I've personally not run into any "in-the-box" processing that would induce DC offsets. Normally, that would be a problem caused by a piece of outboard gear or an A/D-D/A converter or whatever that has a faulty voltage supply to an audio OP-AMP circuit, or a bad transformer in the audio signal chain, for example. I've seen a case where a mic-pre had an issue like this, and I've seen where a mixer did this on a buss, but never anything inside the box.
 
I suspect that once you've remove the offset and kept the mix/master in the box to the end, you should be fine with only the one DC offset correction. I've not had any issues beyond that. However, it doesn't hurt to re-check for it when exporting for mastering or at the mastering stage. Hope this helps!
 

brconflict
Danny, are you suggesting that because (and I've mentioned this in other threads) this isn't what pays my bills, that I'm merely a part-timer? And you're suggesting that a predominant amount of my experience and knowledge is from books? And third, I had to at least show you some element of working connections I have in the industry because you had initially discounted me (albeit not naming names) and my opinions from the start as either misleading, inexperienced, or simply incorrect. If it's your prerogative to maintain your clout or to discount others' experience, you're welcome to it. I'll stand back and just watch next time (EDIT: Meaning I will just stay out of the conversation if I'm not helping).
 
I will add this. Yes, we both agree on the same results from different angles. I don't think that's a bad thing, either. I'm just going to say that we do make money Mastering, and if our customers are happy and coming back or bringing friends, then there must be something we're doing right.

Are you kidding me with that first paragraph? I'm triple reading it just to make sure I didn't read it wrong. Ok let me breathe and read it again....
 
*smacks head* Again...are you serious? LOL! You don't have to stand back and watch. I never said or implied that to you or anyone else. Worst case scenario, if my initial post made a few people take a look at themselves, I did my job. Do you not see quite a bit of misleading information on these forums at times? Do you care about people being mislead while you spend time here? I do. I shouldn't care but I do. Sue me for being passionate and wasting way too much time in my life being influenced by the wrong people. One day, I made up my mind that if I had a little credibility, I'd use it for good and help people whenever I could.
 
For the record, I know nothing about you...I have no clue what you do for a living though I saw you talking to the Bakers about software and mentioned what you had experienced in your life with bug reporting and the release cycles etc in a company you worked for maybe? I didn't insinuate anything about you or personally attack you or your credibility, did I? You've got to be kidding me with this!
 
You flashed a few names at me, mentioned that you heard/learned things from interviews and that you learned things in books. I said there's no need to play that card with me and books and interviews have never taught anyone "how to". All this because I made a mention that some of the stuff in this thread was disturbing to me and you took it upon yourself to think it was you. *smacks head* (I'm starting to get a red mark on my forehead)
 
I never asked you to justify yourself and thought I explained myself in a rather courteous manner in my last post considering you felt the need to mention my confidence and how you wouldn't say "you don't know what you're talking about." What that means, I have no idea but I took it as you maybe WANTED to tell me I don't know what I'm talking about but chose not to this time. I'm only confident because I've had successful results and have taught others to have the same or similar results. I've never been stuck-up nor have I held back information when someone was in need. I don't come on here all pompous with all the answers while failing to lead/teach by example. Wait...you'll take that as me insinuating that's you. Forget I said that. Please stop insinuating. I think you're blowing the entire thing out of proportion.
 
There's an obvious communication problem here that can't be fixed so let's stop talking before one of us goes off. You can even have last word and reply to my post as sarcastic as you need to. If you are going to assume every little word/phrase I say is attacking you and take things to heart that were not said to you, we really have nothing more to talk about. One thing I absolutely hate...is when a person insinuates or assumes something. I spoke about a genius (or a guy that thought he was one) on a forum one time. Some other guy thought I meant him so much, he started one of the biggest firestorms of my internet life. He was so wrong, it wasn't even funny. This seems so familiar...
 
John: Thanks. :) I honestly have no ego believe it or not and could care less who's better than me or more in the know. I'll be the first to say I like to think I'm decent at what I do because I believe in me and have consistent results. But I'd never say "I'm great" or better than anyone else. I never EVER look at things that way and will send something to another engineer before I'll just take money. I'm harder on myself and more competitive with myself than I would be with any living person. I don't listen to others and compare, I listen and worry about bettering what I did last time. 
 
