Assembly 2015
category: parties [glöplog]
I usually do check for infra frequencies of all my tracks before shipping, and it's completely my fault I forgot to do it in this instance. Of course with intros there are limited possibilities with "mastering", but I think I could've had just enough space for a couple highpass filters (one for each channel) in the master. Can't always rely just on your ears, shame I forgot to check the spectrums in the hurry I had... few of us have monitors that reveal the problem audibly.
I want to clarify I'm not laying too much blame on the sound guys, I think the hall audio most mostly fine (and still a ton better than it ever was at Hartwall). I still have a lot to learn about taking large PA into account in my mixes :). I do like spiikki's idea too.
One thing I have to mention though are the interpolation artifacts. Most of the tracks mask them pretty well, but the Assembly cap of Hydrokinetics makes it quite apparent. It appears you're running your sound output on Windows at probably 48kHz. Most of the audio used in productions is in 44.1kHz though (with 4klang for an example you have no other choice). There is a known issue with Windows that it uses linear interpolation to do the SR conversion, which causes very audible aliasing artifacts especially with low frequency content. I noticed this already in the compobase when I went to check the intro ran ok. I mentioned it there already (to sivu, I think) but didn't really get any response.
There's a fix available too, but I think only for Windows 7 (finally fixed in later versions, perhaps?), but even that one isn't pushed through the update system. More info: https://support.microsoft.com/en-us/kb/2653312
Though this is the most likely culprit, it's not necessarily caused by Windows. You could have a similar problem somewhere else in your signal path too.
I want to clarify I'm not laying too much blame on the sound guys, I think the hall audio most mostly fine (and still a ton better than it ever was at Hartwall). I still have a lot to learn about taking large PA into account in my mixes :). I do like spiikki's idea too.
One thing I have to mention though are the interpolation artifacts. Most of the tracks mask them pretty well, but the Assembly cap of Hydrokinetics makes it quite apparent. It appears you're running your sound output on Windows at probably 48kHz. Most of the audio used in productions is in 44.1kHz though (with 4klang for an example you have no other choice). There is a known issue with Windows that it uses linear interpolation to do the SR conversion, which causes very audible aliasing artifacts especially with low frequency content. I noticed this already in the compobase when I went to check the intro ran ok. I mentioned it there already (to sivu, I think) but didn't really get any response.
There's a fix available too, but I think only for Windows 7 (finally fixed in later versions, perhaps?), but even that one isn't pushed through the update system. More info: https://support.microsoft.com/en-us/kb/2653312
Though this is the most likely culprit, it's not necessarily caused by Windows. You could have a similar problem somewhere else in your signal path too.
Spiikki that low-cut might be a nice option indeed if it's feasible for the team.
Personally I'm with the "fix your mix" school. Tbh I see no reason why couldn't this be implemented within softsynths too. Most softsynths do have filters, no? And if yours (Joe Scener, the average compo participant) doesn't, surely if you esp. coded the damn thing you can add one, no? :)
Imo it only needs to be sth like a 12dB/oct high pass filter on the master out with the cutoff set to ~50Hz. Shouldn't be too difficult implement and will definitely help with the rumble. Without any factual stats to draw on, I'd still say most demo creators rather would prefer a clear mix rather than have the fruits of their labour overrrun by all the party PA low end boom.. Just because "last 1% of the project, didn't bother to check".
And no I'm not ranting here :P
I'm also not saying the above is a "100% WIN" recipe for success. If you can check the mix on a subwoofer setup, then it's a recommended work stage to fine tune the highpass cutoff & roll-off to better match what you're working with. You could also get sth like Audacity and Voxengo Span and check how the low end roll-off of your song compares against commercial records.. Or better yet: You've been to the party, so use compo entries from any Assembly that you know had bass issues on the PA. Other people fail with their mixes too, so take notes and turn that to your advantage for next year.
Seconding Navis, I too do think that Messukeskus is slight upgrade to Areena in terms of overall sound reproduction (smaller space, different shape). Wasn't at the party this year (no time for an entry) so can't really comment how things might've been this time around.. Simply assuming that the rig or its settings haven't changed much from last year :)
Personally I'm with the "fix your mix" school. Tbh I see no reason why couldn't this be implemented within softsynths too. Most softsynths do have filters, no? And if yours (Joe Scener, the average compo participant) doesn't, surely if you esp. coded the damn thing you can add one, no? :)
Imo it only needs to be sth like a 12dB/oct high pass filter on the master out with the cutoff set to ~50Hz. Shouldn't be too difficult implement and will definitely help with the rumble. Without any factual stats to draw on, I'd still say most demo creators rather would prefer a clear mix rather than have the fruits of their labour overrrun by all the party PA low end boom.. Just because "last 1% of the project, didn't bother to check".
