Error when updating your profile (08 Sep 2018)
When you try to update your profile, you get an error from the server. This is a know bug, which is caused by incompatibility between the website template software and the forum software. The developers are working on a fix. Hopefully the problem will be fixed before October. Until then please contact a moderator if you need to update your profile. He or she will help you.
Roland E-A7 BS!
Thanks in advance
How did you make the track? Is it a live 'capture' of a style? Which style?
Are you using USB or a MIDI interface?
Do you have Sonar set up to NOT 'echo' the MIDI in to the out? And do you have the E-A7 set to NOT allow MIDI 'through'? MIDI feedback loops can result in some pretty strange stuff!
One thing missing from the readout you supplied was any timing information. Hard to tell the data density without bar/beat/tick info...
Can I presume that the columns are for Ch/Type/Value/Data/ and lastly note length?
As a rule, controllers like this do not tend to disrupt timing. They are very short data packets and on the whole (I've never experienced it since the early 80's synths) don't disrupt timing. But I'll tell you what often DOES, and it's very interesting that your readout doesn't include any (is there a display filter on?)...
Generally, at the start of any style, there's a pretty big sysex packet that gets sent out on 1.1.000. It's the header information, any Makeup Tools data that is in the style. This may be edits you have made, or any offsets from the original programming. Depending on your MIDI interface, connection type and gear (some keyboards respond to sysex faster than others) you may often hear some timing stutters.
I notice that your event list edit display has none... So, is this readout from the middle of a track, or the beginning? Where is the sysex? Is Sonar hiding it from you?
This looks like the readout of a bassline (Ch2. is usually reserved for that). So this doesn't look like a particularly dense data stream (basslines tend to be one note at a time!). What is interesting is all those identical pan/expression/effects/chorus packets, but without timing info it's hard to tell how often these repeat. Are they happening every four bars, every chord change, every time you change Variations?
On the whole, you can tell that they are all identical (with the exception of the CC5 value=0 portamento and the CC84 data, which has to do with 'jumping' to the correct note when a new chord is input fractionally late), so, on the whole, it shouldn't be that difficult to erase all but the first packet of data. I'm not familiar with Sonar, but many DAW's allow you to filter the display by specific CC#'s, so just have it display CC11, CC10, etc. one at a time, and erase all but the original packet. Just be vigilant about if there is a change in value (you don't want to miss those!). That ought to thin out the data a bit (although from just this one track, it is hard to imagine this little data interrupting the timing). Have you tried seeing if the timing gets iffy when you just solo this track?
Sysex, OTOH, yes, that can DEFINITELY mess with timing. The nature of sysex means that the entire packet needs to be sent in one go, and if there's a bunch of notes all supposed to come out on 1.1.000 and a big sysex header, this can often mean the notes stutter a bit.
There's a fairly simple fix, though, as long as the sysex is only at the beginning... Drag your entire song ahead one bar, then select ALL the sysex and cut and paste it to 1.1.000. That way, the sysex plays BEFORE any notes start, and that's that! No more timing issues. I often find SMF's have this issue too, if there's sysex coming out on 1.1.000 and a count-in supposed to start on 1.1.000 as well. Usually, the first click comes in a bit late and messes with the timing.
BTW, here's a little tip for getting super accurate 'captures' of styles when using external sequencers. SLOW DOWN! The slower you set the tempo, the more time the interface or USB bus has to receive it (although MIDI clock is variable, the data transmission rate in baud of MIDI stays the same). Try dropping the tempo 30% or more while you record the arranger's style Parts (probably best to not play your part at the same time, do it as an overdub after you get a good 'capture') and you will often get a tighter timing 'capture'. Then simply speed the DAW back up to project tempo afterwards!
Not only will the 'capture' be tighter, but because of the slow speed, your playing and timing accuracy will be tighter playing the chords and pressing the Division buttons, which leads to less of those fractionally late Portamento=0 CC84 issues.
