're telling them without overpriced gear they can never make great sounding music which is false

thanks grant, was not aware of that. somehow nerve-racking when other sites pick up on my stuff. cheers.

7 is Howard Jones - Pearl In The Shell ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-sTRuoXEbvjE.html

Would be cool to see the preset with each song too.

Jaime Gonzalez I second that. He gave the first 2 and the 4th away in the first video but I’m curious to know the rest.

I am with Jamie Gonzalez. I’d love to know which preset each was.

truth

The shape of the waveform in this case just tell you the amplitude. Anyway, there's no significant differences

Actually no. If you look at ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-iHycGwuWcn0.html you can clearly see that there is a case of inverted polarity going on. So they are not "identical" as in photo copies. Still they can sound the same if the polarity is inverted if they are just "mirror images" of each others (as much plus as minus of each other). But then you can see from other patches that they aren't identical (when no inverted polarity). Still they sound very, very close almost identical.

I could tell the difference on some (real thing sounded just a touch nicer) but on others not.

i thought so too, it's quite surprising how the plugin matches the shape of the original, even when there is some very low frequency DC offset.

agreed. thanks robin, same to you.

you could get a TX7, much less weight; and having a computer, you can easily program DX7/TX7 using DEXED as programmer software.

Dexed is even better

Mbc ‍ Agreed.

I think it is much, much more down to how the DX7 internally computes the algorithms. I don't think that they "open sourced" every part of the engine. Yes you can reverse-engineer it, but there still will be parts of it where you have to guess how much and in what way one part reacts to another (over a whole range of values). Add to that the rounding errors in the original DX7 which couldn't use large floating point numbers (like in a modern DAW and VI) for every calculation, which would result in a wave that would be different since there are loads of these calculations(often multiplications for example) going on. I think that it's very easy to always blame it on the converters when these comparisons are made.

I really think the differences is neglicable. Yes it would be better if they'd completely reverse enginered the original Yamaha chips and build a fpga version, but that just wouldn't be feasable if they are to make money of it. If you just need a single or a couple DX7 patches for a song and you don't have an original DX7 in your setup no matter what people say, you are not gonna be able to hear the difference in the mix if the person used a DX7/DX7II or a software product like DX7V or Dexed. Yes you can hear the differences in an a/b test. There is more body in the Original, the Dexed is thinner and sharper, where the Arturia is wamer compared to the real DX7. Yes You will know what you used, but a normal listener would neither be able to hear the difference, nor would he bother if you told him. Who cares if the result sounds good.

i agree.

+Edward Cotter You mean the Octave Plateau Voyetra 8 in all of its "stair step" "zipper noise" glory? Thats why I got rid of mine. The envelopes were digital and way less than 16 bit, you could actually hear the slow decay "stepping down" as the sound faded. The oscillators did have a wide bandwidth sound.

Richard Hill I don't know... Always sounded great to me. New Order managed to do some lovely, airy things with it. I hope you got good money for it though. They go for absolutely insane prices now when you see them(not often) which is why I'd love a VST version.

i also appreciate the convenience of the plugin, but remember it's not just the sound, you got the physical presence of the real thing, that keybed... etc.

In addition when you track in the studio or live line through a system you would never hear the difference Most of us have preamps of decent quality many interfaces of course have them built in

I have a dx7 and would never use the words warm or rounded to describe it. To me it's usually people that have no real experience with hardware that act like hardware snobs and say silly things like that in comment sections(this comment section has quite a few examples of stupidity on display) this is a great plugin and done well enough to fool anyone, especially within a song. I look forward to to purchasing it.

@@leepicciotto Exactly. I had an original DX7 when they first came out. This goes way beyond the now laughable 8 bit cpu. Does it sound exactly the same, 95 percent. Can it sound a million times better with all its extra features and vastly superior quality sound mode, 1000 percent it can.

Jamie Bales I have lots of real analog hardware and it all beats plugins. Some more so than others. I don’t have a DX7 so can’t comment on that but I want to get a TX7. I have a few analog synths and even the very best plugins don’t compete. There is another factor in this though. The A/D conversion when you record this stuff. I have a Mytek interface which is very high quality. The differences then become much more apparent. Recording studios don’t use the cheap converters than many people use on RU-vid and you also critically need high end playback speakers or headphones. Then it becomes night and day. If you are listening via low end quality gear then you won’t hear the difference.

And if you just bought the whole V Collection 5 collection a month ago, you can now upgrade to version 6 for $199 more. I thought that was too high. But I could not say no.