And no I'm not ranting here :P
I'm also not saying the above is a "100% WIN" recipe for success. If you can check the mix on a subwoofer setup, then it's a recommended work stage to fine tune the highpass cutoff & roll-off to better match what you're working with. You could also get sth like Audacity and Voxengo Span and check how the low end roll-off of your song compares against commercial records.. Or better yet: You've been to the party, so use compo entries from any Assembly that you know had bass issues on the PA. Other people fail with their mixes too, so take notes and turn that to your advantage for next year.
Seconding Navis, I too do think that Messukeskus is slight upgrade to Areena in terms of overall sound reproduction (smaller space, different shape). Wasn't at the party this year (no time for an entry) so can't really comment how things might've been this time around.. Simply assuming that the rig or its settings haven't changed much from last year :)
Noby: ossom entry btw :)
el-bee: well, boys brought twice the sub than last year, so.. khrm.
what a party! finally back home and had a proper sleep.. see ya again next year! <3 :3
Spiikki: *oops*? :D
Spiikki: not only boys were bringing the sound system, btw, but there were girls as well at every stage and role of the work. Besides, more subs and their different positioning facilitated a more even coverage with less interference issues. While higher output helped a lot, it was more like a side effect.
At Revision we play everything as linear as possible (really only minimal eq on the master to fix the worst of the room characteristics) but we've got the infras on a second fader next to the main fader, and we know why :)
Nemppu: true that!
I don't remember that horrible experiences with the 4kB compo as with the demo compo. More than half of the entries had a lot of bass-rich beats that messed the whole sound leaving shadow sounds covering the rest of the music. EMIX had the most intense example of this phenomenon, irritating almost completely through the demo, while it sounded really pleasant on the YLE broadcast afterwards.
In total this was the worst listening experience of a demo compo I've ever had, even worse than last year with the skriik effect. I'm not sure if it's more about where the listener sat, about the guy controlling the playback mixing console or the mixing and mastering of demos. Possibly all of them. Although some four-square songs didn't get that horribly distorted, eg. Encore.
I had similar problems mixing-wise earlier in Hartwall Areena as well. While I didn't find any good mixing literature on electronic music, I had to do it with trial and error and tried to replicate what it sounded like at Hartwall areena by applying a strong hall reverb on top of the mix. After that I tried to do everything I possibly could do with the rest of the song to make it not to turn into a horrible mess and at the end removed the master hall reverb. That idea seemed to worked close enough since my songs didn't get ruined by the Areena acoustics anymore.
In total this was the worst listening experience of a demo compo I've ever had, even worse than last year with the skriik effect. I'm not sure if it's more about where the listener sat, about the guy controlling the playback mixing console or the mixing and mastering of demos. Possibly all of them. Although some four-square songs didn't get that horribly distorted, eg. Encore.
I had similar problems mixing-wise earlier in Hartwall Areena as well. While I didn't find any good mixing literature on electronic music, I had to do it with trial and error and tried to replicate what it sounded like at Hartwall areena by applying a strong hall reverb on top of the mix. After that I tried to do everything I possibly could do with the rest of the song to make it not to turn into a horrible mess and at the end removed the master hall reverb. That idea seemed to worked close enough since my songs didn't get ruined by the Areena acoustics anymore.
That said, someone should totally record an impulse reponse for all popular party places *with people sitting in them* so that musicians can just put a convolution effect on their master and check what it would sound like. ;D
Saga Musix, as if all those people would sit still silently to record the impulse sweeps :)
Saga Musix: as if all the musicians had speakers capable down to 25 Hz. ;)
Also, how about getting the compo crew involved more? One more Revision tidbit: The PC compos are all played by the compo orgas who have some weird future technology called "a sheet of paper" with the list of entries on it and helpful notes like "a lot of bass", or "WARNING: a minute of silence followed by OH MY GOD MY EARS BLEED" and basically everything else the AV crew needs to know that goes besides "channel to 0dB and cut away as soon as it gets black". You wouldn't believe how many problems this solves.
This of course requires the compo orgas to actually watch and listen to the entries in silence with ample time and good equipment; but parties the size of Revision or Assembly can easily accomodate for that.
This of course requires the compo orgas to actually watch and listen to the entries in silence with ample time and good equipment; but parties the size of Revision or Assembly can easily accomodate for that.
Quote:
This of course requires the compo orgas to actually watch and listen to the entries in silence with ample time and good equipment; but parties the size of Revision or Assembly can easily accomodate for that.
