Quantcast
Are Intersample Overs really an issue? - Gearspace.com
The No.1 Website for Pro Audio
Are Intersample Overs really an issue?
Old 23rd January 2011
  #1
Gear Nut
 
🎧 10 years
Are Intersample Overs really an issue?

Seriously, more than half of the commercial CDs that I look at with the SSL X-ISM meter have them. If they were such an issue, then why wouldn't most engineers just lower the ceiling a little?
Old 23rd January 2011
  #2
Lives for gear
 
wado1942's Avatar
 
🎧 10 years
A single oversample may in of itself not be such a bad thing but most converters don't have any headroom. One sample hitting zero can cause many samples before and after that one to overload due to the IIR filters used to turn the information into waveforms. This is true of many ADCs as well.

I think it's safe to say that nowadays, typical commercial releases are perfect examples of what NOT to do. The concept of quality sound has been abandoned 15 years ago and sheer loudness has become not the #1 priority, but THE ONLY priority. Which is why it's all pretty well unlistenable.
Old 23rd January 2011 | Show parent
  #3
Lives for gear
 
Cellotron's Avatar
 
Verified Member
3 Reviews written
🎧 15 years
ime it really depends on not whether there are any intersample peak overs present (i.e. an occasional single one will not generally ever get noticed) - but if they are regular or constant. Generally they are also most noticeable in effect as a crackling when you convert to mp3 or other lossy audio compression codec but will likely not be audible when playing back directly from the uncompressed PCM.

I've found that how much headroom you need to give to completely avoid them definitely depends on what type of final limiting you did. With an intelligent limiter that rescales the peaks to avoid flat topping you can often get away with -0.1dBFs - while with heavy clipping I've found you have to pull things down to -0.5dBFs or sometimes more if you want to truly avoid their presence. In general I've found -0.3dBFs makes for a decent compromise.

Best regards,
Steve Berson
Old 23rd January 2011 | Show parent
  #4
Lives for gear
 
Mark D.'s Avatar
 
🎧 10 years
I agree that occasional individual intersample peaks aren't an issue. The more the worse. It's a bigger issue in mixing where many might be getting them on a track and plug-in level before any bus (which they're pobably getting them on too). Yes, they'll be there on future MP3 renderings, it doesn't change the rule that doing less damage wherever possible nets a product that's better able to handle future adulterations.
Old 24th January 2011 | Show parent
  #5
Gear Addict
 
🎧 10 years
Quote:
Originally Posted by wado1942 ➡️

I think it's safe to say that nowadays, typical commercial releases are perfect examples of what NOT to do. The concept of quality sound has been abandoned 15 years ago and sheer loudness has become not the #1 priority, but THE ONLY priority. Which is why it's all pretty well unlistenable.
What I like is how you're always looking on the bright side!
Old 26th January 2011 | Show parent
  #6
Lives for gear
 
Table Of Tone's Avatar
 
Verified Member
8 Reviews written
🎧 10 years
15 years ago?
How about 50 years ago!

Some of those old 45's had WAY more crunch and were even less dynamic than most of today's releases!

Many of the old jukebox speakers didn't handle anything over 4khz so they didn't sound as apparently crackled.

Try listening to em on a linn sondek LP12!

I just wish everyone would just stop crying about ISP's N loudness when the real culprit is the overuse of lookahead limiting in general.
Old 26th January 2011 | Show parent
  #7
Lives for gear
 
🎧 10 years
Intersample peaks are a problem in exactly the same way that a glaring, flickering, gaudy neon sign in low orbit around the sun is a problem.

You don't want intersample peaks, but they are impossible to notice floating in an infinite sea of awful. Remove the clipping and a typical modern recording is upgraded to "the worst sonic train wreck you've ever heard".
Old 26th January 2011 | Show parent
  #8
Lives for gear
 
echoRausch's Avatar
 
🎧 10 years
I always try to prevent intersample clips with MD3 Brickwall set to a treshold of 0.0 dBFS as the very last part of the chain. The ceiling is set to 0.2-0.3 dBFS and X-ISM doesn't show any intersample clips anymore (as well as TC Brickwall itself).

Depending on the input the TC sometimes shows gain reductions of up to 3 dB but usually they are lower than 1 dB.

Don't think this upgrades to
Quote:
Originally Posted by Cheebs Goat ➡️
"the worst sonic train wreck you've ever heard".
Old 27th January 2011 | Show parent
  #9
Lives for gear
 
Nordenstam's Avatar
 
Verified Member
🎧 15 years
Quote:
Originally Posted by Table Of Tone ➡️
I just wish everyone would just stop crying about ISP's N loudness when the real culprit is the overuse of lookahead limiting in general.
Different ears! To me, the inharmonic crackling of clipping is typically worse than the slower pumping of limiting.

As for ISP's not being a problem. In my experience, the typical processing options that create lots of ISP is the sort of processing that lends itself poorly to translation to lesser equipment and/or lossy encoding. I hear the crackle of clipping conversion (both lossy and DAC) lots and lots of times! Exactly as much as one should expect given the frequency of extreme loudness treatment on modern disks..
Old 27th January 2011 | Show parent
  #10
Lives for gear
 
Table Of Tone's Avatar
 
Verified Member
8 Reviews written
🎧 10 years
Quote:
Originally Posted by Lupo ➡️
Different ears! To me, the inharmonic crackling of clipping is typically worse than the slower pumping of limiting.

