They are used for different things. For capture, you are concerned primarily with not clipping…hence faster peak meters. For mixing/mastering you are concerned with loudness and averages, hence K20, RMS or LUFS equivalent.
I recommend the free Klangfreund LUFS meter (or paid-for multimeter) I referenced in an earlier post as well as Robin’s meters. I’m assuming that at some point there will be both a true-peak and LUFS meter built into Ardour but I might be mistaken…For now, it is all still possible with plugins, the benefit, of course, being that I can use the same meters cross-platform and in whatever DAW I’m using.
As @x42 suggested to me, you can always normalize to -1 dB true peak as part of the export function of Ardour. However, the down-side here is that whatever LUFS level you have carefully worked out during mastering may be brought down somewhat on export if only relying on regular peak meters as part of the mastering. Same goes for using a regular limiter with no intersample peak detection. Whether it is significant enough of a difference is probably all dependent on your material.
My own preference is for getting the desired peak, integrated loudness and LRA before I go to export but each to his or her own.
@anon60445789 I think I have understood these differences well, but there is something I don’t understand, why the values are the same between different meters, for example:
In the image above left the meter “True peak” indicates a maximum value of -7,5 and below left the meter K20-RMS also indicates a maximum value of -7,5.
What is the difference between True Peaks and K20-RMS?
I think the -7.5 reading you are seeing on the true peak meter is actually RMS as it includes both types of measurements (the true peak is read at the bottom of the plugin). K-20/RMS meter is using that same RMS averaging so that’s why you’ll see identical values. I don’t use that true peak meter but I’m sure @x42 can jump in if I’m misrepresenting his plugin.
The numeric display in Ardour is always digital-peak relative to full-scale (dBFS), regardless of the bar-graph meter-type. So in this case there is simply no difference between digital peak and true-peak.
The K-meter spec recommends to display a digital-peak along with a user-resettable digital-peak-hold.
The former allows to determine the crest-factor (difference between peak and RMS). That is a reasonably good indicator of the dynamic-range.
The dilemma is that string quartets and Renaissance music, among other forms, have low crest factors as well as low natural loudness. Consequently, the string quartet will sound (unnaturally) much louder than the symphony if both are peaked to full scale digital.
I recommend that classical engineers mix by the calibrated monitor, and use the average section of the K-meter only as a guide. It’s best to fix the monitor gain at 83 dB and always use the K-20 meter even if the peak level does not reach full scale.
Sorry, getting confused between your true peak meter and true peak + RMS meter. So many meters to remember!
To everyone: if there are further suggestions for USB interfaces, standalone recorders and microphones keep 'em coming! I wanted to get into further detail about the editing portion of classical workflow in Ardour…what is your workflow for the actual splicing and dicing (region splits and crossfading)? Do you cut out applause and, if so, how do you deal with early enthusiastic clapping? Do you crossfade into room tone in between tracks? How do you normally approach reverb, EQ, compression, limiting in a classical album?
There is a shortcut for split, “S”. It works with selected regions only and it works with both mouse and playhead edit points. I usually use it with mouse and I find it fantastic. When in playhead mode, you can splice the region while the music is playing - so you can watch the score and make edits without watching the screen. The only thing that slows me down is that after making a cut I need to select the new region to cut again (instead of automatically have the following one selected, or am I missing something?)- kind of logical, the programmer doesn’t want to guess my next step - but for lots of cuts there is cut tool anyway, and for selecting portion of a region while the music is playing there are “,” and “.” shortcuts and then right click on selection->Separate… or create custom binding in Alt+K->Editor->Editor->Separate, which work wonders when one needs to make edits while following the score.
EDIT: I was wrong about “S” binding! When the track is selected (header highlighted) and no particular region is selected, the “S” shortcut will work multiple times.
I use to cut applauses, but I find that if I do so the tracks often end abruptly, so lately I started to leave a couple of seconds of clap with quick fadeout. Not ideal, but those are charms of live environment.
I avoid compression when possible (exception being soprano singer with huge dynamic range), I much rather use gain drawing (to the limit where difference in background noise becomes obvious), I try to avoid reverb (not always possible)… and use EQ very sparingly, 1-2dB here and there… except for mics like Behringer C2 you mentioned, which lack bottom end. Well, all SDC cardioids do to some extent - or maybe they are designed with close micing and proximity effect in mind. Anyway, in my experience 3-4 dB boost @~150Hz with SDC cardioids make recording fuller and “warmer” without sounding unnatural.
… I’ve been digging some older (10yr-ish) recordings and upon listening I figured there are some really good ones, and when recalling the sessions in my mind I figured they are all recorded with a pair of Rodes nt5 - the one almost everyone thinks is a piece of harsh junk. When comparing Rode and Line Audio CM3 at home, side by side, Rode can’t possibly win, but when comparing remote placed pair in the real-world recordings, I wouldn’t say CM3 is leagues better. It is different, but in some situations you might be better off with Rodes’ qualities. And then there is omnidirectional capsule available for the NT5 which makes it fantastic for an AB pair.