Because at Assembly we need to also capture the entries, which takes time as well as exporting and importing the video playlists, the time is actually exactly what we don't have.
We do watch the entries but as there is very limited time, some things might not be noticed early enough. We have a good audio system where we watch the entries but it's still not even nearly the same as the PA. Using actual analysis tool for checking the spectrum of the audio track (before it's played) would be a good idea imo.
My apologies to the audio guys, it's not their fault, or about tuning of the audio system.
The problem is, it is not possible for demo makers to prepare for a huge audio system like that. I think it would be better if the whole party was thought of as more like a live electronic music festival where there's going to be a lot of unpredictable self-made weirdo gadgets connected to the audio system. Who knows what garbage signal they're going to output, so you definitely don't want to let that through unfiltered and without any judgement and moderation all the way down to infra frequencies. It needs mixing to begin with.
Yle does not let audio go unfiltered and without discretion and moderation to their broadcast to viewers, so why should Assembly?
The problem is, it is not possible for demo makers to prepare for a huge audio system like that. I think it would be better if the whole party was thought of as more like a live electronic music festival where there's going to be a lot of unpredictable self-made weirdo gadgets connected to the audio system. Who knows what garbage signal they're going to output, so you definitely don't want to let that through unfiltered and without any judgement and moderation all the way down to infra frequencies. It needs mixing to begin with.
Yle does not let audio go unfiltered and without discretion and moderation to their broadcast to viewers, so why should Assembly?
Most fascinated: So many complaints about the audio, and not one that basically all the demos somehow were in mono? :-)
(I wasn't at Assembly, I can't judge how the PA was. It was fine on my headphones listening to ASMTV.)
(I wasn't at Assembly, I can't judge how the PA was. It was fine on my headphones listening to ASMTV.)
rimina: how about using a multi-band compressor/limiter?
Myself, I usually run all my mixes through an audio processing plugin called Voxengo Soniformer, and one of its "mastering" presets. If there's too much bass, it will tame it down automatically without me having to do anything.
It is just completely unreasonable to demand that demo makers provide professionally mastered mixes that work everywhere including huge sub bass systems. Not to mention small intros and oldskool machines. Assembly should be the place where demos look and sound better than anywhere else.
Talking of which, the stuff that was shown on the big screen before the Amiga demo compo looked absolutely awful, with motion horribly distorted. You spoiled Extension! BLASPHEMY!!!!! I wonder what kind of video processing is able to do that.
Myself, I usually run all my mixes through an audio processing plugin called Voxengo Soniformer, and one of its "mastering" presets. If there's too much bass, it will tame it down automatically without me having to do anything.
It is just completely unreasonable to demand that demo makers provide professionally mastered mixes that work everywhere including huge sub bass systems. Not to mention small intros and oldskool machines. Assembly should be the place where demos look and sound better than anywhere else.
Talking of which, the stuff that was shown on the big screen before the Amiga demo compo looked absolutely awful, with motion horribly distorted. You spoiled Extension! BLASPHEMY!!!!! I wonder what kind of video processing is able to do that.
Quote:
The problem is, it is not possible for demo makers to prepare for a huge audio system like that. I think it would be better if the whole party was thought of as more like a live electronic music festival where there's going to be a lot of unpredictable self-made weirdo gadgets connected to the audio system. Who knows what garbage signal they're going to output, so you definitely don't want to let that through unfiltered and without any judgement and moderation all the way down to infra frequencies. It needs mixing to begin with.
Yle does not let audio go unfiltered and without discretion and moderation to their broadcast to viewers, so why should Assembly?
The fundamental point is that it still is a demo competition. As such, it is not desirable to twiddle with entries and definitely not ok to do something that is not under any control of the contestants. The top groups put out demos that sound really good at almost any system and it isn't fair to fix some of entrys and not touch others. If you don't know how to mix your entry for large system and space then maybe you don't deserve to win.
I remember going through similar discussions several times in the past and it has always ended up with point that this is a competition, not a concert. If it were about getting most out of every entry a lot of things would be done different, now the best solution has been to not touch them in any way and playing them through a sound system that has reasonably straight frequency response. One possible idea for next year is to hand out measured response of the system, although it most likely does not make any difference. The PA has fairly straight response from ~16kHz to 25Hz and there will be some entrys that have unreasonable amount of bass and/or really loud and annoying tones at 800Hz - 4kHz.
zouwl: come on, that's just bullshit. Because it "is a" demo competition", so therefore it "must behave like a". Do you have some kind of rule book you're reading functional requirements from?
