Jump to content

MXStudio RC released


sunrat

Recommended Posts

Here is a brand new unofficial respin of MX Linux tweaked for optimal audio performance and containing a nice collection of audio software and plugins.

Developer is this guy called sunrat, although probably a more correct title would be "hack-togetherer of lots of other peoples' work". B) ;)

It's tweaked from a full version of the August MX snapshot, with Liquorix kernel, KXStudio repos and most of raboof's realtimeconfigquickscan script. It will run as a live CD/USB although I prefer to install it to HD/SSD for best experience. There will be an RC2 in a week or so as I already have a list of minor changes and requests for additional programs.

I still have to add some documentation to explain what it is, but you can get the .iso file from here - https://sourceforge....o-sunrat/files/

Discussion thread is here - https://forum.mxlinu...?f=100&p=458605

  • Like 4
Link to comment
Share on other sites

That's awesome sunrat!

 

I've recently added KXStudio repo to my custom Lubuntu remix for a studio pc, but I'm still learning my way around that. I'm mostly interested in audio tools, and coming from a Windows/Cubase background, I'm finding that a LOT of this audio stuff is over my head...at least for now. JACK, bridges, sinks, busses, plugins, CALFs...my head is swimming.

 

Right now, I'm just focused on learning how to record a single simple track in Ardour, and start building up layers from there, with a good clean mix. At least Ardour kinda sorta looks like Cubase and Cadence seems to have solved the JACK>PulseAudio issues with no sound, despite a visible signal. Still fumbling around...

 

It seems there a LOT of recording engineers i linux-land who play instruments. However, I'm just a guy who plays an instrument(s) who wants to record some stuff. And there IS a difference between those 2 statements. Just curious...which are you?

 

Care to provide a rundown of your MXStudio audio applications? And do you have any of YOUR recordings that you care to share with the group. Or at least me...via a PM? Linux and original music are about the only 2 interests/hobbies that still stoke the passions. I'd LOVE to hear your stuff! :D

 

EDIT: Nevermind on the rundown...I read your link and you already provided that info. My bad. The rest of my post is still applicable though! :-D

Edited by Hedon James
  • Like 2
Link to comment
Share on other sites

Thanks @HedonJames!

I'll post the full list of tweaks and extra packages next time I'm on the production system. There's quite a few things under the hood apart from what's posted in the MX thread.

Good that you got the PA>JACK bridge working in Cadence, that's easy to do but also easy to miss.

Pretty much all the recordings I have are from live performances and I probably don't have the rights to share. I can share some Cosmic Psychos with you for sure as I know them well. Hardcore Aussie pub punk rock, probably right down your alley! ;)

  • Like 3
Link to comment
Share on other sites

Probably are right down my alley! I'd say hard rock, heavy metal (without the grrrrrr vocal stylings), punk, and 70's style funk are my preferences....although I distill them all into a single category that I call "groove rock", which only means something to me. All that said, LOVE to hear some Cosmic Psychos!

 

What about you? You play instruments? Or just record others' music? IMO, the guy who does the studio recording is the invisible band member...I've heard sound engineers make crappy bands sound fantastic; and I've heard sound engineers make fantastic bands sound crappy. If there's a "weak link" in a band, it's usually fairly easy to cover or compensate, but there's no cover/compensation for a weak sound guy! JMO...

  • Like 1
Link to comment
Share on other sites

Yeah I'm a sound engineer. I mainly do monitors these days and can just hook up the Digico SD10 to a computer via MADI to USB and record multitrack to mix down at home. Just for fun. Worked at a nice studio many years ago but it closed.

I'll cut a few tracks out of a Cosmic Psychos set and make an Ardour session to upload for you. I mixed FOH for them for about 7 years. Check out their YouTube channel (NSFW) and on Wikipedia.

  • Like 1
Link to comment
Share on other sites

I'd really enjoy that! And I'd REALLY like to be putting you on speed dial. I remembered there was another home studio/recording junkie on here, but couldn't remember who. Your signature shoulda been a clue...DOH!

