This image pretty much sums it up.

I realized recently that I hadn’t put in a lot of work on my two Ardour projects. This is partly because I have a part time job that was working me full time for awhile, but there was something else to it, too. I dreaded working on them. A file which had become corrupted and stalled the mixing of a project had left me at a dead end in one, and in the other, the process of trimming audio and getting it to match the video had become so tedious that I simply just didn’t enjoy it anymore. And I typicall enjoy the tedium of work like that – there’s something kind of Zen about it to me.

While I still stand by the fact that Ardour is miles ahead of any other open source DAW I’ve tried, I must say it is unusable for regular stable audio production. And that’s okay – I want to see Ardour become better. To a degree, instability and the need for workarounds is somewhat expected in the land of open source. But in the meantime, it has a lot of stability issues to work on. Its interface has drastically improved between 2.x and 3.x, but the tools within it could still use improvement. Trimming is clunky and I’ve had tons of problems with plugins crashing (which I’m not even sure if the problem is Ardour in that case). When editing an 18+ track session on a laptop, I began to hear fuzzies and clicks in the playback while mixing (but at least it wasn’t in the final bounce).

In all, it became so dreadful to me that I’ve scrapped the nearly complete mixing session. I could have done it three times faster in Pro Tools, and I’m honestly going to get it done faster if I go straight to PT rather than try and fix little details in Ardour. At least with the video project I can export it and keep some of the careful timing that was done there.

In the end, it was big waste of my time and energy, when I could have been doing something more productive. Ardour and I can still be friends, but we’re not working  together anymore.

If you aren’t familiar, it’s a big ol’ event about Circuit Bending (let’s take electronics and repurpose them sonically though modification), complete with 2 stages, over a dozen acts, workshops on circuitbending & DIY electronics, and some interactive Art. I’ll be running sound there, and I’m assissting with a collaboration between Josh Gumiela (who helped me build Vidarr) and Kelli Hix, Curator of Moving Images at the Country Music Hall Of Fame.

The Ball itself is on September 29th at the Brick Factory here in Nashville. Workshops at 10 am, with music at 8pm.

You can find out prices and more details by reading Tony’s big post about it all on Theatre Intangible. You can also visit the Kickstarter Page to donate.

I’ve managed to pick up some part time work with some friends-of-friends, and I think the next month should show some new work. I’ve got some interviews late in the month, who knows where I’ll end up.

I’ve decided that for my review(s), I wanted to use it for my two main projects that I prefer: Mixing a prerecorded track (I have nowhere to record right now anyways) and adding sound to picture. **Note: I ended up abandoning this project due to frustrations with Ardour’s stability. Blog Post Here.**

Installation

Installation requires you to install some extras. Long story short, it’s a driver setup to hook up your inputs/outputs into the software. I used JackOSX. It’s pretty simple, and it runs a little background app call JackPilot. Once I read some manuals for all the programs involved, I started in. I also wanted to work on sound for video (For Mechanical Monsters). However, Ardour does not support video natively, and I think I read somewhere there is no intention for it to do so. However, you can also run a program called Jadeo, and simply use JACK as the sync clock, and it works just fine.Certainly, I had a busy taskbar.

GUI & The Mixing Experience

The GUI parts of Ardour are very easy to understand if you’re used to Pro Tools or Logic. The basic edit/move, trim, and drawing tools seem fairly obvious, though you don’t get the handiness of the Pro Tools Smart Tool. You’ve got a Mix window and an Edit Window. Your tracks have labels and controls in a very similar fashion. Ardour uses Regions in a similar way, and they even have a menu for them. I like how the text changes color if a file was missing, or if it is an edited sub-region. There’s far more options for control  and window focus than I would ever use, but it is nice to have them.

