| Author |
Message |
3phase
Joined: Jul 27, 2004
Posts: 1184
Location: Berlin
Audio files: 13
G2 patch files: 141
|
 Posted: Mon Jan 22, 2007 8:07 pm Post subject:
|
  |
|
I just created that file...nothing on modulations yet but the dx oscillators are set..envelopes was all ok... but scaling went wrong an vel sens parameters wasnt applied...
I had to rearrange the operators to ease the hand editing
| Description: |
|
Download (listen) |
| Filename: |
Brass1.pch2 |
| Filesize: |
3.81 KB |
| Downloaded: |
1823 Time(s) |
| Description: |
|
Download (listen) |
| Filename: |
rom1ab1.pch2 |
| Filesize: |
4.26 KB |
| Downloaded: |
1733 Time(s) |
|
|
|
Back to top
|
|
 |
ian-s
Joined: Apr 01, 2004
Posts: 2672
Location: Auckland, New Zealand
Audio files: 42
G2 patch files: 626
|
|
|
Back to top
|
|
|
ian-s
Joined: Apr 01, 2004
Posts: 2672
Location: Auckland, New Zealand
Audio files: 42
G2 patch files: 626
|
| Posted: Tue Jan 23, 2007 1:05 pm Post subject:
|
|
|
One more thing I overlooked, the KBT needs to be switched off when op.RatioFixed is selected (they are not separate functions on the DX7). I have added the code to fix this and the LRDepthMode swap, missing parameters. Here is a zip with all 4 factory roms, talk about nostalgia. I don't remember there being so many duplicates though.
Should I post the updated dx2g2.py ?
Last edited by ian-s on Wed Jan 24, 2007 10:53 pm; edited 3 times in total |
|
|
Back to top
|
|
|
ian-s
Joined: Apr 01, 2004
Posts: 2672
Location: Auckland, New Zealand
Audio files: 42
G2 patch files: 626
|
| Posted: Tue Jan 23, 2007 1:13 pm Post subject:
|
|
|
| I think the pitch EG is possible, just need to work out rate tables that can be scaled by the difference between levels. Might be able to use the operator envelope as a reference. |
|
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004
Posts: 1184
Location: Berlin
Audio files: 13
G2 patch files: 141
|
|
|
Back to top
|
|
|
ian-s
Joined: Apr 01, 2004
Posts: 2672
Location: Auckland, New Zealand
Audio files: 42
G2 patch files: 626
|
| Posted: Tue Jan 23, 2007 7:24 pm Post subject:
|
|
|
| 3phase wrote: | | to really catch the DX7 a big moddel is needed |
Perhaps, but I don't think many DX users would actually miss the awful sound of FM at 14bit, 50K. Clavia's model I think manages to increase the quality, in a good way.
Part of the reason for the death of the DX was that people began to hear past the novelty of the realistic sounds to the underlying digital nastiness.
Still the beauty of q's design is that if you really want lofi in the final effects, you can add it yourself to the DX7.PCH2 template and it will be there for all the patches you convert.
The LFO/Pitch EG section of the DX7 was a pretty poor design compared to the rest of the synth, which probably accounts for how rarely it is used in real patches. I think the entire LFO section can be dropped without much harm, just use the G2's built in performance LFO.
The pitch EG might be used in 1:1000 patches and from memory, was difficult to get anything useful out of it. It could be approximated with an EnvMult in bipolar L3 mode. The rate conversion would need to be proportional to the level Delta, so a rate value of 69 would be 20ms if the difference between L1 and L2 was 100, or 4ms if the difference was only 20. That is assuming the relationship is linear which it isn't for the level EG  |
|
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004
Posts: 1184
Location: Berlin
Audio files: 13
G2 patch files: 141
|
| Posted: Tue Jan 23, 2007 10:58 pm Post subject:
|
|
|
| g2ian wrote: |
The LFO/Pitch EG section of the DX7 was a pretty poor design compared to the rest of the synth, which probably accounts for how rarely it is used in real patches. I think the entire LFO section can be dropped without much harm, just use the G2's built in performance LFO. |
I dont agree...Its rarely used in all this public domain trash patches but in all the better sounds the lfo is important..I also dont see how you want to replace the lfo with the G2 performance Lfo...
