You are here

MIDI Channel 10 is bad for drums - fact or fiction?

For fans of synths, pianos or keyboard instruments of any sort.

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby Wonks » Wed Oct 23, 2019 11:06 am

Once you use a standard DIN hardware MIDI connection, then you are constrained to the 31250 bits per second of the MIDI standard. So if you have several notes all being sent at the start of a bar on other channels sending messages out on lower channels before channel 10 will delay the sending of the drum note message until those other lower channel note details have been sent.

'Running Status' MIDI messages allow multiple note on/off messages to be sent in sequence without having to send the MIDI channel number at the start to reduce the number of message bytes sent (a MIDI message byte is 10 bits long as it has a start and stop bit as well as 8 bits of data). So using this 'running status' form, a six note chord would take up 13 bytes, instead of 18 bytes if each note was sent as a separate entity. 13 message bytes is 130 bits, so that message would take 130/31250 = 4.1 ms to send. Add in a few more messages like that on other channels below channel 10, and you can see that you can easily get noticeable delays of 20+ms. What's more, with less messages sent on other channels, that delay time from the beginning of the bar (or 1/4 or 1/8 note position etc.) can change each time so it may be 5ms or it may be 30ms. Very noticeable on sharp transient drum sounds, less so on soft attack pads.

But this all depended on the sequencer used. Some used the channels in order from 1 to 16, others prioritised channel 10, then 1-9 and 11-16.

But within software to software MIDI transfers, the software isn't limited to physical serial transmission speed limitations , so it gets transferred within microseconds worst case.
User avatar
Wonks
Jedi Poster
Posts: 10132
Joined: Thu May 29, 2003 12:00 am
Location: Reading, UK
Correcting mistakes on the internet since 1853

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby BillB » Wed Oct 23, 2019 11:33 am

Wonks wrote:But this all depended on the sequencer used. Some used the channels in order from 1 to 16, others prioritised channel 10, then 1-9 and 11-16.

Thanks, Wonks, that is interesting and pretty scientific! Especially that some devices may prioritise channels. I guess an easy test would be to set a note on exactly the first beat on every channel, run it into MIDI-OX and see if they go 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16 or 10,1,2,3,4,5,6,7,8,9,11,12,13,14,15,16...

It may not matter if all the drum data goes via a dedicated port, but it would be good to know if it makes any difference in a mixed channel data stream...
BillB
Frequent Poster
Posts: 756
Joined: Mon Apr 07, 2003 12:00 am
Location: East Yorkshire

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby Mixedup » Wed Oct 23, 2019 12:47 pm

Hmmm... interesting. Back in the arse-end of the '80s, when I started making MIDI-based music on an ancient IBM clone, my sequencer's (Prism) manual was very clear that it assigned a higher timing priority to MIDI channel 10 than any other channel, and recommended using it for drums. I can't say I noticed any timing issues working that way, but I was a young whipper-snapper back then and probably knew no different...

(I acquired an electric guitar in 1990 and promptly eschewed all things MIDI for a good decade or so :beamup: )

Did they have it wrong, or did different software treat things differently back then?
User avatar
Mixedup
Jedi Poster
Posts: 4388
Joined: Wed Sep 03, 2003 12:00 am
Location: Cambridgeshire, UK

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby desmond » Wed Oct 23, 2019 12:58 pm

Mixedup wrote:Hmmm... interesting. Back in the arse-end of the '80s, when I started making MIDI-based music on an ancient IBM clone, my sequencer's (Prism) manual was very clear that it assigned a higher timing priority to MIDI channel 10 than any other channel, and recommended using it for drums.

There was a time *before* GM, you know! :tongue:
(Ie, the MIDI sequencing years 1983-1991 - yes, GM standard came about in 1991)

The whole "MIDI Channel 10 for drums" thing largely came about because of the GM standard. After that time, some sequencers may perhaps have given priority to channel 10 events. Before that, however, in general higher tracks had timing priority, as tracks were usually processed from top to bottom.