The Mixing window is comfortable as well. I thought it was cool that you actually could place pre- and post-fader plugins above or below the fader itself. An odd thing that I noticed is that the visible track options between the Mixing window and the Editing window are completely independent. I noticed this when I hid 3 of my extra muted vocal tracks in the Mixing window, and yet they were still present in the Editing window. Same thing with groups – they’re completely independent of each other. Reflecting on this, I’m betting that driving idea is that since you are focusing on different things in these environments, you would want to have independent control. From my Pro Tools & Adobe Audition background though, it was schizophrenic. If these are indeed the same session, why aren’t the controls together? This is especially perplexing to me since Track Order & Color are shared between windows.

However, Ardour has a really cool feature in the mix window where the Master Audio Channel floats on the right hand side over your other tracks, rather than have a master volume indicator in the transport area or menu GUI areas.

Ardour has a weird visual for Panning. If you take a look at the piece of screenshot I’ve got here, you can see that there is a black box at the bottom, with a green line. The green line can be moved left or right for each channel, much like the knobs on a Pro Tools fader or a real mixer. However, I find the lines to be hard to read. Perhaps if they had a solid color linkage to the center, it would be better. But I do like the differentiation from a real life mixer – we are using software, so why limit ourselves to what we have in the physical world? I didn’t feel the same way about the gain faders though – there’s a lot of grays in Ardour, and the light gray on dark gray faders just kind of get lost in it all.

General Editing like splitting regions, placement, and so on works well enough. Again, I miss the Ardour comes with a nice little feature where you can download sounds directly from Freesound.org, which I was a big fan of, but led to complications. You can’t simply sample a sound before you download it, and after selecting like 6 of them, when you download, it downloads all of them. I think. All I know is that I thought I was downloading the files once, but my regions list showed multiple copies of everything.

This, in turn, led me to discover  some problems with the Regions List. If I wanted to remove a region from the session, it wouldn’t always actually remove the file. I even went to the session folders and deleted the region. Then, reloading Ardour, I would tell it to remove the pointless region reference (highlighted red now), save, and quit. When I reloaded the session, the region was still there. Lord have mercy on your soul if you dare to want to remove multiple regions at once. You can select a a range of them, but when you tell it to remove, it will remove some – but not all of them, usually. Sometimes it would only remove one. I’m not even sure what was really going on here.

Routing In the Box, as I’ve experienced it so far, is a huge pain in the ass. Let’s take setting up a an Aux fader for a Sub Mix Fader for example. In Pro Tools, you click the output, and a menu comes up which you can hover over to find your inputs and outputs, with the option for paired connections. For Ardour, you click the output, and go to “edit” (Master Fader and Out 1&2 were the only other options). This brings up a window, where you have to remove the current output routing, and add in the new one from a menu. You must do the left and right channels separately, during removal and adding. It’s a really tedious process, and groups don’t make any difference – a 7 channel sub for drums took seemingly forever. I suppose the idea here is that it allows maximum flexibility, but what it ends up doing is creating a crummy process for doing a simple task. Pro Tools has the same options, even sending out to multiple outputs – but it doesn’t clutter you with options like sending the left to one channel and the right to another right off the bat.

I ended up having another issue with the routing… sometimes, I had to ‘Solo’ the Master Track for it to receive audio. Other times, Ardour simply refused to play anything. On my current mix, suddenly the Drum Aux has suddenly stopped sending audio to the Main Out*. The majority of the cases I booted Ardour though, it ran stable.

Groups aren’t greatly implemented as well… you are only allowed to assign a single group to each track. This was kinda disappointing, as I usually have a couple groups that overlap (i.e.: Drums, Bass instruments only, Strings).

Automation took some getting used to. Just like Pro Tools, there is both Region Automation for easy fade ins/outs, as well as Track Automation, for controlling the track throughout the session. You can also automate plugins, though I didn’t do that for my sessions. Track automation comes with touch/write/play/manual modes, which do about what you would expect. Recording automation while playing a track back creates a lot of data points, so you might want to go back and clean it up by hand if you do it.

It took me a few hours of trying what seemed obvious and going back and rereading the manual until I figured out how to manually draw automation. The “Draw Automation” tool doesn’t really seem to be able to create new points. Instead, you use the Select & Move Objects tool (the Hand) to manually create and connect points.  Deleting points manually is also tedious. It can be really hard to mass select a number of points, and it took me a bit to setup the hotkeys in a way that felt familiar and right.

