Ha! Definitely get some more opinions. I can only talk about what works for me based somewhat on observing others and doing vigorous internet searches
A K-meter or K-meter-esque approach using LUFS would explain why modern releases donāt all hit the exact same integrated level. I chose to set +4 on the K-meter as my loudest part even if the instrument isnāt quite able to give a real fortissimo (88-90+ dB). The 0 to +4 dB (or -20 to -16 LUFS momentary max) zone should be used sparingly. A purist would probably let the master sit lower down. One thingās for sureā¦hitting 0dB peaks for a harpsichord is just weird even on a CD, especially if youāve just been listening to actual fortissimo full orchestra. Listeners will instinctively reach for the volume knob as their ears hear it as too loud.
There may be more so listen carefully over headphonesā¦It is a ādigitalā glitch for which there could be several causes.
Reverb: As much as I like the Samplicity Bricasti IR I wonder if there are other options to exploreā¦either real space IR or algorithmic. Iām not saying what you have is bad but you might want to rule out other options as well as play with settings like pre-delay etc. Something doesnāt sound quite authentic but, of course, you are recording in your house and adding a very different IR on top.
Compression: If you are happy with your parallel compression that is great. It doesnāt jump out at me at all. Does the dynamic range compare well with that commercial disc? Do the peaks sound energetic and climactic?
My takeaway is that your recordings showcase your amazing talents while demonstrating solid engineering practices. Who could ask for anything more? Well, a beautiful church away from traffic
Thank you so much for taking the time to listen to my recordings @anon60445789. And thank you very much for your words.
Technical problem: Right. I hadnāt heard that. Maybe thereās always a lot of noise in my house, itās hard to appreciate but yes, you can hear it clearly. I had to fix it using the first bars of the piece, theyāre the same.
Sometimes Iāve got some clicks, not too many, but itās a problem, it can ruin a take, and if you donāt have more the problem can be serious. I donāt know why, but they were much clearer and stronger, this one was more subtle.
Do you have any idea what could cause them?
Reverb: We were discussing different types of reverb on a separate thread, do you remember? I tried some Dragonfly configurations but the Samplicity Bricasti reverbs were the ones I liked the most, and specially the one I use, Vienna Hall. Do you think I should buy the Impulse Records pack and try some?
The plugin that worked best for me was the Robin Gareus one, with other plugins it didnāt sound as good, I donāt know why. The problem with Robinās plugin is that you canāt practically control anything, only the output gain, which I have around -20 dB.
Itās true that there can be something unreal in the sound, sometimes when everything is a little stronger there are even some harmonics in the high frequencies that are a little uncomfortable, I already talked about it when I had the other mics, maybe itās a problem of my sound or my flute. Anyway the final result doesnāt displease me at all, as you say itās nice, and in general all the comments from friends are quite favorable.
Compression and Loudness: I am happy with this compression, I think it practically does not affect the musical result. Comparing with other recordings the dynamic range of mine I find it similar. I do see differences in the audio level, if I compare with wav files extracted from CDs my recording is usually the same or a little lower than the others. If I compare with Spotify my recording is very low level, maybe Spotify raises the level of the audios.
The repertoire on the CD may be another problem, itās interesting but theyāre not great worksā¦ but thatās another issue.
And yes, I think the pending subject is finding a silent church.
Are you free of x-runs when recording? What are your buffer settings when recording? So you are saying you have had to deal with evener stronger versions of this glitch during your recording sessions? We might have to troubleshoot your usb device and other settingsā¦
Then I would leave the reverb exactly as you have it! The Convology halls and churches are not true stereo IR but they are from real spaces unlike the Bricasti stuff. Using the Robin Gareus IR plugin might be part of your glitch problem. I realized that he designed it for cabinets and other short IRs. I couldnāt use it for longer IRs without getting into serious CPU loads. Perhaps try the same Vienna Hall in Klangfalter (install lv2 and vst but I believe Ardour blacklists the lv2 version due to block size issues). Note also that unless you use the browser panel, to manually insert a regular stereo vs true stereo impulse you need to use the 1st and 4th slots or else you will get some weird channel effects
I canāt remember now but how do you deal with room tone before a track? Did you have digital black before the first note or did you fade into room tone a couple of seconds before the music begins?
Since I switched to the new 4.20.0-16.2-liquorix-amd64 kernel I will only have had one or two xruns.
Other newer kernels gave me problems with wifi.
I did have 2 or 3 strong clicks during the recordings.
Some time ago I installed the Klangfalter but it was the lv2 version and I couldnāt use it. I have installed the vst version and now it appears, but when I select the ViennaHallQuad file the program stops responding and the CPU starts going crazy. I didnāt quite understand the last part of what you were saying, maybe thatās what you meant.
Edit: I just figured out what you meant. Itās already working. Iām trying it out
Edit:
Iāve been testing the Klangfalter a little bit and even if I try several parameters itās difficult to get something as smooth and natural as the reverb you get with the Robin Gareus plugin.
