Tag: Jamulus

This month’s Jamulus singing was the best yet. We did not have the tempo problems that have plagued us in previous Jamulus singings. We were more in tune than we have been. We got through some passages — fermatas, time signature changes — that used to feel difficult when we couldn’t see the leader.

And at the end of the singing, we all commented on how good we sounded, and how together we were. Several of us wondered aloud why this was so much better. Susan pointed out that we’re listening to each other more. I commented that singing on Jamulus requires a somewhat different set of skills than singing in person. Mark said he felt he could just sing without fiddling so much with the controls. Whatever the reason, or reasons, the general consensus was that we’ve finally gotten good at singing on Jamulus.

We’re still all looking forward to singing in person once again, obviously. Nevertheless, Jamulus singing is now quite pleasurable.

A somewhat smaller group this time, with some excellent singers. We had a rocky first half hour, then we seemed to get into a groove.

We were doing well enough that I decided to take a chance and when it was my turn started us off on Worcester by Abraham Wood, one of my favorite First New England School tunes. It went surprisingly well. It probably helped that I was leading the tune as well as singing bass, because the bass line really drives the whole piece so I could set the tempo just by singing (it also helped that Jerry, an excellent musician with rock-solid time, was also singing bass). It also helped that I chose a moderate tempo, about 112 b.p.m.

Jerry then led Billings’ Easter Anthem. Here again, the bass line can drive the whole piece. He started us at about 120 b.p.m., and when all four parts came in the tempo would gradually slow to maybe 112, but then in the duets or bass solos, Jerry could pick up the tempo again. Interestingly, the recitatif at measure mm. 77-81 seemed to me to stay at tempo without slowing noticeably.

There were lots of other successes tonight as well — those two just happen to stand out for me. It was really the best music making I’ve yet done online.

Bay Area Jamulus singings from the Denson book were initially scheduled for the first and third Wednesdays. The BASH Board decided we’d sing from the Shenandoah Harmony on fifth Wednesdays.

Jerry came prepared with a list of tunes he wanted to sing — most of which were by First New England School composers like Abraham Wood , Jeremiah Ingalls, and Billings. I love First New England School music, so I was a happy camper. (Also, while the Shenandoah Harmony, like any collection, contains a fair number of duds, I feel that the editors did their best work with the 18th century music.)

Most of us were seeing all this music for the first time, and sight-singing together on Jamulus was a bit challenging. We finally got into a pattern that worked: whoever was leading the tune would count in two full measures at their preferred tempo. It was a forgiving group, and everyone was ready to start over if started to sound like a train wreck. Singing on Jamulus forces you to really listen to the other parts — you can’t fall back on watching someone keeping tempo, or pointing to you to bring in your part. That there were only seven singers probably helped with our sight-singing; Jamulus can sound a little chaotic until you adjust the volume of each individual singer, and with only seven singers adjusting volume was easy.

I feel that singing on Jamulus revealed something of the “singability” of the compositions. First New England School music can be challenging, but it is eminently singable: the music fits comfortably to singers’ voices; Johann Fux would have approved. Some of the recent compositions, while good as music, feel less singable to me; but this isn’t a hard and fast rule, for when we sang Neely Bruce’s 1990 tune Millbrook, I felt the bass part (at least) was quite singable.

For the last half hour, we switched back to the Denson book, and sang some familiar tunes. It was tiring to sight-sing for an hour and a half, and it was good to end with music we all knew well.

All in all, this was a good singing. There were the usual technical problems: Mark got kicked off the session and couldn’t rejoin; Carla tried to log in and wasn’t able to. But overall, a good two hours of music-making.

My work schedule finally allowed me to join the regular twice-monthly Bay Area Sacred Harp Jamulus singing.

There were about 15 log-ins, and three of those log-ins had more than one person singing. So we may have had twenty total humans singing together. We had one singer log in from Kansas, and his total latency as reported in Jamulus was about 74 ms; I did not notice that he slowed us down. Another singer logged in from San Diego, and I’m not sure what his latency was. But most of us had latencies in the range of 30-50 ms. This tended to keep tempos slower than usual.

Singing using Jamulus is getting easier. I’ve gotten used to the lack of visual cues; it’s still annoying, but it’s no longer disconcerting. I’m better able to judge how to stay on tempo: it’s a fine line to walk because on the one hand you need to have rock-solid time and stick with that time no matter what, but you also have to listen carefully in case the overall group is slowing down, in which case you have to adjust your internal metronome.

However, I still get tripped up by things. For example, this time I thought I was watching my volume level on the Jamulus controls, but at one point when I had to look down at the music, I unconsciously raised my volume enough to push my audio feed into the red zone. Jamulus has no room for error — you go into the red zone, your audio feed sounds horrible, and all the other singers have to mute you. I adjusted my mic volume down, and that solved the problem. And it wasn’t just me — another singer had the same problem.

I also wonder what will happen when we try music we don’t know well. So far, we’ve been sticking to tunes that all of us know well. When we start learning new tunes, for example some of the new music in the Shenandoah Harmony, will it just turn into chaos?

Nevertheless, singing with other people in real time — even in not in person — was enormously uplifting. The pandemic can really get you down, and this singing was a good antidote to that.

One challenge with Jamulus is reducing latency, and in some major metropolitan areas (like the Bay Area), you can have a great deal of latency to singers who live close to you, even if you both have good fast Internet service. Because if the packets have to travel up to a Tier 2 network, then up to a Tier 1 network, then back down to a Tier 2 network, and further down, it’s going to take a while — and the Bay Area has so much Internet traffic from small users up to really big users that things are just going to get slow at times.

In online forums, I came across one way to reduce the potential latency for all singers: host your Jamulus server in the cloud. The idea is to host your server on something like Amazon Web services that has a data center near you.

So Mark, one of our Bay Area singers who’s also a software engineer, set up a Jamulus server for us on AWS’s norther California data center. We tested it this afternoon with four log-ins totaling five singers, and the latency was better than I’ve experienced using Jamulus servers that are hosted locally in Palo Alto or Mountain View.

Not to say the latency was low. I probably had the highest latency, ranging from about 60-70 ms to 40-50 ms by the end of the session. This was counter-intuitive, because although the location of the AWS server is a Big Secret we sort of know where it is, and I was probably physically closer to it than anyone else in the session. Cyprian, who joined us from the North Bay, had better latency than I did, though he was probably 75 miles farther away. Also of interest: Adam joined briefly from Seattle, and his latency was about the same as mine.

Because of the latency, we had to keep the tempos slow. Then we often let the tempo slow down while singing a tune. And once or twice a tune just turned into chaos.

Nevertheless, it was so good to sing with other people in real time that all the frustrations were worth it. I hadn’t realized how much I missed singing in four part harmony.

I had to go back to work, and only stayed in the session for an hour. But it was good enough that I’m looking forward to doing it again.

David set up a Jamulus server at his house in Berkeley, and we both logged on this afternoon to try it out. The latency was perhaps a little high, and it was strange at first singing without being able to see David leading, but we got used to it pretty quickly.

We sang a couple of tunes, then noticed that someone had dropped in to listen. We wound up chatting with Rob from Rancho Cucamonga. A trained musician and a guitarist, he was interested to hear us singing solfege; he’s just discovering Jamulus and was listening to different sessions to get a feel for it. It was sort of like when you’re singing Sacred Harp in person, and someone hears you and comes in out of curiosity. This could be a side benefit to singing on Jamulus!

David and I sang a few more tunes, then we both had to go. While it’s not the same as singing in person, it was really good to be able to sing together with another person in real time — it was also really good to be able to sing with David once again, who is a really good singer.