Plugins are something I want to write about, though keep in mind, they don’t come with Ardour itself. You have access to a range of LADSPA and LV2 Plugins, as well as VST on Windows (though you’ll have to compile it yourself). Some of these plugins (I’m looking at you, Steve Harris EQ) caused Ardour to crash. Some of them were pretty nice, although not well laid out. I blame this on software engineers, who, let’s be honest, see the world a very different way. On the good side, you have access to those nice Apple plugins that come default on a Mac, like a pretty good EQ and very nice MultiBand Compressor. Perhaps a further review will be in order later.

Summary

The Good

Ardour’s got a lot of functionality. It is a far better multi-track DAW than Audacity. You can do better routing and sub-mixing than in Garageband. The layouts are familiar enough, and its featureload is compareable to what I would use in most music editing software. Depending on your donation, it could be a really bang-for-your-buck DAW.  With open-compatibility options for plugins, it is likely that you’ll at least find something out there that does what you like how you like. Even though you can’t do video natively, it isn’t too hard to bring it around to working. Ardour sounds fine to my ears, and it handles things well without too much latency.

The Bad

Ardour suffers from serial clumsiness. Simple processes like internal routing, automation, and grouping have issues with usability and intuitive design. None of this is prohibitively bad, but it creates a set of obstacles to deal with that can slow you down at the beginning especially. There are some stability and oddball issues that I have while using it, but I’m not sure if it comes from Ardour or JACK or something else.

Please also note that Ardour’s problems are like those you see in many large-scale complicated open-source collaborative programs. On their own site even, they clarify this is work done primarily by one programmer, with two supporting coders, and a dozen or so contributors. So, if you’re used to dealing with open-source projects, you probably won’t mind the little hangups here and there. However, the audio issues I had near the end of my music mix were fairly unacceptable. But, thems the ropes of  GPL software sometimes.

The Takeaway

I don’t think Ardour is going to become my DAW of choice anytime soon, but it will be something I try for some projects when I can’t be with my Pro Tools rig. The automation and routing issues, if you hadn’t guessed, drive me up the wall the most out of anything else and honestly are the biggest thing holding it back. At the same time, I feel that Ardour has a lot going for it in terms of complexity – Even with all the issues I have with it, I’m kinda rooting for it to improve.  It’s likely something I will donate to in the future, and for any sound editor, I’d suggest giving it a shot.

*Looking Forward

As I was wrapping up this review, I began to wonder about future versions of Ardour, and what might be in the works. In a chat session on IRC with the OSX Support group, I found out that the soloing/routing in Ardour 3 has been completely overhauled. I opened up my other in-progress session to take a look. My complaints on Grouping have not only been resolved, but improved – you can easily create groups and see them, and create a subgroup fader with a simple right click, which is really cool. The pre/post fader section is gone, which is ok. The visuals are much more polished – the buttons have a lot more separation, and the individual elements and controls are much easier to see thanks to a little bit more color. The routing menu is different too, and while you still have to manually disconnect if you’re trying to bus, it’s much more pleasant to the eye, and easier to use. Take a look for yourself.

In that vein, I’m starting work today on a new project: a complete teardown and rebuild of the audio in a Supes cartoon from 1941. It’s titled “The Mechanical Monsters.”  You can find it on Archive.org and YouTube. I’ve embedded it below. One of the things I noticed after my first viewing of this is how much of the cartoon is based in music. There’s sound effects for the most important visual events, but the majority of the drama is portrayed through music. Granted, some basic googling showed me that Sammy Timberg, the composer at Fleischer Studios, was a pretty talented guy, and did a lot of music around that period, so it’s no surprise that they featured his dramatic tunes so prominently in the episode. In older cartoons, before they did vocal dubbing, the whole cartoon was usually musical, kind of like a prehistoric music video. So, it may be a part of the natural evolution of cartoons on that point.