Totally get that "could not say no" feeling! But I also think that the upgrade price is too high - common complaint I have seen on blogs, social media, etc. I think for the 'Collection 5' that $99 would have been so much more equitable, with $199 for earlier releases, etc.

Blind tests are a bit problematic, as some people know what artifacts to listen for. But most would have no way of knowing. And withint a track, with effects and other sounds surrounding it, I don't think it would make any real difference.

Where do you find the sysex?

hi, thanks, you can download factory presets from bobby blues.

damon, both of those plugs are available as free trial versions, so you could easily do your own comparison :)

that setting was on vintage yes, i think i did a demo showing the difference, somewhere, sometime. the only difference i could hear is added hiss.

+1

It's called an eq,use it instead of complaining about a damn DX7 emulation. It's digital bro,you can use it just like the real thing. What a stupid comment.

That makes sense.

you'd both need the midi and the audio right? it's a 45 minute long track before editing... but I can make it available if you still want.

Hi Woody, Midi only is ok, I can compare with the audio from the video.

all presets in this demo from the rom1 factory bank

Ah, found them. They're Arturia recommended patches, so they are prioritized at the top of the library and don't appear alphabetically with the rest. Thank you!

Sure the DX7 mode should be called DX7 Emulation. And they could have added a Controlled by DX7 mode, that is in reality the same as the full mode, but just so that users having a DX7 or midi files from one, would select the right mode.

My main bone of contention is when presented with MIDI notes which max-out at 100 velocity, in 'DX7' mode the VST does not play its full range of velocities. So, contrary to what one of its co-developers said in the comments of the previous video, it would not sound like a real DX7 if played directly from a real DX7, whose velocity maxes-out at 100. Instead you have to select 'full' mode. Then, it would. The two modes appear to operate exactly opposite to what you would expect. I think it's a bug and Arturia just don't like to say so.

The "DX7" velocity mode, is "DX7 Velocity Emulation Mode". Since the DX7 doesn't use the full velocity range, the "DX7 Velocity Emulation Mode" re-scales the velocity data it recieves so it corresponds to the value the DX7 engine in the DX7 hardware would receive for its own keyboard. In this case the midi data comes from a DX7. So what happens when DX7 Velocity Emulation is performed on the output of a DX7 that doesn't have full velocity? Well the values will be lowered. That is the case since if it recieves 100, the software will think that is 100 out of 127, not 100 out of 100 and re-scale it to a lower value that is a certain percantage less than full value of hundred on the DX7. The DX7 sends the value 100. It doesn't send a message that means Max Velocity (that is not how midi works). And since the maximum actual value of midi is 127, there is no way for the software to know, that 100 was the maximum output of the device that was used to create that midi data. There is no such communication going on. The Full mode allows to send values to the DX7 V engine that goes beyond what the DX7 with its own keyboard. This could be used for experimentations. But would not be the option you would want if the goal is to emulate a DX7 as closely as possible. They could have called it "Full Velocity Experimentation Mode" or something like that. But it happens that since it doesn't re--scale the midi data, it is also the mode that should be used if the midi data comes directly or recroded from a DX7. They could have added "Controlled by DX7 Velocity Mode" that works the exact same way, so that people that use it with DX7 midi data, will know what to select. I would say it behaves as it should. Very few people will control a DX7 emulation software with the DX7 as a controller, since they then already have the DX7 at hand. And due to the velocity limitaion, it would not be the best controller for other synths. The issue is that the naming isnt perfeclty clear, at least not without reading the manual... I haven't read the manual so I don't know how they describe these modes in it.

Jörg, I encourage you to download the free demo and test it out, you can experiment with the new features and play around with sysex etc.

Rod Beauvex - “What A Fool Believes” - Doobies

Ahh, thanks. Piano playing that riff, that's what threw me off.

I would say that the actual issue here is that he is using the rather unexpected scnario of having the output midi data of a DX7, compared to what most people will have (the output of a keyboard with full velocity).

yes, i think you nailed it there Jon, I believe the real issue her is that this setting is saved per patch, and not globally.

Thanks to both for clarifying!

The hardware is a pain to program. in most peoples minds. Dexed and Arturias are better for programming. Dexed vs Arturia is a bit of personal preference.

yeah i agree with jon there. i loved the real hardware, was a pleasure to own and play. but when doing multitrack recording it's a pain with all the cables, and impossible to edit the tracks after recording. so i found myself using dexed even when the real dx7 was in the room.

yes it takes a few seconds for me too. difference between commercial software with heavy frameworks, DRM etc and a lightweight dll. I don't even know how large the dx7v installation is, whilst dexed is just 2mb. dexed wins in that respect.