As for ISP's not being a problem. In my experience, the typical processing options that create lots of ISP is the sort of processing that lends itself poorly to translation to lesser equipment and/or lossy encoding. I hear the crackle of clipping conversion (both lossy and DAC) lots and lots of times! Exactly as much as one should expect given the frequency of extreme loudness treatment on modern disks..
Do you really think it's just that?

How about mic's not being on -10 when they should be, or mic pre's being hit too hard when tracking, or desks being overloaded in a bad way, stuff being tracked in the digital red, etc?

The list goes on.

That's really the issue here.

Records are being recorded on the cheap and often very badly!

Hit that with lookahead limiting and those crackles will come up with many other added side effects!
Old 29th January 2011 | Show parent
  #11
Lives for gear
 
Nordenstam's Avatar
 
Verified Member
🎧 15 years
Quote:
Originally Posted by Table Of Tone ➡️
Do you really think it's just that?
Nope, but it is a distinct problem on its own. It is our problem as mastering engineers. Or not, depending on viewpoint. What's certain is that tracking and mixing is usually beyond our control.
Old 30th January 2011 | Show parent
  #12
Motown legend
 
Bob Olhsson's Avatar
 
Verified Member
🎧 15 years
They can be a huge issue with some MP3 and AAC encoders.
Old 30th January 2011 | Show parent
  #13
Gear Addict
 
TimDolbear's Avatar
 
🎧 10 years
Yes they are a big deal. A CD that has them, example the last AC/DC album from about a year ago, if you load this and listen to it, then gain change the wave file down -.4db and listen again, its a completely different experience. All of the grit and ear bleeding disappears. The music just sound 100 times better.

CDs are overcompress and volume wars and da da da... thats 2nd in my mind to intermod dist.
Old 30th January 2011 | Show parent
  #14
Gear Maniac
 
Laarsø's Avatar
 
🎧 10 years
Back to Bob ("and them")

It occurred to me that two of my CDs that should be really good examples of what _to_ do (as opposed to what _not_ to do) ought to be Loaded Back into the DAW for analysis. I happened to have a demo of soundBlade and the SSL X-ISM.component in my Library, so, I checked out what was going on with Stanley Clarke's The Bass-ic Collection (©1997 Sony ME, Inc/Epic/Nemperor; Mastered by B. Gardener) and Beck's Guero (©2005 Interscope; Mastered by R. Ludwig).

Guess which CD, the jazz or the rock, has numerous regular ISO's (inter-sample-overs which trigger the "analog" light on the SSL X-ISM meter)? That's right, the Stanley Clarke CD is toast. Check out the clap track on, "I Wanna Play For You." Ouch. Don't beat me, Daddy.

Whereas the Guero CD is totally within the box of the digital carrier. There are NO ISO's in the Ludwig CD that is 8 years newer than the Bass-ic Collection. They must have been Free-Bass-ing back in '97.

I had "I Wanna Play For You" on vinyl the year it came out. What a gorgeously rich funk/jazz jam that that album is supposed to be.


Anyway, while we're talking about loudness, for me, the War is over and we won. iTunes spins various releases for me all day without invoking the auto-mastering preferences that change the levels or change the sample rates - no need. When a quieter CD starts playing than the previous one, I either adjust the knob, or relish the change in wallpaper. Quiet is a sound, too. I think most consumers of recordings don't care about level changes between sides, either. Only coked up executives who don't even know what's what (or Ryland?) would think that clipping the code is better sounding than turning up the monitor gain.

I have extracted all my CDs with Max, using full paranoia mode and have imported them that way (uncompressed) into iTunes, which streams nullable clones of the tracks to the mastering room DAC (through the wall in the machine room). I have started deleting the CDs that have digital samples at FS for long enough to trigger the "over" light (3 consec.), but now I am thinking about deleting the CDs that are regularly relying on ISO's for bright loudness' sake. That's just wrong, dude... Sorry Mr. Big Bass.

So, it's back to Bob. And Doug. And Wally. And J. Palmaccio. And DC. And (your name here?)....




Laarsø
Old 30th January 2011 | Show parent
  #15
Lives for gear
 
echoRausch's Avatar
 
🎧 10 years
I think the problem of intersample peaks/overs is was not very common until 2003 (?). I think in 1997 it wasn't known or at least not a known issue - so it's not surprising the CD of 1997 has IS-Overs.