 

I'm finding a LOT of this tech-talk in the Ardour forums to be "over my head mumbo-jumbo" and I'm back to googling the meaning of words and phrases that are defined with other words & phrases that need to be googled. It's a rabbit hole... And stuff like above where we discussed my complete lack of sound issue, eventually solved by Cadence. You are correct, it was such a simple setting that needed tweaked....bridge JACK to PulseAudio...but I'm not exaggerating when I say it took me 3-4 nights of 2-4 hour marathon sessions of troubleshooting, forum cruising, trying different solutions that didn't work, etc...

 

The actual solution itself...took about 15 seconds of reading the GUI options and clicking. Grrrr..... Back on point...mind if I poke you on occassion for technical questions, maybe even some instruction?

 

At this point, I'm recording everything (bass, guitars, vocals) through a DI box plugged into a Lexicon Omega. I also recently sold my drum kit to purchase an Alesis DM-10 e-kit so that I could bypass the mini-project of recording acoustic drums, figuring out which mics give the best responses, and the location of each; which frequencies to scoop/attenuate; adjusting the gates for minimal bleed; recording kick, snare, toms, cymbals and overhead ambient on different tracks for level mixing; etc... Within the past 2 weeks, I've been trying to dial in a kit that I like, voices and levels, so that I can record drums directly into the Omega, onto a single track, with no interference whatsoever from my barking & howling dogs when someone knocks on the door. LOL! And while I'm currently using the Omega, it's over 10 years old now. It still works, but the light signals give false readings for clipping, signal levels, etc... So I recently bought a Focusrite Scarlett 6i6, which seems to be an impressive little tank-built brushed aluminum box that will take a beating without any wear & tear. But I'm still using the Omega because I need a base of familiarity, due to all the other changes in the home studio. New computer...replaced an old Dell Dimension 3100 single-core with a new franken-built computer with an AMD 6-core; new OS...replaced Windows XP with Linux; new software...replaced Cubase with Ardour and a bunch of KX Studio software that I don't have a clue about, yet. You get the picture...I wanted to keep ONE element familiar and I figured the recording interface and internal routing options was probably the best one to maintain, for now at least. But ultimately, once I start gaining familiarity, I'll swap out the Omega for the Scarlett and I'll pair the Omega with a virtualized VM of my former Dell Dimension machine, with Cubase projects on it. I've got a few Cubase projects on there in varying stages that I'd like to migrate from Cubase to Ardour. That should be no problem for some of my "rough track riffs", but there's a few on there that are close enough to being finished that, honestly, they just need to be finished as Cubase projects and close the book on Win/Cubase. At least, that's the plan...it'll keep me busy for all my forseeable future "hobby time". I'll bet you can relate?!

 

Back to the Linux/Ardour machine, once I get to that stage, maybe even send you an Ardour session with rough mix for you to critique and offer pointers for better sounding projects?

Link to comment
Share on other sites

I'd be most happy to help with anything I can, and would be interested to hear your stuff and attempt a mixdown of one of your Ardour sessions.

I almost exclusively use Harrison Mixbus which is based on Ardour, and use mostly Harrison plugins which are not free so sessions in progress from that would not be compatible unfortunately.

I'll sort out some raw Cosmic Psychos for you to play with as well. Not today though, head is too fuzzy after a 12 hour shift and non-stop 4 hour show which was a benefit gig for legendary Melbourne bluesman Chris Wilson who has cancer.

Here's an old video of Chris with Shane O'Mara on guitar:

http://www.youtube.com/watch?v=ksnjf5kvKUA

 

I've known Chris for over 30 years and shed a tear just watching that, and a few covert ones last night as well.

Link to comment
Share on other sites

I did listen to the Cosmic Psychos on Youtube, and you are correct...I did enjoy that! Not sure if I can say it on here, being a family-friendly forum and all...and it IS the proper name, but I especially enjoyed F*&KWIT CITY. Right up my alley! :shifty:

Link to comment
Share on other sites

  • 2 weeks later...

Hey Sunrat,

 

Starting to get a little more familiar with the Ardour interface and settings...sorta...but wanted to ask some questions. Keep in mind, as stated above, I'm not an engineer who plays an instrument; I'm just a guy who plays an instrument who wants to record and make it sound like I knew what I was doing. A sound engineer would certainly know better, but my goal is for the average listener to think it sounds professional.