But anyway, back to your issue... Try this:
Set your DAW to NOT transmit sysex. Then play a captured style sequence that previously stuttered a bit. Does it now play smoothly? If so, there's your answer. If not, first thing I would do is make sure you aren't accidentally looping your MIDI ins to outs and vice versa.
Give this a try and get back to us....
I made the track by playing live into Sonar as a MIDI project Via MIDI input (Emu). This is the exact setup I have been using for 22 years ... only the E-A7 has replaced the G-800. I have disconnected all other instruments and sound modules, so it is just E-A7 into Sonar for simplicity.
All Echo is off
MBT= (Example) 12:04:029
"Can I presume that the columns are for Ch/Type/Value/Data/ and lastly note length?[/color" Yes, that is correct
It is exactly the grouping of these controllers on all tracks that cause the stumble. That is precisely where the song pauses.
I have disabled all Sysex commands ... both when Sonar starts and on each track of a project with no difference.
"What is interesting is all those identical pan/expression/effects/chorus packets," Yup, that is happening on all tracks at the same time and that is what is causing the stumble, in my opinion.
"Are they happening every four bars, every chord change, every time you change Variations?" Depending on the song, yes to all on one degree or another.
"Set your DAW to NOT transmit sysex" Yup, done that to no effect.
I called Roland headquarters in LA (raised a little hell) and they had their tech support manager call me. I sent him several screen shots and we are working on the problem. Fortunately, he has both the E-A7 & Sonar. He tried exactly what I did and didn't get the problem. He is contacting Japan. I'll wait and see. All 210 songs done on the old G-800 work perfectly, except for a very few incorrect instruments on a few songs, which is so easy to fix.
Thank you so much for all the time you put into answering this and I will keep you posted in case anyone else has this situation. It just might be a defective unit.
Perhaps there's another (fairly) simple step you can take...
Rather than hooking up to your computer through a 20 year old MIDI interface (you didn't mention whether you use an up to date computer and OS with your vintage MIDI interface) which may have out of date drivers, could you try using a USB cable to connect the E-A7 with Sonar? Perhaps that might cure it (I bet the Roland guy is using one, or at least a different MIDI interface with perhaps more up to date drivers!).
What is weird is, the first post does NOT have the controller package after every note, but the new one does. I wonder what is different between the tracks?
Just to be sure, make sure 'Soft through' (p.65 in the E-A7 Reference manual PDF) isn't enabled.
I know it's probably obvious (did the Roland guy make you go through these steps?) but have you tried backing up your User Data, then doing a Factory Reset and trying to record a style before you reload your user data?
One last ditch idea... Try a different DAW. Reaper is very good, free to try, and may be able to confirm or disprove that it is the DAW that is the issue. Reaper with USB connection definitely gets rid of two of the variables, and puts the spotlight directly on the E-A7.
Me, I'd put my money on the interface as the likeliest culprit, given that the Roland guy wasn't getting the issue (and it hasn't been reported here yet), and some kind of MIDI echo mismatch. It is hard to think of an OS problem (in the E-A7) that would result in this if no-one else is getting it.
In the meantime, it strikes me as easy that, if you get the display to show ONLY those controllers that are constantly repeating, select from the second one of them onwards and erase. After you have done all this on all channels, does the timing get better, or can you see that the notes themselves are moved to inconsistent ticks?
Expecting an old MIDI interface to keep up with much higher data density of modern instruments, particularly when USB is so easy and convenient is a tall order. I have always preferred arrangers with sequencers in them to do my capture (the BK9 was the only BK with a proper sequencer on board) before I transfer to a DAW. I tried many comparisons between using a MIDI interface and the onboard sequencer, and ALWAYS got better timing accuracy with the on-board. I wish more manufacturers would acknowledge this...
I'm afraid comparing to the old G800 (I had one, quickly replaced by a G1000 I used for over ten years) isn't really relevant. So much has changed, so much more is being handled by CC and sysex codes that weren't in the G800, it isn't really fair to compare..!