Nah, two different Yamaha DX7 would sound identical. They are all digital. And at the output stage they use the same DAC. I think you confuse this with what happens with analogue synths where there could be a great difference. There are no analogue filters in the DX7:s. In fact the whole synthesis chain is all digital. I would be extremely surprised if anybody could hear a difference between two DX7:s (the originals not the later upgraded versions). But on the other hand there are some differences in how the emulation sound here compared to the DX7. Yes most of the sounds on them sound virtually the same, but the "The greatest love of all" sounds clearly (Well, slightly) different on them. Definitely more than a DX7 vs another DX7.

Magnus Lööv Yep, I agree that two DX7s should sound almost the same, but I just meant that we are VERY close to the sound in VSTi plugins like Dexed now. And, I appreciate that the SID chip was probably a bad example in this case. lol I know that the DX7 / YM2151 is basically a fully digital chip, and uses an on-die ROM table for a quarter of the root sine wave etc.

I spent quite a lot of time in recent months reverse-engineering things like the LA32 chip (MT-32 / D-50), and the SID chip, so yeah, the SID was a bad example, because the output filters can actually sound quite a lot different between batches. hehe There is an "average" of the SID sound though (for both the 6581 and 8580), but that exposes one of the main issues with "cloning" it on modern hardware like FPGAs - everyone remembers their own SID chip sounding slightly different to the "average" one, because it probably did. For the LA32 btw, rather than staring at the die shot or logic analyser captures for weeks, I was trying to think of a much faster way to be able to watch the register writes in realtime... I finally tried something I've been meaning to test for years, by hooking up the CPU bus of an LA32 board (in this case from a Roland D-10) to the FPGA board. I could then have the FPGA display an exact "mirror" of the LA32 registers via VGA, which made it SOOOO much easier to figure out what was going on (as I played notes on the MIDI keyboard, and twiddled the parameters etc.) imgur.com/pBFNEzw imgur.com/33wKe1L

(for anyone still reading. lol)... It turns out that the LA32 registers are in groups of two bytes, so pretty much everything is handled as 16-bit values internally. For each of the 32 voices, the CPU simply writes the values for the filter cutoff / resonance, the pitch of the voice, whether it's a PCM or "synth" voice, the PCM sample index (if used), and some other bits for the chosen algorithm etc. Interestingly (as partly mentioned in the MUNT emulator project), the LA32 also assists the main CPU by generating an interrupt every time the "ramp" values for the TVA reach the "target" values. (using the top set of 32 two-byte values, shown in red. One "word" per voice, with the first byte being the starting ramp value, and the second byte being the target)... imgur.com/r1Ienjx (So, 8-bit values for the ADSR / TVA envelope.) As soon as the ramp value reaches the target value, the LA32 triggers an interrupt, then the CPU then just updates the next values in those regs, so it can do the full ADSR cycle. As far as I can tell, it doesn't do the same thing for the filter regs, so the timing for all of the TVF stuff is just handled directly by the CPU itself. I'm hoping to eventually implement a 100% clone of the LA32 on FPGA, so more people can enjoy things like the D-50 sound on real hardware, without having to fork out the expense of the original synths. ;) (preservation is also a big part of this, as I'm hoping more and more of the digital synth designs can be reverse-engineered and captured for all time.)

Good ears! I definitely also noticed that the HW DX7 is a little sharp too. This is probably adjustable with a "tune" knob on the original yeah?

my hunch is that this is a bug, this is a global setting that should stick as you change presets.

Woody Piano Shack - I'm a relatively new user of the full Arturia V Collection 5, and was trying to decide if the DX7 V was a good enough plugin for me to do the upgrade to V Collection 6. Due to this demo video I decided to do the upgrade which is now installed. It is interesting to see that many of the presets default to Velocity "DX7" and many default to Velocity "Full". Since upgrading my demo DX7 V to the paid version, it is easy enough to just save the preset after making the switch to Velocity "Full". It then is always recalled as desired. Thanks again for your hard work making these demos...

i'm not sure it is a bug. The full mode is useful for experimenting with full velocity. Or it is useful when the midi-file or controller is the output of a DX7. But for most cases people would probably want to emulate the DX7, from a full velocity keyboard.

interesting that it's saved with the patch. that's a good suggestion to overwrite, would have saved me some time.

jon, typically you would set that according to which controller you are using, then you don't want it to change every time you select a new patch. seems like should be a global preference to me.

we think the dx7 might be slightly out of tune, it never had a 0 setting, it's just a slider...

ahh.. that might be it then ! I think I might be my old brain adding in my head the other artefact - that of the pitch going down slightly throughout a note being played... :)