Like I said, Assembly should be the place where demos look and sound better than anywhere else. Now it was the opposite. Everyone should have stayed outside watching the stream, it would have been more enjoyable. "I saw it at the party place in the compo" is supposed to mean having experienced the demos in the most awesome and juicy way imaginable, with all possible extra goodness you can't get at home. Every party should try to achieve that sort of experience and status.
Do you know why a lot of music lovers use so-called hi-fi speakers at home, and not studio monitors? Because they want to _enjoy_ music as much as possible, not to spot as many mistakes and errors as possible.
I disagree with the whole analytical police laboratory measurement idea. You can do your crime laboratory analysis as much as you like at home, or in your mobile laboratory or something. The party place is for partying.
Like I said, Assembly should be the place where demos look and sound better than anywhere else. Now it was the opposite. Everyone should have stayed outside watching the stream, it would have been more enjoyable. "I saw it at the party place in the compo" is supposed to mean having experienced the demos in the most awesome and juicy way imaginable, with all possible extra goodness you can't get at home. Every party should try to achieve that sort of experience and status.
Do you know why a lot of music lovers use so-called hi-fi speakers at home, and not studio monitors? Because they want to _enjoy_ music as much as possible, not to spot as many mistakes and errors as possible.
I disagree with the whole analytical police laboratory measurement idea. You can do your crime laboratory analysis as much as you like at home, or in your mobile laboratory or something. The party place is for partying.
I like an unpredictable soundsystem. I'll leave the sound mastering to the professionals, though a full dynamic range is a beautiful thing.
Quote:
zouwl: come on, that's just bullshit. Because it "is a" demo competition", so therefore it "must behave like a". Do you have some kind of rule book you're reading functional requirements from?
I don't really have a personal opinion for that and it is not in my hands. If it were for me, I'd put a multiband comp for the signal and push against it if I'd feel that it would do some good. Problem is that it is not up to me and that is not fair from so many different points I can't list them all in here. If the scene wishes that the asmorg sound crew does things for entrys then I'd suggest to be in touch with Abyss and Pehu. If something like that is going to happen then it has to be properly done and informed in advance.
First major problem that comes in to mind is that after a general smoothing compression it is quite easy to mix your entry for maximum loudness and "punch" of the sound (which ear perceives as better sounding). Because the amount of compression cannot be predicted the sound coming out is pure lottery and with already good mastering it is likely to do more harm than good. Those who know what they are doing get louder (or at least perceived as louder) playback and others with uneven frequency content get forced to be more linear and are played more quietly. The strong ones are still strong, weaker ones are still weak and the whole process involves very much randomness.
The idea of available at request high-pass filter in capture is a good one, other idea is that while submitting entry you can choose between getting as-is playback or to give permission for sound techs to do things under their own discretion.
First of all, this was the best Asm in a while. Thanks to everyone for that. I especially liked the high quality of the 4k compo and the incredible scope (length) of the demo compo. It's been a while since we've had a treat like this.
Besides, discounting audio system issues, I'd guess the stomach-shattering 'brown noise' feature of noby's entry played to its appeal. Like it or not, that was something you cannot overlook or forget.
Also, Sesse, at least our 4k just was mono because that saved a couple of bytes.
Besides, discounting audio system issues, I'd guess the stomach-shattering 'brown noise' feature of noby's entry played to its appeal. Like it or not, that was something you cannot overlook or forget.
Also, Sesse, at least our 4k just was mono because that saved a couple of bytes.
I'm not talking about extreme compression or anything but some sort of a safety system, just trying to be fair and reasonable. What is not fair is to set up a sound system that goes linearly to 25 Hz, and expect 4k intros to be "mastered", let alone for a system that nobody has access to, including the compo juries. It's not fair for the contestants, it is not fair for the compo audience. It just does not make sense.
I'd much rather see compos become more like concerts than like analytical measurement laboratories. Scene people playing for scene people. What's wrong with concerts, I don't get it. Every person at a party makes a contribution to the whole experience already by being there. Every party is a unique one-time thing, you have to be there, be with the people, you live it, feel it, hear it, taste it, and it can't be repeated.
I'd much rather see compos become more like concerts than like analytical measurement laboratories. Scene people playing for scene people. What's wrong with concerts, I don't get it. Every person at a party makes a contribution to the whole experience already by being there. Every party is a unique one-time thing, you have to be there, be with the people, you live it, feel it, hear it, taste it, and it can't be repeated.
Trilkk: I agree it was a real experience! When we're old we'll be remembering old times like, "were you there when they played Noby's Hydrokinetics on the huge sound system, and the earth was shaking and people's teeth were falling off?"