BTW, once the issue is fixed, DON'T disable the sysex any more. It includes all kinds of really important stuff that is audible if missing (EQ's. MFX settings, Makeup Tools drumkit edits, etc.). But if you get a timing hiccup on 1.1.000, do my trick of moving the sequence a bar later, and then move the sysex ONLY to 1.1.000. That will cure it..!
Strictly speaking, if you are playing a style into the DAW to record, the light for MIDI activity from the E-A7 to the computer should light, and the light for activity from the computer to the E-A7 should NOT. And vice versa if playing from the computer to the E-A7.
That should give you a definite visual confirmation that all is working as intended.
"Last tip... Does your MIDI interface have activity lights on it for MIDI in and out?" Yes, both work as before
Yes, I tried the MIDI USB with the same results
"What is weird is, the first post does NOT have the controller package after every note, but the new one does. I wonder what is different between the tracks?" Actually, they are from two different songs. I can't explain the difference ... like you stated, It is just weird.
"Just to be sure, make sure 'Soft through' (p.65 in the E-A7 Reference manual PDF) isn't enabled." Soft Thru is off when recording and on when playing back, because the data for the other instruments needs to get thu. I have MIDI Sets that do this automatically and I have verified that they still work.
"have you tried backing up your User Data, then doing a Factory Reset and trying to record a style before you reload your user data?" Yes, same result
"Me, I'd put my money on the interface as the likeliest culprit, given that the Roland guy wasn't getting the issue (and it hasn't been reported here yet), and some kind of MIDI echo mismatch. It is hard to think of an OS problem (in the E-A7) that would result in this if no-one else is getting it." I don't think it's the interface, nor, Sonar are creating all those extra events. The ol' crap in ... crap out deal. Besides, wouldn't an echo problem echo the notes as well? Again, after all you, I and the Roland guy (waiting for him to get back) have done it just might be a defective unit.
DiKi, thanks again for all the effort you have given me. I'll keep you posted.
I guess the last step to try and track this issue down is, have you got a friend with a sequencer? Any chance of taking the unit over to his/her place and do a quick test recording there? That would remove the computer and interface completely from the equation and prove beyond a doubt that the issue resides in the arranger.
BTW, have you re-installed the latest OS? I know previous OS updates for other Roland's (the BK series was one of them) had OS update procedures that often left them 'hung' for a considerable time, and if you interrupted the process, sometimes things went south. Do you recall the last time you updated the OS if anything unusual happened? Did you have the E-A7 on a UPS while you updated (that's a REALLY good idea if you have one, to prevent any voltage swings from interrupting the process)?
No, unfortunately, I live on a mountain top in the Mojave Desert. One person came over and by the time he left, my system was all screwed up.
The unit came with the latest OS version. They suggested that if you have the latest, do not reinstall it.
I made a workaround and just tried it. First, I recorded the song. Then selected all the tracks, except the Hammond one, Then I used "search by filter", selected controller #'s 11 & 91-93. Once selected, I hit delete and that solved my problem A bit awkward, but. I am now running until I hear from Roland. It is definitely all those controllers on all tracks.
Also, I would not remove those portamento and CC84's that are only occasional (they are there whenever an inputted chord is a tiny fraction late, to 'jump' the incorrect note to the correct one without retriggering the voice).
"Also, I would not remove those portamento and CC84's that are only occasional (they are there whenever an inputted chord is a tiny fraction late, to 'jump' the incorrect note to the correct one without retriggering the voice)" I did not know that ... thanks for the info.
Thanks, again, for all your time. If you lived close to me, I'd buy you dinner :~)
If I ever find myself on a mountain top in the Mojave, I'll be sure to drop by!
"If I ever find myself on a mountain top in the Mojave, I'll be sure to drop by!" By all means. I'm on the outskirts of Palm Springs, CA and there are plenty of restaurants there. You have been very helpful. Take care.