important feature of the dx 7 is that you can adress the lfo individually to the operators and control the modulation amount by after touch or breath controler... this can be extremly important to bring DX sounds alive
| Quote: | | The pitch EG might be used in 1:1000 patches and from memory, was difficult to get anything useful out of it. It could be approximated with an EnvMult in bipolar L3 mode. The rate conversion would need to be proportional to the level Delta, so a rate value of 69 would be 20ms if the difference between L1 and L2 was 100, or 4ms if the difference was only 20. That is assuming the relationship is linear which it isn't for the level |
I agree that the pitch envelope is a bit funny..but its actually the part that is the easyest to reproduce...
i had the strong impression that the G2 multienv in bipolar linera just behaves like the dx one...i was a bit irritated about the minimal attack time from around 300 ms... and checked withe a less deep sweep... but..
seems to be wright... or am i missleaded? |
|
|
Back to top
|
|
|
ian-s
Joined: Apr 01, 2004
Posts: 2672
Location: Auckland, New Zealand
Audio files: 42
G2 patch files: 626
|
| Posted: Wed Jan 24, 2007 12:23 am Post subject:
|
|
|
| 3phase wrote: | I dont agree...Its rarely used in all this public domain trash patches but in all the better sounds the lfo is important..I also dont see how you want to replace the lfo with the G2 performance Lfo...
important feature of the dx 7 is that you can adress the lfo individually to the operators and control the modulation amount by after touch or breath controler... this can be extremly important to bring DX sounds alive
|
Sounds interesting, I would love to hear some of your non trash sounds.
| 3phase wrote: | i had the strong impression that the G2 multienv in bipolar linera just behaves like the dx one...i was a bit irritated about the minimal attack time from around 300 ms... and checked withe a less deep sweep... but..
seems to be wright... or am i missleaded? |
It has been more than 20 years since I last programmed a DX7 so I could be wrong. The pitch EG may behave differently to the Operator EG.
A simple test would be to set the Pitch EG levels to 50, -50, -50, 0 and the rates to 50,50,0,0, on the DX7 with a held note, this should produce an upward rise in pitch followed by a downward fall. If the time taken for both rise and fall is the same, then the EnvMult will do the trick. |
|
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004
Posts: 1184
Location: Berlin
Audio files: 13
G2 patch files: 141
|
|
|
Back to top
|
|
|
ian-s
Joined: Apr 01, 2004
Posts: 2672
Location: Auckland, New Zealand
Audio files: 42
G2 patch files: 626
|
|
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004
Posts: 1184
Location: Berlin
Audio files: 13
G2 patch files: 141
|
| Posted: Wed Jan 24, 2007 8:34 am Post subject:
|
|
|
thanx..but the problem wasnt the steps  ...
its the way the DX 7 is adressing keys that get into the glisando...
probably you are wright that this is allocation mode related...
maybe this just cant be done the same way on the g2..or with a high module effort...
However ..DX1.35 gives all basic parameters an is tuneable by the converter ...
I however thinkl that there should be a version with a propper DX modulation matrix aswell..or at least the mod wheel because there are many patches that reactt on the modwheel with a total soundchange...
my version 1.37 is buggy and to big... but i just played an hour on the DX7..to bad that i forgit how to safe on the machine... i had quite some fun with applying the mods to operators...
its funny..after 20 years computer trouble programming the DX from the front dont hurts as much as it used to hurt in the past... |
|
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004
Posts: 1184
Location: Berlin
Audio files: 13
G2 patch files: 141
|
|
|
Back to top
|
|
|
ian-s
Joined: Apr 01, 2004
Posts: 2672
Location: Auckland, New Zealand
Audio files: 42
G2 patch files: 626
|
| Posted: Wed Jan 24, 2007 1:03 pm Post subject:
|
|
|
That looks much better, and you could get 7 voice polyphony on an unexpanded G2.
I refreshed my memory with the original manual pdf that Jan posted. You are absolutely right about the expressivity of the modulation routing and that many patches take advantage of the individual AMD sensitivity for each operator, I think routing the LFO to AMD is pretty unusual. Most players, when being expressive, would prefer to control the modulation with there breath/hand/foot directly rather than relying on a mechanical sounding LFO wave. What I’m going to do then, is just add 1 constant level module which is morphed to my controller of choice patched to the AMod input.