The first ISP Limiter was the TC 6000 back in the year 2000 I think. I don't know if it was able to limit ISPs back then or if there were some software revisions in between... but correct me if I'm wrong...
Old 30th January 2011 | Show parent
  #16
Motown legend
 
Bob Olhsson's Avatar
 
Verified Member
🎧 15 years
The first limiter to address intersample peaks was the WAVES L1's analog mode back in the '90s. It was also the first "look ahead" limiter.
Old 31st January 2011 | Show parent
  #17
Lives for gear
 
Table Of Tone's Avatar
 
Verified Member
8 Reviews written
🎧 10 years
Quote:
Originally Posted by Lupo ➡️
Nope, but it is a distinct problem on its own. It is our problem as mastering engineers. Or not, depending on viewpoint. What's certain is that tracking and mixing is usually beyond our control.
It may indeed be out of our control but we're the ones that get blamed for it every time!
Old 31st January 2011 | Show parent
  #18
Lives for gear
 
Table Of Tone's Avatar
 
Verified Member
8 Reviews written
🎧 10 years
Quote:
Originally Posted by TimDolbear ➡️
Yes they are a big deal. A CD that has them, example the last AC/DC album from about a year ago, if you load this and listen to it, then gain change the wave file down -.4db and listen again, its a completely different experience. All of the grit and ear bleeding disappears. The music just sound 100 times better.
That album is at digital zero and there is no sign of the ripple caused by SRC.
It's also a pretty loud album, especially for George Marino!

The above would indeed be the case when making MP3's or AAC's!
Old 31st January 2011 | Show parent
  #19
Lives for gear
 
echoRausch's Avatar
 
🎧 10 years
Quote:
Originally Posted by Table Of Tone ➡️
It may indeed be out of our control but we're the ones that get blamed for it every time!
I don't think it's out of our control. Who when not the ME has control about ISP's?
Old 1st February 2011 | Show parent
  #20
Lives for gear
 
Table Of Tone's Avatar
 
Verified Member
8 Reviews written
🎧 10 years
Quote:
Originally Posted by echoRausch ➡️
I don't think it's out of our control. Who when not the ME has control about ISP's?
I was referring to the recording process!
Old 1st February 2011 | Show parent
  #21
Lives for gear
 
Waltz Mastering's Avatar
 
Verified Member
1 Review written
🎧 10 years
...and mixing

YouTube - Pensado's Place #3 - Black Eyed Peas, Pitbull, T-Pain, and guest Dylan '3D' Dresdow

You start seeing the red lights around 5:20
Old 1st February 2011 | Show parent
  #22
Lives for gear
 
Table Of Tone's Avatar
 
Verified Member
8 Reviews written
🎧 10 years
Quote:
Originally Posted by Waltz Mastering ➡️
Yep!
There ya go!
Old 1st February 2011 | Show parent
  #23
Lives for gear
 
echoRausch's Avatar
 
🎧 10 years
Though I think intersample clips are an issue I don't think it's very important for mixing/recording. If you're mixing/recording in a good way you will have enough headroom so there will be no (intersample)-overs. And even if there would be some intersample overs it's way less important. If you're doing a bad job while recording/mixing you will have overs as an issue, not just intersample clips.

But when it comes to mastering it plays a role, the end product has not much headroom left and if you can avoid intersample overs it will be good. Just to secure a good translation to all kinds of playback systems.
Old 1st February 2011 | Show parent
  #24
Lives for gear
 
Alexey Lukin's Avatar
 
Verified Member
🎧 10 years
Quote:
Originally Posted by Bob Olhsson ➡️
The first limiter to address intersample peaks was the WAVES L1's analog mode back in the '90s.
Strangely, this feature has been dropped in L2 and L3.
Old 1st February 2011 | Show parent
  #25
Lives for gear
 
Table Of Tone's Avatar
 
Verified Member
8 Reviews written
🎧 10 years
Quote:
Originally Posted by Alexey Lukin ➡️
Strangely, this feature has been dropped in L2 and L3.
The only real answer is to master quieter!
Old 1st February 2011 | Show parent
  #26
Gear Nut
 
🎧 10 years
I have never had intersample overs with the L2 dropped to -.3.

At least, according to the SSL X-ISM detector.
Old 1st February 2011 | Show parent
  #27
Lives for gear
 
Table Of Tone's Avatar
 
Verified Member
8 Reviews written
🎧 10 years
Quote:
Originally Posted by Mara*** ➡️
I have never had intersample overs with the L2 dropped to -.3.

At least, according to the SSL X-ISM detector.
Never had a problem just bringing the ceiling down to -0.3 either.
Even -0.1 if you're not that loud!

No L2
Old 2nd February 2011 | Show parent
  #28
Motown legend
 
Bob Olhsson's Avatar
 
Verified Member
🎧 15 years
Quote:
Originally Posted by Alexey Lukin ➡️
Strangely, this feature has been dropped in L2 and L3.
When I asked, I recall being told it was because they'd found most people preferred the sound of just dropping the overall level a bit even if it required a touch more limiting to get the equivalent loudness.
📝 Reply

Similar Threads

Thread / Thread Starter Replies / Views Last Post
replies: 274 views: 76529
Avatar for E-Irizarry
E-Irizarry 2nd May 2021
replies: 1114 views: 128258
Avatar for wheever
wheever 16th December 2021
replies: 26 views: 4604
Avatar for robshrock
robshrock 14th October 2021
Post Reply

Welcome to the Gearspace Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…

Forum Jump
Forum Jump