Years ago I had an acquaintance in Boston who worked for a classical record label; her job was to edit concert recordings to remove extraneous sounds (where possible) such as coughs, sneezes, inadvertent taps of the conductor’s baton on the music stand, chair squeaks, etc. The end result was that she could no longer enjoy actual live concerts anymore because she couldn’t hear the music; her brain was trained to focus on everything but.
As I mentioned before, my aim is to apply classical recording techniques to traditional music, since it seems the classical approach is most suited to that kind of capture, and I’m especially interested in capturing the ambience of the space in which the music was played.
Some of my favourite field recordings were made in pubs or in people’s kitchens, where you can hear glasses being set down, silverware rattling, conversation and laughter in the background. I have an old recording of the East Galway fiddler Lucy Farr and the flute player Chris Ferguson playing and chatting together in Lucy’s flat in London, and at one point one of them lights a match and you can hear the flame swell up. In another the uilleann piper Tommy Reck is playing by the fireside with his baby daughter crying in the crib nearby.
A lot of listeners want those sounds removed but for me it’s an essential part of the experience, for traditional music at least – you’re hearing the music in its normal context. Sometimes I think live classical recordings risk become too sanitized; we’ve gotten used to a pure musical experience when listening to classical recordings, but a live concert recording captures a moment that included the audience as well as the musicians and I’d be inclined to retain most of the audience sounds except those that are truly distracting.
I’ve found the same thing. I wonder whether there is a setting/hack to change this?
Yes, I love the shortcuts. For me, they are the building blocks of a potential (simple) source-destination editing feature. At the very least, it would be nice to copy the source to clipboard in this way and then select your destination in and out in a similar way (as you say, using ears and those shortcuts) and then paste with the default short fades that ripples the destination material underneath as appropriate. I think there are three issues: 1) It is hard to explain in words, 2) If there is full understanding of its editing power, I think it is not a priority and 3) It messes, somewhat, with the layering system in place but I think I have a work-around for that.
Not sure whether this is standard practice because I simply went with my gut…I cut all applause and dealt with any single early claps in RX as necessary. I then found a reverb to sound as natural as possible and this often meant automation at the end of each track to boost the reverb tail. That was my workflow with GVerb+ but I found that reverbs like Seventh Heaven needed less of that tweaking. It isn’t perfect by any means and I’d prefer to keep applause but that is the will of my client. The other thing I want to experiment with is cross-fading to room tone to see if that helps the reverb tail. The problem for live concert recording is that I never have a chance to capture 30 seconds or so of silence given all the milling around. There’s always this trick I suppose: http://hearspotbark.com/blog/2015/1/26/using-logic-pro-to-generate-air. Works in every DAW, pretty much.
Having owned the NT5 pair (purchased the additional omni caps later on) I think I didn’t give them enough of a chance. The omni caps are magnificent but I was probably using the cards too close to the source. The KSM141s are probably quite similar. I nearly was put off buying them but realized that any weird frequency bumps are not that important when distance miked.
Agreed. My ear is now focused on hearing every possible squeak, sweet wrapper and any other distraction. Same goes for any mis-tunings. Luckily, my ears want to work very differently when I’m making the music. My brain has a forgiving tendency to smooth out many of the rough edges during performance so that I actually enjoy it a little more. Weird, but true.
I’ve come to appreciate this far more as the years go by. I think for Folk you are absolutely right that it is part of what people enjoy hearing. For classical there seems there is an expectation that noise is minimized and for studio recordings it is relatively easy to achieve. For live concerts, especially when microphones often have to be placed directly above audience (and often near the excitable family members who sit up front) I find it a real challenge to successfully eliminate distractions. I keep telling the groups that they should really consider a “studio” album even if they only end up recording a few tracks for tour promotion / website etc.
Here’s a loudness and metering question. I often see pretty large discrepancies between the Ardour/Mixbus export analysis and what’s reported in my loudness plugin on the master bus (I use Bute Loudness Suite). For example I’m working with one fairly short file where the Bute plugin shows integrated loudness of -20.7 LUFs whereas the export analysis in Mixbus shows an integrated loudness -16.9 LUFs. What are possible causes of this discrepancy and which of these readings should I trust? I can upload the exported file, which is pretty short (its shortness itself might be the issue but I’ve noticed discrepancies with longer files as well).
The Bute plugin is what I use myself on Windows and it always seems in step with Youlean, Klangfreund and whatever else I throw on the master bus for analysis purposes.