 

I was tinkering with an imported mp3 track and recorded an e-kit track with it, just to see how it sounds. Was fairly pleased with the rough track run and saved the file. 1-2 weeks later, I wanted to start writing myself some notes on "how to" and opened the file to play the track and re-trace my actions. I heard a pretty serious latency that wasn't there on the day I recorded and played back. This was pretty alarming until I realized I've tinkered with so much stuff, I'm not 100% positive of the baseline I started with. But I THINK I might've imported the mp3 at a 48000 sample rate, and 1024 buffer, which resulted in a 43ms latency, according to Cadence. However, I think I changed the sample rate to 44100 for the e-kit track, to match my Cubase LE settings (trying to duplicate a previously known "good" configuration from old software for use in new software). Would this have caused my latency? If not, any thoughts on maybe what did?

 

I don't understand this "sample" rate stuff and how that affects track synchronization. I sorta understand the buffer rate, understanding that a lower buffer is less latency, until you reach the point where the CPU outstrips the buffer cache; is this correct? Right now, I have Cadence set for a 44100 sample rate, with a 512 buffer, resulting in a 21ms latency. I have no idea if this is good/bad or otherwise. Intuitively, 21/1000 of a second delay probably isn't detectable to the human ear...or is it? Can you suggest some baseline settings for a good quality recording? If it matters, I've got an AMD FX-6200 (6-core) CPU, 16GB RAM, and an older discrete AMD Radeon GPU (perhaps 6450?). I can provide specs if needed, but I think I have PLENTY of horsepower; which leads me to believe I don't have something configured correctly, or optimally.

 

I'll also note that I'm running a stock kernel 4.0.0-134-generic from ubuntu. I looked into real-time and low-latency kernels and understand why low-latency is a better choice for audio production, but I haven't switched over because I'm under the impress that 21ms latency is pretty good. I noted you are using a liquorix kernel in your MX Studio. What was your thought process on that? Why liquorix over low-latency? And should I swap out my generic kernel for low-latency or liquorix? If so, what type of performance gain(s) can I expect? Or are there other reasons?

 

Lastly, I've still got my cloned WinXP/Cubase LE image in a VirtualBox VM, hoping to finish some old Cubase recordings from within the VM. HOPING! I repeated the process of recording a test track in a Cubase project, but it was even more out-of-sync latent than the native Ardour results. This isn't surprising, as it's a VM. And I only mention it in case any of your thoughts or advice would be impacted by knowing this in advance. For now, I'd like to get a little more fluent with Ardour and KX Studio tools on Linux. We can tackle the WinXP/Cubase VM at a later date, if that's even a realistic goal. Not sure if latency in a VM is something that can even be addressed...

 

Any insight you're willing to provide is most welcome, as I feel like a "newbie" with all this audio tech jargon. Educate me? Thank you in advance!

  • Like 1
Link to comment
Share on other sites

securitybreach

I'm just a guy who plays an instrument who wants to record and make it sound like I knew what I was doing.

 

Well maybe that is your problem.... ;)

 

Hedon, I am not familiar with most of what you posted but since it's geeky, I approve ;)

Link to comment
Share on other sites

Haha! Thanks SB...I'll take it! And just for clarity, I'm not necessarily looking to know what I'm doing...I'm just looking to make it SOUND LIKE I know what I'm doing. :pirate:

Link to comment
Share on other sites

securitybreach

Haha! Thanks SB...I'll take it! And just for clarity, I'm not necessarily looking to know what I'm doing...I'm just looking to make it SOUND LIKE I know what I'm doing. :pirate:

 

I know the feeling...fake it till you make it ;)

Link to comment
Share on other sites

Haha! Thanks SB...I'll take it! And just for clarity, I'm not necessarily looking to know what I'm doing...I'm just looking to make it SOUND LIKE I know what I'm doing. :pirate:

 

I know the feeling...fake it till you make it ;)

 

Yes...you DO "get me"! :beer:

  • Like 1
Link to comment
Share on other sites

@HedonJames - you should check out this video which explains sample rates and bits in an easy to understand way.

https://xiph.org/video/vid1.shtml

 

Sample rate is how many times a sample of audio is captured per second. If you play it at the wrong sample rate its speed will be wrong.

Latency is the time it takes for your computer to process audio. For mixing or a first recording you can set it high, but if you are recording new tracks or overdubbing it makes a lot of difference. If your system can handle it, 128 or even 64 buffer should make it better. Alternatively you can use hardware monitor through your soundcard if it supports it. Ardour has an xrun indicator on it's top panel. You will get xruns if you set latency/buffers too low so keep an eye on that.