This gives expression at very little cost and has the advantage that individual variation/patches can use different controllers if required, unlike the DX7 where it is global. I updated the DX7 patch I posted in the G2 patches/synth section like this. The choir and to a much lesser extent, the string pad uses it.
Because the DX7 LFO is essentially mono, could you perhaps put all your LFO and routing stuff (not the pitch EG) into the FX area and feed it back to the VA via a bus? |
|
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004
Posts: 1184
Location: Berlin
Audio files: 13
G2 patch files: 141
|
| Posted: Wed Jan 24, 2007 2:32 pm Post subject:
|
|
|
| g2ian wrote: | | Because the DX7 LFO is essentially mono, could you perhaps put all your LFO and routing stuff (not the pitch EG) into the FX area and feed it back to the VA via a bus? |
Too bad that the fx busses are not bi directional... and i dont want to use a main bus
I think its better to leave that to the user..you either can remove all the lfo stuff or place it elsewhere..
The G2 can do much better modulations anyway..But its important taht any parameter in the sysexfile find an adress in the G2...
The lfo is only part of the am mod circuit to the operators..The EG Bias parameters are the main part..you can decide if they carry lf mod or not but they act as a master after the lfo..
in the patch there are 2 EG bias assigned to modwheel and aftertouch..but you can assign anything there... the crossfader below needs to be fully clockwise..when than the lfo mods are switched off you have diredt AM modulation of the operators..
In the dx 7 you can switch and adjust that independently but its a good compromise to have the G2 patch more simple like this... The user can evaluate if the settings of the patch are promissing for operator am mod and can additional circuit manmually if desired |
|
|
Back to top
|
|
|
ian-s
Joined: Apr 01, 2004
Posts: 2672
Location: Auckland, New Zealand
Audio files: 42
G2 patch files: 626
|
|
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004
Posts: 1184
Location: Berlin
Audio files: 13
G2 patch files: 141
|
|
|
Back to top
|
|
|
ian-s
Joined: Apr 01, 2004
Posts: 2672
Location: Auckland, New Zealand
Audio files: 42
G2 patch files: 626
|
| Posted: Thu Jan 25, 2007 1:30 am Post subject:
|
|
|
| 3phase wrote: | | the minimal attack time is aroundd 300ms |
What L1 and L2 settings did you use to get that time?
On the operator EG there is a non linear relationship between rate and level so that L1=50,R1=65,L2=99 takes longer than L1=0,R1=65,L2=49.
I presume this is because of the expo curve and was hoping the pitch EG would have a linear relationship to match the linear curve.
So for example if R1=99 and abs(L1-L2)=50 takes 150ms, then abs(L1-L2)=25 should be 75ms. This would mean dx2g2 would just need to calculate the ms based on a table[0:99] * abs(ln-Ln+1) and do a closest fit to the G2 EnvMult rate. |
|
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004
Posts: 1184
Location: Berlin
Audio files: 13
G2 patch files: 141
|
| Posted: Thu Jan 25, 2007 7:59 am Post subject:
|
|
|
I only can say that the pitch eg i ve in my DX version patch sounds pretty much like the one in the real DX..so i guess its linear...
and the fastes setting wich is 99 translates to 322 ms on the G2..
The levelmatching is done with the extra mixer module next to the pitch EG..0=0 and 99=127 on the G2 while 50=64...so you have the range and its centered...
You however should note that L4 is setting the startpoint of the Pitch EG aswell...so it can have an imideate jump to any level..just like in the DX |
|
|
Back to top
|
|
|
ian-s
Joined: Apr 01, 2004
Posts: 2672
Location: Auckland, New Zealand
Audio files: 42
G2 patch files: 626
|
| Posted: Thu Jan 25, 2007 9:37 am Post subject:
|
|
|
I'm not sure how to explain this better. The G2 EnvMulti uses T1,L1,T2,L2 etc as apposed to the DX which uses R1,L1,R2,L2...
This is not simply a different choice of letter.
In the G2, T=Time. This specifies the number of msec for each segment.
In the DX, R=Rate. This specifies the rate of change or slope of the segment, the actual time in ms depends as much on the distance travelled.
You can match the times of T1/R1, but the G2 setting will only be valid for the specific L1 to L2 setting.