But I got to to thinking about it, and I’m curious if the transition from radio had any difference in it. In the Adventures of Superman radio play, they can’t simply have music play while Supes does his thing. Instead, they have to convey action through sound effects and dialogue – music can be cued, but it can’t be the sonic focus. When you have a visual, you can have scenes where the action explains itself, and the music can dramatically play along and accent it. I think this is really observable in the Superman Cartoon that Fleischer studios put together. Its really the sonic centerpiece. In my own exploration, I won’t be able to do the same… I’m not really a composer. We’ll see where it goes!

As I waited, I reflected on my last major audio play (If on A Winter’s Night, A Traveler) and some of the issues I faced while producing it. When I first started working on it, my plan was to include most of the narration from the book. To me, part of the attraction to Italo Calvino’s writing in the book is that it is intensely verbose and self-aware. He plays on assumptions, lets the mind wander back and forth about the time period, (is it a new train station with electric trains? or an old one with steam trains?) and actively teases at the ongoing theatre of the mind in the reader. I figured the best way to include this part would be to do a spirited reading of it underpinned by ambiences and sound effects – at least for the scene setting. Once it hit dialogue, I’d drop the voiceover and let the setting narrate itself.

Well, as time went on and my voice grew hoarse, I began to realize that even just using that setting dialogue, it was a lot of speaking, and it put me at a crossroads. The production just didn’t feel right. It was sounding like an audio book, not a radio play. This was odd to me – I hadn’t really ever considered the differences between the two – or even considered them in the same realm. After all, radio plays are an old art, full of rich sound and voices, filling the theater of the mind. Audio Books are just a well produced narrator (granted, with a great voice) in a booth reading you a book since you’re either too old or too busy to sit and read.

But then again, that’s not really true is it?

For one summer, I interned at Full House Productions in New York City –  now called CityVox. At the time, the market was a little slow, so most of their staff were handling the editing and recording work. I made copies, dealt with archiving, and hung around during sessions. I got to sit in with Lou Verrico, who does a lot of work for Ken Burns. Anyways, I remember I sat in on a lot of the voicework sessions for Last of the Mohicans (I’m 80% sure it was this one). At the time, it hadn’t occurred to me that what they were doing was so close to the traditional play. In fact, they included radio play elements such as ambience and some sound effects via post production.

With Traveler, I ended up realizing that my focus on narration from the original text was misplaced. Calvino wanted the reader to be explicitly aware of their imagination, and while the narration is an excellent device for this in the book, it doesn’t translate into the audio medium as well. So,I dropped the narration for  a large section of the beginning, and instead focused on using the Theater of the Mind, sculpting a dense environment of sounds inside the cafe and train station. I was careful to make sensible panning – as the character turns, the steaming espresso machines remain in the same spot of the room. Footsteps change on different surfaces, and you can hear the protagonist’s wheeled suitcase as he struggles with it. I like to think that I healed off on narration long enough that the listener must become very focused on their surroundings and question exactly what they are hearing. I even mixed it up a little – in the beginning, the character arrives on a steam engine – in the end, they leave on an electric train. I think it ended up being a pretty good approximation of what Calvino intended… though it’s only a single chapter of the book.

So, what separates the two?

Any “Radio Play” that I’ve done has been completely mixed and edited in post production. I don’t think there are many that are performed live like they used to be anymore, except for Keillor’s Prairie Home Companion. Does that make my work an “audio book”, since it isn’t live? I suppose it is missing the live element that would connect it directly to theater. However, I feel the main difference for separation could be seen on the focus. To me, audio books have a more specific focus on the text of the story – for example, a single narrator, who maybe does some light voices and intonation changes, like a father reading to his kid. Or really anything without a narrative. Radio plays, mean while, focus on the narrative and the theater of the mind – more rich in ambience and sound, and less focused on telling you what is happening, unless it is necessary. Of course, like all things, it’s more of a spectrum between these two than a black-and-white issue.

What do you think?