Liquorix kernel is low-latency. I chose it over realtime as I needed to use Nvidia drivers which is difficult on RT kernels. It seems to work pretty well.

  • Like 1
Link to comment
Share on other sites

securitybreach

Haha! Thanks SB...I'll take it! And just for clarity, I'm not necessarily looking to know what I'm doing...I'm just looking to make it SOUND LIKE I know what I'm doing. :pirate:

 

I know the feeling...fake it till you make it ;)

 

Yes...you DO "get me"! :beer:

 

Hehe, right

Link to comment
Share on other sites

@HedonJames - you should check out this video which explains sample rates and bits in an easy to understand way.

https://xiph.org/video/vid1.shtml

 

Sample rate is how many times a sample of audio is captured per second. If you play it at the wrong sample rate its speed will be wrong.

Latency is the time it takes for your computer to process audio. For mixing or a first recording you can set it high, but if you are recording new tracks or overdubbing it makes a lot of difference. If your system can handle it, 128 or even 64 buffer should make it better. Alternatively you can use hardware monitor through your soundcard if it supports it. Ardour has an xrun indicator on it's top panel. You will get xruns if you set latency/buffers too low so keep an eye on that.

Liquorix kernel is low-latency. I chose it over realtime as I needed to use Nvidia drivers which is difficult on RT kernels. It seems to work pretty well.

 

Interesting stuff sunrat, but a lot of that is still over my head. I think I get the basics though:

 

1 - probably doesn't matter if I use the default 48000 sample rate, or choose 44100, as human ears probably can't hear the incremental improvement. I chose 44100 as I noticed that ripped CD tracks had that sample number in their metadata. I always wondered why the 44100 wasn't a multiple of 8 or 16, as it "should" be.

 

2 - the smaller the buffer, the lower the latency; but if the buffer is too small, the CPU runs out of material to process (causing clicks, pops, crackles, etc..?); is there an "acceptable" absolute threshold of latency, such as anything less than XXms? Or is it relative to the machine/hardware?

 

3 - I DO monitor with headphones from the Lexicon Omega soundcard; this suggests my sample/buffer settings are source of latency?

 

4 - Why choose liquorix over low-latency? Are they the same thing, but just different names in different distros? FWIW, I read an article recently that stated, unequivocably, that RT is NOT a good choice for audio production, as the prioritization of CPU tasks may actually degrade performance in a multi-tasking operation, such as audio/video production. The author suggests that RT kernels is best suited to IOT and simple-operation devices where the processing hierarchy is less dynamic, more stable. Not sure if the author is correct, but in layman's terms, this made sense to me.

 

As a matter of reference, my old audio rig was a circa 2005ish Dell Dimension 3400 with single core Intel CPU at 3.4Ghz and maxed out at 4GB RAM, with the Lexicon Omega. Sample rate was 44100 and I THINK buffer was 1024, although I'm not certain what the sample rate/buffer settings were, as the Omega would check settings during initial setup and either say "settings insufficient...adjust them" (I'm paraphrasing here), or simply move on to the Cubase LE software, as it passed the config test. I can't find the settings in WinXP and it's been almost 15 years since I set it, and forgot it. But those settings produced multi-track audio that I was satisfied with.

 

Point being, I'm certain that my circa 2014ish frankenputer with AMD FX-6200 and 16GB RAM has a LOT more muscle for audio production than that old Dell. But latency suggests I don't have it configured correctly yet. And I mention that as a reference for my WinXP VM with Cubase on it...I've allocated 2 cores and 4GB RAM to that machine, and auto-capture the Lexicon Omega device. So in theory, that's an even more powerful machine than the bare metal; but in reality, the latency is worse. So I'm thinking that Virtualization layer is probably the latency culprit for the WinXP VM, although I'm holding out hope that once I get the bare metal Linux/Cadence/Ardour config dialed in, that I can duplicate those settings in Win/Cubase VM. Hoping...