This is confirmed by the factory patch conversion I did, patches like Decent and Flexitone have big variations in frequency and the times sound about right (from memory) but the Voices patch which has a very small frequency deviation sounds obviously too long.
What is needed is a table of times measured for a level deviation of 10 units (ie L1=50,L2=60) that would allow dx2g2 to do a true ms calculation based on both the table and the difference between consecutive levels.
Does that make sense?
On the bright side, thank god Yamaha didn't include KVS and KRS on the Pitch EG, that would have made it impossible. |
|
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004
Posts: 1184
Location: Berlin
Audio files: 13
G2 patch files: 141
|
| Posted: Thu Jan 25, 2007 11:53 am Post subject:
|
|
|
I just checked it..you are wright..but how can we scale it roughly to get a good match without making a monster table..
my settings are fore the extrems..so +50 (+64) than the minimal time is 322 and max time is not long enough
when being on just a minimal setting like 8 units up on the dx 7 (+11 on G2) we have minimal times around 14 ms... while max is around 6,5 sec
would this be enough have a rough frame of settings scaled to avoid drastical missmatchings? |
|
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004
Posts: 1184
Location: Berlin
Audio files: 13
G2 patch files: 141
|
| Posted: Thu Jan 25, 2007 12:20 pm Post subject:
|
|
|
other possebility would be to use a dx operator as pitch eg..but thats pricy..but you could edit than the same way as you would do on a dx 7..
theoretical... the operator is actually quite cheap ..seems that more than one module in a patch gets optimized somehow... but...
he enevelope is unipoar exp not bipolar lin...
So an evelope followe alone doesnt to the job.. :-/
checked it..dont works.. |
|
|
Back to top
|
|
|
ian-s
Joined: Apr 01, 2004
Posts: 2672
Location: Auckland, New Zealand
Audio files: 42
G2 patch files: 626
|
| Posted: Thu Jan 25, 2007 1:05 pm Post subject:
|
|
|
The EnvMult will do, we don't need a monster table, just two tables and a bit of code.
First the program uses the DX rate value to look up a table as measured for a fixed interval (say 10 DX level units). It then works out a factor to multiply this value by based on how many actual units are between the two levels. Finally, it finds the nearest match in a table of G2 envelope times and uses that as the G2 T setting.
If the T is greater than 45 seconds it just gets set to 45 seconds, not that critical.
This would do fine till those nice folks at Clavia releases V1.5, the one with the fixed filter, DX pitch EG, sine bank etc.  |
|
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004
Posts: 1184
Location: Berlin
Audio files: 13
G2 patch files: 141
|
| Posted: Thu Jan 25, 2007 2:58 pm Post subject:
|
|
|
| g2ian wrote: | | This would do fine till those nice folks at Clavia releases V1.5, the one with the fixed filter, DX pitch EG, sine bank etc. |
Dx pitch eg? a constant rate envelope is really missing... thats hard to patch..
The dx operator with an envelope foloower is somehow a trich to schieve such things..but only with ex curves and not bipolar...
I am just sitting on the PEG table..its not so easy because yamaha choose to scale booth time parameter and and level setting exponential..the enelope it self appears to be linear..but its stepped..it dont plays smooth to the end..
So in any case we only get a rough equation... because the G2 one is smoother its probably not too bad |
|
|
Back to top
|
|
|
dasz
Joined: Oct 16, 2004
Posts: 1644
Location: victoria, canada
Audio files: 29
G2 patch files: 56
|
| Posted: Thu Jan 25, 2007 3:49 pm Post subject:
|
|
|
qfingers,
you and 3phase (and all other q2tools contributors) are A W E S O M E 
I just converted my entire Papa Nord library and about 85% of the patches were dead on. Whoa, now I do not miss my Papa Nord at all.
One question, on some patches (most of them), the osc pitch is converted to Hz, could one instead ensure the converted pitch is in semi's?
/Dasz |
|
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004
Posts: 1184
Location: Berlin
Audio files: 13
G2 patch files: 141
|
| Posted: Thu Jan 25, 2007 4:39 pm Post subject:
|
|
|
could you please show a patch that dont converted well?
last chance to find a bug before 1.3... |
|
|
Back to top
|
|
|
|
Forum index » Clavia Nord Modular » g2ools utilities
FAQ
Calendar
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
Chat Room