One change since the first Infinity Blade game is in the Voicing, as other reviewers have noted. In the first game, everything was done in an alien non-understandable language with subtitles. In this one, the voices are English. Some reviewers feel this removes some of the “exotic” atmosphere from the first game. It’s understandable, but I have to say, they really did knock it out of the park for voice actors. Jim Cummings does the voice of The Archivist, and for the short part that it is, I immediately recognized it and had to go look him up. A “Fred T” is also credited, who I suspect is Fred Tatasciore, who’s done work for Epic before.

The music of Infinity Blade II is improved. They’ve added some more battle themes, so it doesn’t get as repetitive. It also helps that the combat system is a little more dynamic, so you have less time to really notice it. However, I did notice the same “end of battle” theme was kept – when your opponent is low on health, the music gets a little dramatic. I’m quite familiar with it, since in the first Infinity Blade I would sometimes end up waiting for an opening forever. I would hear it loop over and over and over. While it is important to provide a sense of continuity with the previous game’s music, I felt it could have been a little more adventurous at times. However, main thematic songs and ambient tones fit well enough. There’s a pretty good interview with the composer up on the Epic Games website, so there isn’t much more I can say about music for the game that isn’t already said there.

Now we get to the sound effects and ambience. Here are the good parts: sword sounds and environmental sounds are all well and good. They have as much variety as needed, and sound good. There’s even a bit of running water in a courtyard that I still think is dropped in just right.

Here is where I want to get critical. During battle, when you damage your opponent, they of course cry out in pain. Human knights sound human, and Monster-y ones sound monstrous enough. There’s a few different sets of noises. Some, such as the Wooden Jester, have special sound sets. Others, such as the various versions of the Wretch (a big lumbering beast that sometimes appears with different sets of armor) share sounds. This is where I begin to take problem. As most of the game is run like a beautiful version of Mike Tyson’s Punch-Out with swords and a bit of plot, you spend a lot of time fighting. So, you get used to hearing these sounds, as they are part of the gameplay – the grunts before a swing may be your cue to dodge, and a loud cry indicates a successful hit. The problem, to me, is that the variety available begins to run thin after a few hours into the game. Perhaps it is because I have a tendency to play in marathons: I pick a game, and play it exclusively for weeks.

This problem of repetitiveness is accentuated by one of the new weapon choices – dual wielded swords. When you enter into a combo while dual wielding, most of your moves consist of spins which hit with one weapon right after another. For each sword hit, the enemy makes a pain sound, and the beginning of one sound doesn’t end the previous, so you have a lot of situations where two vocalizations are being played on top of eachother. I think the intention was supposed to be to have an indication of both hits, but the continuation of both sounds is physically impossible, and just kind of lazy hit recognition at best. Strangely, even fans of the game don’t see it as a bug, at least not in the Master Bug list thread they’ve started and maintained on the Epic Games forums.

This got me thinking about how the audio works in the game, and how much space should be allocated to it. The Apple Developer Guide for the App Store seems to indicate a maximum size of 2 Gigs. Infinity Blade II comes in at just a little over 1 GB. It’s a gorgeous game, and with the large number of resources it probably needs for the Unreal Engine and so on, I’m frankly surprised they managed to fit it all in. It wouldn’t be hard to fit the extra sound in. A 5 second stereo uncompressed .Wav file comes out to about 1 MB of sound.

With what little I know about games, I’d bet it’s likely that they are probably using some form of compressed file (Ogg or Mp3 if they are paying the license), and these are probably mono sounds, since there is no real conveyance of space within the game. Maybe they are generous and using FLAC or ALAC. Either way, it’s likely that each sound is using even less space. There is a lot more that could be fit in there!

I’m reminded of when I attended a workshop by Doug Quin, a nature recordist who did sound design for Jurassic Park III, Werner Herzog, and a few other films. Quin talked about how they had to be careful about using sound from the previous films. Our brain is very good at recognizing sounds – even very short ones. Think of how recognizable some songs are by a single bar, or even a couple opening notes. Your brain can do the same thing with sound effects – have you ever been jarred from a commercial or movie because of a super-familiar sound? Maybe it was the Wilhelm Scream, Youraagh, or the Door Sound from Doom. I even once heard a loop from Garageband in an insurance commercial. Even tiny sounds are super recognizable – if you’ve played Doom, you probably heard it in your head the second I mentioned it.