Basically, there is no "this channel is best" because it's *all* to do with the sequencer's implementation. As long as you have plenty of bandwidth and aren't trying to put too many events on the same rigid clock position, things are generally fine. The moment things get cluttered is where some event jostling takes place, and the exact behaviour will depend on the events, and how the sequencer works.

For example, some sequencers may go in a loop and work from channel 1 up when outputting events. Some sequencers may place recorded events in a list/queue, and output the events in queue order, rather than channel order. Some sequencers may be designed to say "let's handle MIDI channel 10 events first, as we give those timing priority". It all depends on the implementation.

Testing can be fine (if you want to avoid doing anything productive, and yet still *feel* productive :lol: ) but really, make some sensible decisions based on your use cases (eg, lets put critical timing stuff on port 1, and the washy stuff on port 2), or, if you've got a *lot* of timing critical stuff, spread it over two ports - and make music.

If you notice a problem, then investigate *then*, rather than necessarily try to fix a problem you don't yet have.
User avatar
desmond
Jedi Poster
Posts: 8989
Joined: Tue Jan 10, 2006 1:00 am
mu:zines | music magazine archive | difficultAudio

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby Ben Asaro » Wed Oct 23, 2019 1:03 pm

I often have multiple channels going on simultaneously between my TX7, JV-1080, and DR-550 which is on channel 10, all from an MC-500. I’ve not noticed any lag beyond the normal start/stop lag on the MC-500. I mention this because all of this kit is mid-80s or mid-90s vintage, and I’ve not had any call for complaint.
Ben Asaro
Frequent Poster
Posts: 573
Joined: Thu Feb 26, 2009 1:00 am
Location: NYC

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby nathanscribe » Wed Oct 23, 2019 1:06 pm

You could avoid all this MIDI terror completely by programming all drums on your external hardware and syncing over FSK instead. :tongue:
User avatar
nathanscribe
Frequent Poster
Posts: 1078
Joined: Fri Jan 19, 2007 1:00 am
Location: Wakefield, for my sins.
I have no idea what I'm doing.

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby BillB » Wed Oct 23, 2019 2:53 pm

nathanscribe wrote:...MIDI terror...

Pah, MIDI is my friend. I don't want to learn the TR, EMX, and Volca ways of sequencing and saving, I just want to get to know one rhythm programming system to fire off the other devices. That was going to be the Beatstep Pro before I got the MPC, but the MPC makes more sense as the programmer, as it has 16 pads x 4 banks. So any of 64 notes (from 35 to 98) can be assigned to the pads in a sequence.

Perfect as the central MIDI rhythm sequencer, as it can address any hardware drum source I have - excepting that General MIDI has eight drum sounds set below note 35, which I will just have to live without!
[there is even a way around that, but it is too convoluted :headbang: ]
BillB
Frequent Poster
Posts: 756
Joined: Mon Apr 07, 2003 12:00 am
Location: East Yorkshire

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby BillB » Wed Oct 23, 2019 3:00 pm

desmond wrote:If you notice a problem, then investigate *then*, rather than necessarily try to fix a problem you don't yet have.

You're right Desmond, but as I wanted to put some effort into setting up useable templates to make future efforts more efficient, I thought I would ask the SOS sages :D

Sometimes, just the discussion process points out a way - like assigning rhythm parts to one port only. However, as I will also sending multi-channel info on the other port, it is still good to understand the potential issues or non-issues.

As always, I'm very grateful for everyone's inputs. :thumbup:
BillB
Frequent Poster
Posts: 756
Joined: Mon Apr 07, 2003 12:00 am
Location: East Yorkshire

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby The Elf » Wed Oct 23, 2019 5:06 pm

If you're using a multi-port interface and assigning your drum module to its own output (I also can't be bothered programming anything pattern-based outside my DAW...) you avoid the MIDI channel priority problem - then you just have to worry about whether lower numbered MIDI ports are serviced first!
User avatar
The Elf
Jedi Poster
Posts: 13048
Joined: Tue Aug 14, 2001 12:00 am
Location: Sheffield, UK
An Eagle for an Emperor, A Kestrel for a Knave.