I didn't say a word to this man for him to assume so strongly that I was talking about him while twisting every word I've said into a "suggestive" attack/insinuation like he has done. Just because he has a complex doesn't mean he needs to push the insinuations on me. I didn't deserve this. I'm a grown man, if I have a problem with someone, I have no problems expressing my points to that person. I just made a mention that there were things that were missing in this thread and there were things that were wrong to me as this conversation evolved. Some of them, were disturbing to me. Am I really required to annihilate people by naming names and instances where I thought things went wrong? All this so that people with complexes don't have a heart-attack thinking I meant them?  LOL! *throws hands up* This is so not worth my time anymore man. :(
 
-Danny

Danny, it's not that big of an issue, I think we just need to understand each other more. In your case, I overreacted to your initial post and took it personally, I admit. On my side, if you're not addressing a specific individual in a post like that, it does tend to leave other users digging to see exactly who you could possibly refer to, and it doesn't exclude those who you might actually agree with. So, we should both be more careful in the future. If you're good with that, then I am. We can both be more intuitively constructive perhaps.
 
btw, this may be a duplicate post. I'm getting an "Unexpected error" in FireFox for this forum. So, this one I had to post using Opera. 

cparmerlee
Is there any general advice about DC offset?  I had always assumed that if I imported a WAV and removed the offsets, then it was good to go -- that no further processing could introduce offsets.  It seems that assumption is a bad one.
 
If that is the case, then is there any reason why we should NOT remove DC offsets at every step of the process, or at least make sure to do that when we do the final stereo mix?  There really isn't any loss of fidelity by removing the offsets too frequently is there?
 
Other than Process - Apply Effect - Remove DC  Offset, is there any tool in SONAR that does this?
 
Is there any tool in SONAR that tells us how much offset there is?
 
Ideally it would be great to just drop a VST on the master bus right before the limiter to eliminate that issue automatically.  Is there any such VST?
 

Not really. You remove it when you see it. As far as your theory about remove it and you're good to go...it depends. Let me set up a situation for you so you'll see it from my experience. Other people will probably chime in with theirs...I speak for me only and am not competitng with anyone nor am I trying to tell you the advice you have already received is incorrect.
 
I get a song from a client. I ask them to try to get as close to -3 dB as possible. When I import the file, sometimes it has a little DC on it...sometimes a lot sometimes none. It depends what they did on their end. It could be outboard gear, or loads of compression that makes the song appear louder than it is which can raise DC's. I can remove this DC now...which I like to do. Or, I can do it at the end. It really makes no difference sound wise, but along with creating a good master, I like to have good numbers to match all throughout the process. Of course you never rely on them or live by them, but they are important to me. Again, if Ludwig has great numbers on every mix, I need to also. :)
 
Think of it like cooking. If you don't clean your stove immediately or pick up something that dropped on the floor, you have to work twice as hard to clean it. Or imagine being in construction and you hear from the grapevine "dude does incredible work, but he's so messy, I'd not have him back again." I like to take care of things when I see them whether they come from what I do or what others have done. You're not doing anything wrong if you create a little DC offset in your material. It's when it's excessive you need to worry.
 
When I export and then do the sample rate conversion, dithering, limiting etc...the DC will ramp up to maybe .024% depending on how loud I make the mix as we approach -0dB. Limiters will pump up the DC every time depending on what your outceiling is and how much threshold you use. (I like to use an outceiling of -0.1 for waves, -0.3 for mp3's.)
 
I remove the DC and save now as a newly converted 16/44. Sometimes when I don't remove DC's as often as I normally do, the DC may get that -.001% I mentioned once it's saved to 16/44. So I may pass it through Wave Lab DC removal one more time. Some editors will show a DC when others will not. Wave Lab says no at times where Adobe or Sound Forge will show -.001%. It doesn't matter and none of this is worth you going crazy over, I promise. You'll probably never get enough DC to ruin speakers or your material. I've tried to purposely raise it just to see how bad I can get it before it causes trouble. It's something Mastering guys take care of out of principal because it shouldn't be there.
 
If you got a mix back from Ludwig, though .001% is nothing....he doesn't ever leave a % in the stuff I've analyzed. However, Katz leaves a little and doesn't have his numbers as consistent. For example, here are the stats on the most powerful tune on my album mastered by Katz.
https://www.dropbox.com/s/8zrfir2l1fs5o7h/AON.JPG (Keep in mind, an mp3 will show you different stats than a wave. This is a wave file stat and the most true.)
 