Link to comment
Share on other sites

Okay, I think I have a lot of my issues resolved Sunrat! I found an article that I considered to be pretty informative, with explanations that either made sense or weren't too far above my head:

 

https://www.soundonsound.com/techniques/optimising-latency-pc-audio-interface

 

I think I already had a very general and broad-based understanding of this, but this helped provide some of the "infill" and "nuance". As I previously thought, this seems to confirm that the majority of us can't detect anything less than a 12ms latency, which suggests that my latency issue above was probably the result of importing a "click track" at 48000 sample rate, but recording at 44100 sample rate. I have a tendency to start tracks around the 20sec mark, rather than zero, so I can tinker with intros at a later stage if I'm inspired to do that. I think the different sample rates caused the start points to be off by a little. I was able to "nudge" the track to sound in sync, but it was a PITA and shouldn't have even been necessary. I imported a "click track" again, at 44100 and recorded a sample track next to it at the same rate and everything sounded fine. Shut it down, restarted, and checked again, and everything was fine. So I think I just did something stupid in the beginning; but apparently it was necessary so I could get a further understanding of bit depth, same rates, buffers, periods/buffer, and latency.

 

Without providing a play-by-play, I spent a couple hours tinkering with Cadence, Ardour, hardware, and software settings. Long story short, I've got my settings dialed in at 44100 sample rate and 128 buffer, with 3 periods/buffer and a low-latency kernel, which produce a latency of 2.3ms...for all intents & purposes, that is zero latency! However, on rare and seemingly random occasions, sometimes i get 1-2 xruns (according to Cadence), so I'm thinking of dialing the buffer up to 256, or perhaps adjusting the periods/buffer. Your thoughts?

 

And I also thought the concept of recording with the lowest physically possible latency, but then changing the buffer sizes for mix-down to create more headroom for plugins was an interesting concept. I would've never thought of that on my own. Conversely, I've never really dealt with a lot of plugins. I usually record the signal "straight in" from a direct injection (DI) box/pedal, only adding Lexicon/Pantheon reverb on the master bus. But if I'm trying to develop a "best practice" routine, perhaps I should allow for that increased skillset in the future? I'm also thinking of changing my default sample rate to 48000, as a matter of "best practices" with my system, as I've read that 48000 provides more "headroom" without a whole lot larger file sizes. And I'm assuming that 48000 sample rates will get converted to 44100 on exporting to CD/Redbook format?

 

So I think I have a lot of my questions answered...at least enough to generate some forward momentum. But referencing my prior post, I'm still wondering why you chose liquorix over low-latency? I had never heard of liquorix, but google seems to suggest it's a matter of personal preference, with everyone believing their preference is the superior one. Shades of old gnome/kde and nano/vi discussions, LOL! I installed low-latency because it's already there in the Ubuntu repos...should I be looking at liquorix? What advantage(s) does liquorix provide over low-latency?

 

Other than the intellectual enlightening you can provide to Hedon, I think I have a working setup now?! Ardour is a bit off for me, compared to Cubase, but it's similar enough that greater familiarity will come with use. Next up...trying to tackle latency issues with the Omega and Cubase, running on WinXP in a VirtualBox VM on the same computer as Linux/Ardour. If I can solve that, I've got 2 more computers to be refurbished and donated to like-minded musicians who may even want to collaborate?! :clap:

Link to comment
Share on other sites

I thought I had replied to your previous post but it seems not. Good that you've got it happening! I'll just answer a few points you brought up.

There's not much difference between 44.1k and 48k. Generally 48k is used professionally for films and live digital audio (or 96k for live these days). Just don't import tracks of the wrong rate. Ardour should warn of a mismatch. For exports you can choose any format, Ardour will even export multiple formats simultaneously.

For latency, you want as low as possible for when you are recording and monitoring the recorded track at the same time. 128 buffers is good but 256 should be acceptable. xruns can be caused by a number of things, wifi can be particularly pesky. Turn it off while recording. I generally run 2048 buffers when just mixing down.

Liquorix is a low-latency kernel configured with full pre-emption. There's probably no reason to use it over the Ubuntu kernel.

To me, all DAWs are a bit off probably because they follow the ProTools design ideas. I could never come to terms with PT and many other DAWs are the same including Ardour (and Logic and Reaper and Cubase/Nuendo etc.). Mixbus is a breath of fresh air in that it provides a proper Mixer section similar to live desks I use which has EQ, compression and other features right in the channels so you can do a nice mix unlike others which need a mass of plugins to approach that functionality.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...