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby James Perrett » Wed Oct 23, 2019 5:07 pm

Folderol wrote:Probably more relevant is don't use MIDI THRU if you can possibly avoid it - especially not on kit that uses active sensing.

Surely this only applies to gear where the Thru isn't a true hardware Thru. The delay on a hardware thru should be tiny.
User avatar
James Perrett
Moderator
Posts: 8574
Joined: Mon Sep 10, 2001 12:00 am
Location: The wilds of Hampshire
JRP Music - Audio Mastering and Restoration. JRP Music Facebook Page

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby BillB » Wed Oct 23, 2019 5:11 pm

James Perrett wrote:
Folderol wrote:Probably more relevant is don't use MIDI THRU if you can possibly avoid it - especially not on kit that uses active sensing.

Surely this only applies to gear where the Thru isn't a true hardware Thru. The delay on a hardware thru should be tiny.

Yes, I get avoiding daisy-chains of MIDI Thru, but a single Thru-box (4- / 6-way / more), driven by a single output port, should be OK.
BillB
Frequent Poster
Posts: 756
Joined: Mon Apr 07, 2003 12:00 am
Location: East Yorkshire

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby Folderol » Wed Oct 23, 2019 6:45 pm

BillB wrote:
James Perrett wrote:
Folderol wrote:Probably more relevant is don't use MIDI THRU if you can possibly avoid it - especially not on kit that uses active sensing.

Surely this only applies to gear where the Thru isn't a true hardware Thru. The delay on a hardware thru should be tiny.

Yes, I get avoiding daisy-chains of MIDI Thru, but a single Thru-box (4- / 6-way / more), driven by a single output port, should be OK.
In a word, Congestion
P.S. Note I said especially with kit that has active sensing.
User avatar
Folderol
Jedi Poster
Posts: 9179
Joined: Sat Nov 15, 2008 1:00 am
Location: The Mudway Towns, UK
Yes. I am that Linux nut.
Onwards and... err... sideways!

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby BillB » Wed Oct 23, 2019 11:43 pm

BillB wrote: I guess an easy test would be to set a note on exactly the first beat on every channel, run it into MIDI-OX and see if they go 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16 or 10,1,2,3,4,5,6,7,8,9,11,12,13,14,15,16...

OK, I think this is sorted, at least for the MPC.

With identical notes on the first beat, copied to every channel, where track 1 = channel 1 etc, the order of channelised notes arriving at MIDI-OX is
1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16.

However, if Track 1 is set to channel 10, the order of channelised notes is
10,1,2,3,4,5,6 etc.

Same holds true whatever channel is set to whatever track. The notes are output in track order, irrespective of channel. So if I want Channel 10 to have top timing priority, I set it to track 1. If I am understanding MIDI-OX’s timestamp correctly, it seems to take about 17ms to send the full set of 16 note-on commands.

So the simple answer for me is a) stick with Ch10 for drum tracks and b) place them on the first sequencer track(s).

Job’s a good’n, thank you all for your help. :thumbup:
BillB
Frequent Poster
Posts: 756
Joined: Mon Apr 07, 2003 12:00 am
Location: East Yorkshire

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby desmond » Wed Oct 23, 2019 11:56 pm

Good findings. It's about what I'd expect on a hardware sequencer of this vintage - the code is probably fairly simple, and loops through the tracks from track 1 onwards outputting events as necessary.