Here's the same song in Wave Lab: https://www.dropbox.com/s/rogygx8vux8c57i/AonWL.jpg
 
Neither are bad, but I'd not be happy with those numbers for myself or my clients. It does sound good so the numbers don't matter the same as watching an eq graph. Sometimes an eq graph looks terrible yet the song sounds good. Sometimes they look awesome and the song sounds horrible. LOL! To me the numbers thing is just a mark of an ME's work ethics trying his best to do good work like a builder would. You know, doing great work while keeping things neat and clean. I like to have all aspects at all times just for myself though none of it is super important other than to me. So honest, like I told G before...don't worry about them unless you are hearing problems and seeing huge numbers while suspecting that DC's may be the culprit. :) Hope this helps.
 
If you're still a bit unclear, a really good read about it with diagrams here: http://u104.guest.isc.org/pure-data/dc-offset/
 
-Danny

Danny, why -0.1 for wave and -0.3 for MP3?

When you convert to mp3 the data compression reduces the detail in the data set
 
When the detail is reduced the peaks are being described in a cruder fashion.
 
The peaks are extrapolated between discreet samples.
 
Extrapolation that did not exceed the "0"dBFS threshold with a PCM .wav file may not be recreated as accurately when the mp3 data decompresses.
 
If you leave just a little extra headroom when you prepare a mp3 you can avoid being surprised by a peak exceeding your expectations when its data is decompressed back into the PCM stream that is sent to the DtoA converter.
 
best regards,
mike
 
 
 
edit spelling

gswitz
Danny, why -0.1 for wave and -0.3 for MP3?

Mike's post above nails it perfectly. As to *why* I chose those particular numbers for an outceiling....
 
The higher your outceiling in your limiter, (meaning the bigger the number...-1, -2 -3 etc) the more limiting/compression you get when you kick up the threshold to make things louder. I find that leaving the outceiling as close to -0 dB as possible (not to be cornfused with 0 dB without the minus....as 0 dB is officially clipping) allows the material to breathe more dynamically. Realistically speaking, I can get more volume out of a higher number outceiling (-3 dB) by pumping up the threshold, BUT....that would give us square box waves and you don't want that.
 
With mp3's, everything changes due to how it encodes and compresses to that format as Mike mentioned. For example, in my other post where I show you the wave data information.....if I converted that wave file to an mp3, the stat numbers would change. You might have more DC offset, you may pick up some possibly clipped samples...it all changes slightly. This is why I had mentioned not to rely on stats you get when you analyze mp3's.
 
The reason for -0.3 if I can be totally honest? One day I had noticed that every once in a while, my mp3's would show possible clip points yet they wouldn't clip and give me red lights. I noticed that other mp3's from those involved in loudness wars, would clip like crazy. But then again, even the wave files had a few clip points but nothing that stayed in the red.
 
Anyway, we were talking about something on here and bitflipper made a mention of how he dealt with mp3's always using -0.3 as his target. As you know, he's one of the most knowledgeable guys on the forum and I respect him greatly, so you know I had to try what he had mentioned. Low and behold, I have never seen any of my mp3's show possible clips and I've used the -0.3 method on mp3's ever since thanks to Dave. (bitflipper)
 
In the mastering realm, it's rare we actually get asked to convert wave files into mp3's. So it's never bveen anything I've ever worried about. Plus, even when I didn't use the -0.3 method, I never actually HEARD clipping in my mp3's. But the past 3 years, I've had clients that have been doing mixed mode albums with me. They want a regular audio CD and then they want the songs encoded to mp3 as data on the same disc so that fans can just grab the mp3's and drop them on their iPod or whatever.
 
This is also happening with people that want me to do a CD for them as well as upload the album to iTunes. I can literally create their album and their iTunes album in one swipe with one of those (I can't recall what it is right now...lol) m3u files that when you click it, it plays the mp3 files in the exact order they were set for on the CD. You know...like a playlist.
 
So because I'm doing things like this now, it's been great to have learned the -0.3 thing from bitflipper or I would have been sending out semi-clipped mp3's.....or at least mp3's that may have been registering as clipping on some systems. Hope this answers your question a bit more in full. :)
 
-Danny

Is there anyplace inside Sonar or the normally included VSTs to see the DC offset numbers?  If not, are there any low-budget VSTs that will show this info?

Not that I know of....but Analyst has a DC monitor that at least tells you if there is a DC and if it's positive or negative. As you can see in this shot, I'm right about in the middle so there would be no need to be concerned with DC in this song at THIS point.
 
-Danny
 

Thanks Danny and Dave.