Your uploaded file shows -16.9 LUFS via ebur128 so it seems Ardour/Mixbus export analysis is accurate (of course!). The question, then, is whether there is either some post-session normalizing happening during your export (check your export profiles in Ardour) or whether Bute is set up in a fashion to measure something that isn’t the actual output of the master bus. I’d be surprised if there was a bug in Bute given my own positive experiences but you never know…
EDIT: As a test, throw on the free Youlean Loudness Meter in the same place as the Bute and see if the readings are identical.
Well, there is a sort of clipboard: every newly created region ends up in the Editor List under Regions tab. You can audition each entry and if you rename it (double click) cleverly (by putting a bar number or rehearsal mark or something), it can be used as a sort of source. You can then drag and drop it to the destination. Not quite elegant, but almost there.
Strange! Youlean Loudness Meter on the master bus is consistent with the export analysis: it reads -16.8. Bute is still showing -20.7 or -20.6. I’ll have to see if there’s a setting in the Bute plugin that needs to be tweaked.
Yep, as a starter, check that when comparing your analyzer output you are using the same preset like ITU-R BS.1770-4, EBU R128 or whatever as, for example, you might be using different relative gating or RMS window length settings.
Good point! I generally don’t create new regions when selecting source and destination in and outs as it is simply two sets of markers. Ideally I would want to use those shortcuts by ear and then copy/paste into the destination selected via in/out markers. As I mentioned previously, a few edits here and there are fine in Ardour (with similar level of precision given the transparency when dragging) but S-D editing really comes into its own with lots of edits and especially when coupled with a crossfade editor window that you don’t need to leave until all crossfades are perfected. Again, a crossfade editor in Ardour would probably be incompatible with the current layering system (which is wonderful in its own right for all sorts of non-classical stuff).
BTW, loved listening to the snipper of Sean-nós singing. Made the start of my day that little bit brighter! -16.9 or -20 LUFS she sounds great
Yes, she’s a lovely singer! It’s Máirín Uí Chéide; see https://cceboston.org/hall-of-fame/mairin-ui-cheide-hall-of-fame-2003/. She’s a native Irish speaker from Connemara. I did some recordings of her in a borrowed home in Nashville, Tennessee a few years back; I was supposed to record a flute-fiddle duet album with myself and a fiddler with some tracks by Máirín. The CD project was ultimately abandoned but I did get some nice recordings of her singing. It was a beautiful home with nice acoustics; we got pretty hot as we had to turn off all the A/C, unplug the refrigerator, etc., and a few tracks got spoiled by airplanes. I used a pair of Earthworks omni mics for this, quite some distance away from her.
It seems so. By the way Ardour use to have crossfade editor in its own window in version 2. I remember it well, because I used Ardour 2 as stable version almost until version 5 came out. It was quite complex, besides curves you can find in current crossfade menu there was possibility to construct your own fade in/out, which didn’t need to be symmetrical, using arbitrary number of points, it was quite large and it was easy to align regions in phase etc. But, for my needs, the current crossfade style is sufficient.
Reading about that crossfade editor got me into Ardour/Mixbus despite it not being there in the version I ended up downloading (2013 so v3, I suppose). I’ve used it for much of my work since, however I cannot resist the temptation of working in Pyramix (and previously Sequoia) for heavy editing tasks. Also, having spent most of my life in audio engineering using Magix products (starting with a 2005 version of their Music Studio program) it has been difficult to just switch cold-turkey. Perhaps 2020 will be the year…
Hmm. Perhaps my issue is that out of habit I always seem to select a region before making cuts. Or maybe this was a frustration I had with Reaper. Who knows? Good to know about overall track selection for S shortcut.
There’s also the "Edit | Preferences | Editor | Editor Behaviour | After splitting selected regions: " setting, which allows you to choose “no regions” (the default), “newly-created regions”, or “exisiting selection and newly-created regions”. If you’re doing a series of splits one after the other, with no other editing operations between, the last may be useful.
You do need to be a bit careful in this mode to clear the selection before adjusting the start- or end-point at the new split, otherwise you’ll end up altering all the selected regions.
I took a look at that and the author demonstrates both a numeric way, and then later on their last video(part 4) they show how to use the spectral edit just visually without using any numeric input… I have that tutorial bookmarked, it could be handy when I start looking more into spectral editing…
Thank you Robin, it was a bit confusing for me that the meter bars are in one system and the numerical indication in another, besides there is no indication about it, but it is already clear, the numbers always indicate digital-peak relative to full-scale (dBFS)
Thank you also for the article, quite good, the problem is for those of us who have not calibrated our monitors…
-The Universal Audio Apollo preamps seems to be quite good, the problem is its price.
What do you think is best for a live recording, a standolone recorder or a laptop with Ardour?
It looks like it can work well, it’s a matter of proving it. The Samplitude plugin Pro X looks very similar.