Now you have a plan! :thumbup:
User avatar
desmond
Jedi Poster
Posts: 8989
Joined: Tue Jan 10, 2006 1:00 am
mu:zines | music magazine archive | difficultAudio

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby resistorman » Thu Oct 24, 2019 1:00 am

Unless you want the drums a bit behind the beat :D
User avatar
resistorman
Frequent Poster
Posts: 678
Joined: Sun Nov 22, 2015 1:00 am
Location: Asheville NC

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby Mixedup » Thu Oct 24, 2019 8:46 am

desmond wrote:
There was a time *before* GM, you know! :tongue:
(Ie, the MIDI sequencing years 1983-1991 - yes, GM standard came about in 1991)

The whole "MIDI Channel 10 for drums" thing largely came about because of the GM standard. After that time, some sequencers may perhaps have given priority to channel 10 events.

This *was* before GM!
User avatar
Mixedup
Jedi Poster
Posts: 4388
Joined: Wed Sep 03, 2003 12:00 am
Location: Cambridgeshire, UK

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby desmond » Thu Oct 24, 2019 9:10 am

Did you have an MT32 by any chance? That was iirc the first thing that had a convention of channel 10 for drums, largely because it had 8 parts, but ignored channel 1, so it’s 8 parts were on channels 2-9, with the drums therefore landing on 10.

As the MT32 was popular in 1987, a fair few software sequencers directly supported it, and drums started to be expected on channel 10 to maintain compatibility. So that’s where the channel 10 thing really started, and continued in Roland modules/workstations, before becoming officially part of the GM spec.
User avatar
desmond
Jedi Poster
Posts: 8989
Joined: Tue Jan 10, 2006 1:00 am
mu:zines | music magazine archive | difficultAudio

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby Mixedup » Thu Oct 24, 2019 3:03 pm

No, not familiar with that one. I tended to use Yamaha stuff... much of it loaned from the music dept of the school my dad worked at!
User avatar
Mixedup
Jedi Poster
Posts: 4388
Joined: Wed Sep 03, 2003 12:00 am
Location: Cambridgeshire, UK

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby BigRedX » Thu Oct 24, 2019 4:39 pm

I've never used channel 10 for drums and always thought it a bit strange (for the reasons outlined in previous posts) that it was done that way.

In all my programming, as far as possible, drums have always been on MIDI channel 1 and when using multi-port interfaces on Channel 1 of port one. And they've always been on the first track(s) on my Logic arrange page. The only exception to this was if I was using a MIDI device that didn't normally do drum sounds for an additional percussion part in an arrangement.

I think the first MIDI device I came across which defaulted to MIDI channel 10 for drums was the Roland TR707.
User avatar
BigRedX
Frequent Poster
Posts: 737
Joined: Fri Sep 03, 2004 12:00 am
RockinRollin' VampireMan

Re: MIDI Channel 10 is bad for drums - fact or fiction?

Postby desmond » Thu Oct 24, 2019 4:52 pm

BigRedX wrote:I think the first MIDI device I came across which defaulted to MIDI channel 10 for drums was the Roland TR707.

Hmm, interesting. The MIDI implementation chart just says it can be set to 1-16 and can be changed to whatever you want and memorized, but it doesn't say in the manual (that I can see) it defaults to 10 when shipped.

If that's true, that's interesting, and suggests that Roland internally tending to think of MIDI channel 10 as for drums, even before the MT32. I wonder if their other e-drum stuff of that era, like the DDR30*, also defaulted to channel 10..?

*Edit: the manual for the DD3 regarding MIDI channels etc is basically the same as the TR707, no mention of the default either.
The TD7 defaults to channels 10-13.

I suspect Roland wanted to keep the drum channels sufficiently far away from the lower end so that people buying DX7's wouldn't hear drum sounds when playing notes etc, which would be confusing, and they settled on 10 as a sensible choice...
User avatar
desmond
Jedi Poster
Posts: 8989
Joined: Tue Jan 10, 2006 1:00 am
mu:zines | music magazine archive | difficultAudio

PreviousNext

Who is online

Users browsing this forum: Goseba