Maybe itās just chance and I got this result just by chance.
I also just tried the Dragonfly reverb Plate, and I think itās more natural than the others, but I still like what I have more.
Technically there should be zero sound difference between Convo.lv2 and Klangfalter. In any case, if you are editing and mastering with higher buffer settings there will be no problem with convo.lv2. Obviously on export you are doing faster-than-realtime so the audio device is not used. I suspect there is a gremlin in the way your USB device (Audient iD14, right?) is interacting with the audio backend during real-time operation.
EDIT: Do let me know what settings you use for live recording in terms of sample rate, buffer size, number of periods etc. Also, be sure to disable wi-fi before recording! Plus, I assume you are running the CPU scaling governor in āperformanceā versus āon demandā?
Note also that your āDispositivo de Entradaā and āDispositiva de salidaā both need to be your Audient iD14 device. I just been recording 4 stereo tracks of background room noise on my UMC1820 for 30 mins using a 128 buffer and had zero x-runs or glitches. Itās also worth moving between 2 and 3 periods to see if that makes a difference. For the longest time I defaulted to 3 due to internet advice but it seems in recent times that is unnecessary. For the Audient I donāt really know.
I used the brower to select the quad Vienna Hall file and it automatically assigns to the 4 slots (1-1, 1-2, 2-1 and 2-2). Each slot shows a different channel of the quad file. Iām sure youāve done that. If you are running the reverb on a bus, youāll need to disable the ādryā and probably the autogain to get an exact replica of the sound from convo.lv2.
Yes, of course. Itās just that now the Audient wasnāt connected to my laptop.
So itās better to go down to 128 buffer and try period 3?
I didnāt quite understand that part before, so it didnāt sound the same. I selected the file in slot 2-1. Iām sorry for my stupidity.
Now I added it from the browser, which I didnāt understand before, and yes, it sounds quite similar without touching anything. Maybe with the āDryā I like it a little more, I donāt know, it seems more present.
I added it to the same bus, and to compare them I deselect the one I donāt want to hear.
I donāt touch the parameters of the Robin Gareus plugin since months ago, when I made the first tests and commented it in the forum. What I do notice is how much the output gain affects the output volume.
With the autogain the Klangfalter increases the volume much more than I had it. It really sounds good, but maybe I should lower the 4dB I had added with the knob in the flute channel and check all the LUFS levels again.
If you are using the reverb on a bus, you definitely want to disable the ādryā as it should be 100% wet. With dry enabled you are essentially doubling the dry track which, at best, is simply increasing volume.
Great, Iāll be interested to see if the glitches disappear with performance selected and wi-fi disabled. Hereās a script to run to check for other optimizations: https://github.com/raboof/realtimeconfigquickscan
How should I proceed to execute this script properly?
EDIT:
I think Iāve done well. This is the output:
$ perl ./realTimeConfigQuickScan.pl
== GUI-enabled checks ==
Checking if you are root... no - good
Checking filesystem 'noatime' parameter... 4.20.0 kernel - good
(relatime is default since 2.6.30)
Checking CPU Governors... CPU 0: 'performance' CPU 1: 'performance' CPU 2: 'performance' CPU 3: 'performance' - good
Checking swappiness... 10 - good
Checking for resource-intensive background processes... none found - good
Checking checking sysctl inotify max_user_watches... < 524288 - not good
increase max_user_watches by adding 'fs.inotify.max_user_watches = 524288' to /etc/sysctl.conf and rebooting
For more information, see http://wiki.linuxaudio.org/wiki/system_configuration#sysctlconf
Checking access to the high precision event timer... readable - good
Checking access to the real-time clock... readable - good
Checking whether you're in the 'audio' group... yes - good
Checking for multiple 'audio' groups... no - good
Checking the ability to prioritize processes with chrt... yes - good
Checking kernel support for high resolution timers... found - good
Kernel with Real-Time Preemption... not found - not good
Kernel without 'threadirqs' parameter or real-time capabilities found
For more information, see https://wiki.linuxaudio.org/wiki/system_configuration#do_i_really_need_a_real-time_kernel
Checking if kernel system timer is high-resolution... found - good
Checking kernel support for tickless timer... not found - not good
Try enabling tickless timer support (CONFIG_NO_HZ_IDLE, or CONFIG_NO_HZ in older kernels)
== Other checks ==
Checking filesystem types... ok.
** Set $SOUND_CARD_IRQ to the IRQ of your soundcard to enable more checks.
Find your sound card's IRQ by looking at '/proc/interrupts' and lspci.
I seemed to have various things come back as ānot goodā but in discussion with user āsunratā it seems Iām good to go. Indeed, no x-runs or glitches hereā¦
So it seems that it was all the fault of not turning off the wifi and having the cpu āondemandā. I hope there wonāt be any problems next time I record myself. Iāll let you know if I notice any x-runs or glitches here.