Granted, we have to make space for sounds, as mentioned before. Then, you have to get the voice actor to come up with tons of tiny variations of nonlinguistic vocalizations. What if there were another way? Some games I think already play with this – they may use a single sound, and use playback speed to create minor pitch variations. What if it went a step further? Imagine that sounds used repeatedly were split into separate pieces – the Attack, the Decay, Sustain, and Release. Then, these pieces could be randomly (or manipulated by the physics engine of the game) reassembled, stretched, and compressed to create even more sound variation. I just wonder if this could create a processing burden, and if it could be fast enough to not cause lag or delay in the game interaction. As processors and programming moves forwards, I bet it will be possible. But would it be practical? I’m not so sure myself – storage space is becoming cheaper and cheaper, and what’s a little more time with the voice actors when you can use it all? And on that note, I cease my rambling thoughts.

From CultofMac.com

I was pulled in by the idea of a “beautiful” and “elegant” way to manage those little proprietary 30-pin connectors. The designer’s solution, for the record, IS pretty clever – just create 2-part charger that has a cap that mirrors the other side. With the European power plug, it makes an easy set of little posts to wrap between. Seems fine, right?

This is horrible for the cable.

As anybody who has to deal with audio cables should know, this method of cable wrapping is really bad for cables. It twists them against their natural curve, and starts wearing them down internally. It also creates “memory” where the cable stars having bends in it all the time because of the unnatural twists that have formed in the wires beneath the insulation. (Here is how to do it proper)

Now, with things like heavy duty power cables that have really study wires in them, I’m a little more forgiving about improper wrapping. But Apple’s super skinny USB wires are known for this. There are posts on Metafilter, Quora, and the Apple Discussion forums about it. Heck – here is a Kickstarter for a silly attachment to try and prevent the problem. Granted, most of these are focused on that spot right before the plug itself. But it’s foolish to think that the strain of wrapping doesn’t cause anything within, or contribute to the problem as it is. Fortunately, in the last year, Apple has done some work on the cable, and tried to reduce the strain issue.

But look over there. This thing puts strain right on one of the most common breaking points. A better solution, though bigger, would be to increase the size of the Red part of the plug, making it taller, so that the cable has some space before it gets pulled around the outside of the device. At the same time, this just moves the strain point, and it still is improper cable wrapping. I’d really like to see some way to wrap commercial cables neatly with devices that lets you use under-over. Maybe something where you wrap the cable properly, then there’s a clip that holds it tightly to the side.

Wrap your cables using under-over all the time. Always.

(And to be fair, I hate the standard magsafe power adapters for laptops, which encourage the same problem)

I’ve got this odd screechy thing I can do by tightening my throat an inhaling, it creates a sound kind of like a pterodactyl. I did a little demo of it awhile back and uploaded it for free use on free sound. Ragtag Studios, an Indie Developer, used my little creation in the preview for their upcoming iOS title, Unstoppable Fist. Listen for it near the very end… I think. It’s pretty deeply mixed.

This is a nice little extra bit to add to the ol’ list of work. I’ve also got sounds in two Indie Xbox titles, Text Zedventure and Horn Swaggle Islands.

Now, what I want to point out is something very subtle, right at about the :23 second mark. Play it again. Do you hear it?

It’s a car horn, really faintly in the background. A piece of non-visible but believable non-diagetic sound, and believe me, it is no accident it is there. It’s a laugh cue; that little extra bit to push it over the edge. Sound for video is often completely done after the video – dialogue is ADR’d in, and better ambience is recorded. I can guarantee that the sound guy on this put that honk there perfectly to punctuate the end of the joke.

You can see a similar gag in this Sprint Commercial I see constantly on Hulu lately: