Jump to content

Cristiano Sadun

  • Content Count

  • Joined

  • Last visited

Community Reputation

10 Good

About Cristiano Sadun

  • Birthday 09/11/1970

Recent Profile Visitors

288 profile views
  1. +1 The current tempo view is really cumbersome to use, I always end up adding manually tempo changes with the insert new tempo dialog and it'd be oh so much easier to simply have a controlled drawing tool (as opposite to the current mouse-as-pen which affords no precision at all)
  2. This got me curious, so I made a little test - not too much time as I'm in the middle of a big mix, but a big mix is just what we need. I took two projects: one, the real world mix I'm doing, 24 main tracks, around 25 buses, around 50 plugins (give or take, didn't count), some of them 32 bit (I love my Variety of Sound stuff) so using BitBridge, several instances of EZDrummer, SI-bass, Ample Bass P, EastWest HBS. The other is a quick recording of a guitar+vocal idea I had yesterday in a break: no plugins, only audio, no Prochannel active First I closed Cakewalk, and looked at the memory consumption. Then loaded the project, played it, faffed a bit around (nothing substantial, changing the track display from automation to none, removing an inactive plugin etc) and looked at the memory consumption again. Not much faffing for the simple project of course other than playing it. I did the whole three times. The working set and private bytes vary a little bit (there can be garbage collecting threads) and I didn't keep the projects open for long but the experiment gives some indications. On the small project there appears to be a small memory leak that can be attributed to cakewalk (since there's really no third party stuff going on).We're talking ome 7-9K on each opening. While this could of course cause problems if you leave CW always open, it's not likely to give much grief for some dozens of open/close operations. It's also possible that it's cached data and therefore would stabilize if I had run more open/close cycles. On the large project the leak is bigger, with increased memory consumption in both project open and closed state which are an order of magnitude larger. Here the culprit has to be either CW or, more likely, one or more of the plugins and synths. Some plugs load samples (EastWest) and since I just bought HBS a month ago, I noticed how it seems a little buggy if you keep opening and closign the project. Bitbridge (and Herbert's 32 bit plugins) seem rock solid. Data at https://www.dropbox.com/s/nehs7rkx230bmgq/CW%20memory%20consumption.jpg?raw=1 My conclusion at this point would be that while CW does indeed seem to have a small memory leak, the culprit of the behaviour you are experiencing is likely to be some plugin or synth you're using. Something that may have changed from previous version is that CW does more things, so overall the memory required may have increased from Sonar time, so you hit the consequences of problematic plugins more often. It could, of course, be something to do with CW's plugin loading/unloading as well - hard to say. When I have the time I will make an experiment with only 32 bit plugins (which have shown to unload correctly) and audio (no synths) to see if we can draw some more conclusions.
  3. Yes, but, it still seems a bug, as it happens randomly when moving a crossfade, i.e. hotspot H with no modifier keys (and the mouse pointer is what it should be, it changes entering the various hotspots, does it?) It seems to be more of an issue when there are many takes, more than just the usual two or three. The video refreshes more slowly (I can see the crossfade "painted") and suddenly the hole happens, often not in the active clip whose crossfade is being applied, but at the end of the clip immediately on the right. When it happens, it tends to happen consistently but I cannot trigger it at will. A typical workaround is to create a new clip by entirely overlapping the ones that create the malfunction... the problem never occurs when defining the clip - only when moving the crossfade between two existing clips. But the main danger is not noticing that it happens, as then if one saves the project suddenly there's a clip that needs to be recovered from the original audio file, which is a pain in the bottom... Also, not exactly sure when the issue began to appear, but I guess the last couple updates - I've never seen this issue before.
  4. Thank you, will do. It's odd because for me it's always been working flawlessly - other stuff with take lanes is quirky, especially when pasting in regular clips, but that specific "hole" issue is new to me..
  5. Anybody experiencing problems with the magic tool when comping? (version It's always worked flawlessly but now clip resizing seems to have acquired a bug. It happens when I have "cut" a clip with the magic tool in comping mode (which works fine). But then when I move the borders of the clip in the track lane (by floating on the crossfade and moving the mouse vertically until it becomes the "comped clip border moving" tool) and resize the clip with click and drag, CW often shrinks the clip on the right side, leaving a hole between the right side of the clip and the next clip. It doesn't happen all the time. but quite often. I have no special settings different than what I've always used (auto crossfade, haven't touched anything related to the arranger etc), nor touching any key modifier when dragging to move the clip border. If I use the "comp" tool (as opposite to the magic tool in comp mode) all seems to be fine. Anyone else seeing this?
  6. What BitFlipper writes above is important to understand: you have buffers from and to the interface, and the computer must be able to read/write from/to them fast enough. It might be that your ASIO buffers are too small, but unlikely with a new system. My PC is not super-powerful and I run a 32 samples buffer with 4ms of roundtrip latency with no issues. Also as BitFlipper says, there are a few cases of hardware components (or more often, their drivers which are not so well written) which generate lots of interrupts - resulting in high DPC count in Windows, slowing down everything else. Certain motherboard-installed network cards from Intel were notorious for this a few years back, especially in laptops, but in recent years this has been far less of a problem. Sometimes it's surprising, for example the Corsair keyboard driver I had was using an huge amount of CPU for something so trivial as keyboard scanning. So excluding that, first of all, your PC must be optimized for real-time processing. That means disabling anything power-management related (both in the BIOS and Windows). removing any bloatware processes and services (your taskbar should be as empty as possible, and as BitFlipper suggests you can use Technet's process explorer to find out if something is overusing CPU). There's an optimization guide from the fellas at Focusrite which is well put together (https://support.focusrite.com/hc/en-gb/articles/207355205-Optimising-your-PC-for-Audio-on-Windows-10). You want also to ensure the CPU cores are always on (it's an easy tweak in the Windows registry but you can use the free ParkControl utility for example at https://bitsum.com/parkcontrol/ , just turn it off once you've modified the settings). And of course a SSD beats a rotating hard drive every day of the week and twice on Sunday. Also, from the msg you show up it seems you have a motherboard sound card. That's likely gonna create I/O problems, so you want to go in Device Manager and disable all the related elements. Look also at the Windows Services that are running - some are not needed and just steal CPU and I/O resources at the wrong moment. Best of luck!
  7. You really want to preserve the essence of the mix. A way I approach this is that since almost all playback devices worth being called that reproduce the midrange, you really want to make sure that your midrange has a mini-balance that works and keeps the essential elements of the mix intact.. all the rest, is a bonus. I just made a mix where the backing vocals and the tambourines/eggshaker get much less prominent in mono, but the bass and kick keep being clearly heard even on an old Nokia phone speaker .. and that's alright because some elements are nice to have but aren't essential to the feeling of the song - whereas vox, snare, bass and kick can't be done without. So when in mono and on a heavily band-passed speaker, the song folds down nicely to its essence, and still delivers - while if you put it on a great playback system it blossoms with all the bells and whistles. There's to say that lots of it is really in the arrangement, and it's a progressively harder task to pull off (for me at least) the more complex and dense the arrangement is. Most songs that translate best are - at any given moment - stripped down to a few basic sounds (which change over time of course).
  8. Yeah.. as I see it, the whole point of a mix is to delivery what I call an "emotional payload"... the artist wants to convey a certain emotion (and possibly have the listener feel/react in a certain way) and you and the mixing/mastering engineer team wants to make sure that keeps happening over as many different playback situations as possible, from a large disco PA to a phone. And yeah the balance is the most important thing... Say if the way you make the mix has the drums high to get a lot of excitement and energy, you don't want the drums to disappear or be overwhelmed on a mono phone speaker! But the sound itself is never gonna be the same , of course
  9. I'm using a Corsair gaming keyboard and of all things, that driver crashed and started a BSOD cycle. Thanks to Windows RE, I managed to remove it and all is fine. But annoying.
  10. As Craig, I also think of the two stages as separate, and having as little as possible to do at mastering stage is always my objective, with both my own stuff and client's... but not opening an argument here: in the end of the day, what matters is what comes out of the speakers, not how you get there. There's also to say that there are (very roughly, of course ) two classes of mastering engineers - people who want to/like to/are paid for leave a sonic signature on a mix or an album, and people who don't. If you give your mix to one of the former type, you will get back something that has, well, his/her sonic signature. If you give one to the latter, you will get back your mix, as good as it can be without altering the sonics too much. I have the impression that many "names" in mastering (for how much such a thing exists) may be of the first type, but don't know as a fact ..
  11. There's nothing I can think of about mastering vinyl that doesn't fit in that category - from how many songs you could fit to choosing the sequence to fit the changing FR of the medium the more you go towards the center (or, of course, realize a compromise with the artist's conceptual vision), to obviously adapting the song to the small dynamic range. All in the name of making stuff sound as good as it could in the medium. It's not the same of course from a tooling point of view, but that was exactly my point..
  12. Mastering has always been the same, putting on any polish that's necessary to make things sound as good as they can. Which with a very good mix means almost no polish at all. Bit like photoshop - you pull the contrast a little, increase or decrease the saturation and so on, so your printed photo looks as great as it can be (which with a poor photo, will still be not so much). If the photo is truly great out of the box, you're just Instagramming it The specific ways have of course changed over time - just like once it was to choose the most appropriate paper for prints and now it's reduction of color space.. In audio, back in the time for example part of that was to make sure it could fit in vinyl's dynamic range, which nowadays that 's no longer an issue with digital releases. Still, you need another pair of ears. That's because what you can ear, you would fix in the mix. If you didn't fix it , it's because you couldn't hear it, for whatever reason. Mastering in the same room and by the same person who's done the mix is pretty pointless (well, you can always splat a limiter of course or these days calculate LUFS and move the master fader..) As a mixing engineer, you don't need to find someone with special ears: just competent, and different ears. Their monitoring (room included) must be at least as good as yours. Personally, I send my mixes to a good mastering engineer, and invariably I can't really hear much difference when they come back, which is as it should be. Exactly because if I could hear the difference, I would have fixed it in the mix. By definition, what the mastering engineer fixes is something I cannot hear, or don't know how to. But I do occasionally master for other musicians (not as a job, but as a favor, as my interests lie elsewhere) and the process is nothing complex - mastering is far more "mechanical" in many ways than mixing, even if does require more accurate playback, fresh ears and sometimes a vision for the final product (but only if the mix is "vague", so to say). Just like photoshop! Sometimes we have exchanged mixes among friends, mastering each other's - just to have fresh ears on someone else's unknown material. It can work very well, with people who know what they're doing. In these cases my chain is always the same - obviously what links I use and how much. depends from the material. A couple EQs with a minimum/linear phase option; meters, scopes and other analysis tools, a multiband compressor with similar low phase shifts at crossovers; an exciter or an harmonics generator; a stereo toolbox, mostly to mono-ize the low end; occasionally a tape machine emulation. It used to be a limiter, but these days is more often about an integrated LUFS tool and an attenuator. And of course for multiple tracks or CDs all the sequencing, breaks and metadata stuff. Of course desperate mixes may need more, but desperate mixes are best sent back with a list of things to fix.
  13. Well, just a drum kit recorded for a modern sound takes at least 10 tracks for a basic kit.. :D Often in a recording session you keep absolutely everything, you never know what the artist wants or not and what may or may not fit. "Beat it" famously has Michael knocking on a drum case. The session I did for recording bass a few nights ago added some 10 tracks just for it (we took 5 different grooves, each both DI and miked.. and of course they're all in the CW session), each of them has two or three take lanes - I duplicated tracks quickly to more easily differentiate what's what later on. And even if we have already landed the "right" groove, they'll stay in the session as it may be that at a specifc moment it may be good to have an additional gear to shift. Actually one thing that annoys me a little in CW is that when you have several tracks with many take lanes (a good singer will need three, a bad singer may need 30) the saving gets very very slow. It doesn't happen if you "just" have many tracks and busses, so it must be some overhead of the take lanes. Haven't tried with the most recent versions though. As of buses, it's not uncommon to have at least 3 or 4 reverbs to send to, plus for vocals you may have a dedicated reverb, different delays for different sections, parallel compressors, and of course different processing chains for the different sections. Often there's more buses than tracks! It's just that the amount of detail in a modern commercial production is staggering - and the better produced they are, the less of the production is visible.. but it's still there. That's the main difference with homebrew productions (and of course, bad commercial one where the "overproduction" is clearly hearable): the manic attention to detail by someone so skilled that it feels very natural. If you always go for a very vintage sound and you are the artist so you decide what to throw away immediately you can get on with smaller track counts, but it's a bit of a special situation.
  14. Also, you can simply drag the clip down on up keeping Shift+Control pressed, and if you stop outside an existing track, CW will automatically create a new one for you.
  15. A workaround, but have you tried Windows scaling? I've used it in the past with CW and it helped.
  • Create New...