Feb. LG article on Audacity

From Juhan Leemet

I guess you have probably got some feedback on it already?

You have a "busted link": the page refers to "things-amplify.ogg" but the file is actually called "things-amplified.ogg" in the directory. Minor, but "oops!" I don't know if you can correct it on-line? What's policy on that?

[Jimmy] You know, I'll have to ask about that, but I'm pretty sure it'd be OK to ask for it to be fixed. Thanks for pointing it out.

Do you mind if I forward this for possible inclusion in our mailbag? You've brought up quite a few interesting points, and I'm too lazy to rewrite my answers

FWIW, I'm also interested in music and recording, and have been off and on (minus computer career interruptions!) for a long time. I used to play rhythm guitar, hack around on some other instruments, and got reasonably proficient (pickup bands, jams) on bass guitar. Now, after decades, I am (re)learning guitar, in a number of styles: mostly blues/rock, but I'm also interested in fingerstyle picking (folk, old country), but a bit intimidated of (light) jazz. I love the way those guys can "comp" and even do solos with partial chords. Come to think of it good rock guitarists do that, too! I found myself a great local guitar slinger who took me on as a student (blues/rock). Great! BTW, I agree with your comment on recording yourself, as a learning aid.

[Jimmy] Chord fragments are an essential tool! I didn't really get any guitar lessons -- my uncle showed me 5 chords, and how to tune the guitar, a friend showed how to hammer on, pull off, and play a power chord, and a trainee teacher at school showed me the minor pentatonic, and encouraged me to throw in chord fragments from it at will

I also find recording to be useful when writing -- I've lost count of the amount of times a 'mistake' I didn't notice while playing became part of the riff.

Comments on your clips: they are all really low level: at about 1/5 scale on my M-Audio Audiophile-24/96. I suppose I would have tried to get it up to about 1/2 scale myself. I compared it to a cut from my guitar teachers CD and that went pretty well full scale... of course, that is what mastering (post processing) is all about. I would have thought that you could get a bit higher signal level, though. Depending on your sound card resolution, you might lose signal quality. Bits you lose, you will never get back. OTOH, if you clip, you can never get the original back again, either. Balancing act.

[Jimmy] I thought I said this in the article, but I could be wrong -- I had an unusually difficult time concentrating last month, and it affected the outcome of that article.

I was aiming mostly at the absolute beginner to recording while writing, which is why I included so many beginner's tips. The person I had in mind was the person who would try to connect a cable straight from the guitar to the sound card, and who would wonder why they were getting so much gain when trying to record a clean sound.

I'm writing this on my way out the door to work, but I do remember recommending the use of an amplifier or pre-amp. In my own tests, I was able to record at roughly 80% input volume before getting any noticable gain. I could (and should) have mentioned this, but as I said, my concentration wasn't what it could be.

One "funny" thing about your examples: it seems that the compressed clip has a lower level than the amplified one. I'm not sure I understand why, I thought compression was used (with amplification?) to boost the signal and get a more "powerful" sound? With those really short clips it is hard for me to tell much difference. I cannot say I have a trained ear. I'll have to listen more.

[Jimmy] Well, gain is usually added with compression, but the two are actually separate. That's something else I could have been a lot more clear about: I meant to make the point that it's a lot better to get the recording to sound as well as possible at the time of recording, rather than trying to rely on being able to fix it afterwards.

Headroom is one reason I chose the Audiophile card, based on an envy-1712 chip, which can use 24-bit samples. Lots of headroom! With such a card, one could actually master a relatively low level signal such as yours, and lose a few bits from the top, but get them back when you boost up the signal, since you need to trim some lower bits for 16-bit CD quality anyway. I'm happy with that new card, even though I already had a SoundBlaster AWE64 before.

I scanned your 2 articles again, and I see no mention of the gear you are using. I guess you are trying to keep the articles generic, and avoid mentioning (pitching?) specific vendors. Good plan. However, some details of the kind of equipment might be useful to others, while not directly relevant to audacity (but I suppose it is, if you're trying to record with it?).

I guess that is an acoustic guitar you are using? Is it electric-acoustic (plugged in?) or plain (mic'ed?)? Do you use an outboard mixer or preamp? Most seem to like an outboard mixer/preamp (better quality, less noise)? Some of those questions would be particularly important for electric guitar. You might also mention something about headroom and sound quality.

[Jimmy] I was using a Washburn Dime333: electric, with a floating bridge (which is why I had the paragraph about tuning -- that was cut down from a lengthy rant about the pains of trying to tune a guitar with a floating bridge . I did mention that I had plugged the guitar straight into the sound card, and that it would be much better to use an amp or pre-amp. In future, that's what I'll be doing, and I'll be sure to mention what I'm using.

I have no idea what my sound card is though -- all I know is that it's cheap and nasty. I'm looking at getting a proper sound card some time in the near future, but for the time being I'll just be recording using a dedicated recorder, and using Audacity to reassemble and mix the tracks.

This might fall outside the scope of your articles, but a problem I have seen bounced around in various newsgroups and mailing lists is how to add drums/percussion to a recording after the fact, i.e. sync them up. That appears to be a tough problem, and we may not yet have the tools to do it, with Audacity at least. I have played with Hydrogen, the programmable software drum machine. I am very impressed with its capabilities, and I'm only learning how to use it (not a drummer). However, I don't know how to get timing information out of Audacity (I don't think you can yet?) and use it to drive hydrogen "in step" with other prerecorded stuff. I imagine that trying to match the start and beat manually would be almost impossible? Putting the drum track down first would work, but (unless you are a drummer?) that would be really hard to do. I don't think I could. Studios use click tracks?

[Jimmy] That does fall outside the scope of my articles, but into the scope of a future article.

A click track is just a simple drum track -- snare only -- to be used in place of a metronome. It'd be boring if every song kept to the same tempo the whole way through, but for tempo changes a metronome is useless, so you create a click track instead. Audacity is able to use MIDI files, so the best approach is to create a click track in advance, then use that as the basis for the drum track on the drum machine/synthesiser program -- it cuts out all of the problems (I was going to recommend the use of the metronome in Audacity, but couldn't in good conscience do so, because a) I've never used a metronome, and b) Audacity's metronome annoys me

FWIW, one of the things I have wanted to do, to really learn some songs (thoroughly) is to use the original as a "backing track" (apologies to Ashlee Simpson who might have it "patented", ha! just poking fun!). Then layer on my own interpretations of different parts, until I have enough for my own reference tracks, and then drop out the original, and continue layering. Audacity can do the layering OK, but how to get that drum machine synced up and recorded in there? I've heard rumours that Ardour can do it. True?

[Jimmy] Audacity isn't able to do it yet (at least, not in the version I have installed), but it's something the developers are looking at. Ardour, AFAICT, can do it, but it's overkill for my current needs, so I haven't looked at it in depth.

Just getting back into music and recording, and (re)learning a lot of stuff.

[Jimmy] Same here

BTW, is there a good Linux guitar tablature editor, like PowerTab on Windows? Haven't experimented much with lilypond or noteedit, but my impression was that they were mainly/mostly for (orchestral?) scores. Can they do tab?

[Jimmy] I like Songwrite, and my first two 'Songs in the Key of Tux' articles were written about it. It's not as powerful as PowerTab, but I haven't seen any other tablature programs on Windows that are as powerful as it, either. One particular feature that I've never seen elsewhere is the ability to represent harmonics other than the 'standard harmonics' (over the 5th, 7th, and 12th frets) or tapped harmonics.

Lilypond can do tab, and, in fact, Songwrite uses it as a print backend. I recall seeing that noteedit took code from KGuitar to represent chords, but your assumption is correct, both of these programs are primarily intended for creating staff notation, and neither are particularly good at representing tab.

Tablature is something I think I'll have to revisit, because recently I went through an acoustic phase (I don't understand it, really, because I listen to and play mostly metal) and found I had to use two different programs to represent my tablature: Songwrite, to get the correct note durations with it's MIDI playback; GTablature (http://www.solutionm.com/gnometab/gnometab.html) to get a nicer print version.

BTW, there's a set of tools for converting PowerTab files to Linux friendly formats: http://jelmer.vernstok.nl/oss/ptabtools (which I see now supports GuitarPro files).

Also, I recently heard "the amazing SlowDowner" on my guitar teachers PC running Windows. It seems to have much better sound quality than the SndStretch sound stretcher plugin for XMMS. That XMMS plugin seems to create a lot of artificial echo/reverberation, which gets worse as you slow down more. Must be something to do with the (resampling?) technique? Do you know of any other/better playback software that can slow down the recording, without changing the pitch, obviously? Any advice (article?) appreciated.

[Jimmy] Damn. I had something I was pretty happy with, but hadn't used it in quite a while (http://mysongbook.com has some very high quality tabs in GuitarPro format, which (for the most part) render well in Songwrite) and forgot about it when I switched to Gentoo a few weeks ago. I'll make sure to let you know what it is if/when I find it again.

Good articles! Thanks!

[Jimmy] Thank you, and you're welcome.

-- Juhan Leemet Logicognosis, Inc.