Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[djwaudio]

At 24 bit, I guess it's less of an issue since you're pretty far into the noise floor.  I think I can hear the difference with the TC 6000 when using bit pattern transparency, but I have dither on anyway, since the output is being dropped by a few tenths of a dB, and I often use it in up-sample mode.

[Jerry Tubb]

I always try to use:

2.71828

or

1.6180339

or

3.14159265

JT

[Herbeck]

I always try to use:

2.71828

or

1.6180339

or

3.14159265

JT

My condolences and commiserations...


Herbeck

[Jerry Tubb]

My condolences and commiserations...

Herbeck

Number Humor.

JT

[Herbeck]

Jerry, you posted this on Gearslutz today.

http://www.gearslutz.com/board/7888002-post91.html

I find your "sense of humor" very close to trolling and totally pointless.


Herbeck

[MoreSpaceEcho]

have you seen how OTHER people post on gearslutz?

maybe i'm missing something here, but i'm confused about this technique. once you do ANY sort of gain change, the DAW is converting the file to 32 bit float anyway, any 'noise' from rounding errors or whatnot is going to be down around -140dbfs (or wherever), and that's going to be totally swamped by the noise floor of your analog chain...so, i don't get it?

[Jerry Tubb]

Jerry, you posted this on Gearslutz today.

http://www.gearslutz.com/board/7888002-post91.html

I find your "sense of humor" very close to trolling and totally pointless.

Herbeck

Did you "get" my numbers humor?

Best wishes to you Ulf.

Later, JT

[bblackwood]

Did you "get" my numbers humor?

Best wishes to you Ulf.

Later, JT

I always pictured you being more 'jovial', less 'troll-like'...

[Jerry Tubb]

I always pictured you being more 'jovial', less 'troll-like'...

Thanks Brad! Good to see you around.

We did have plenty of good joviality on your forum a few years ago, with the core group of MEs.

And for the on-topic record, when I do gain changes in Sonic sB, I check the "Pure Gain" preference which translates to 3.01dB and 6.02dB.

Nick and I had a discussion on the topic the other day, and even looked it up in Pohlman's book.

The idea is that the digital numbers shift by one place.

I heard about the idea was from Roger Nichols (RIP) about 15-20 years ago.

Sometimes I forget that everyone doesn't already know all this stuff, so I toss a "humorous" zinger into the fray.

Best, JT

[Herbeck]

I heard about the idea was from Roger Nichols (RIP) about 15-20 years ago.

Well of course this is old news. Gregg (Babaluma) had his thread on the subject over at Gearslutz and it was so obvious that people (like yourself) held back on this information while pretending to be helpful. And that was the point that I was trying to make by this thread.

Cheers,
Herbeck

[Hermetech Mastering]

Yeah, my thread at Gearslutz was basically about the WUMP track. At the time, I couldn't access this forum, so I posted it there instead. Should have just waited and posted it here. Herbeck, thanks for moving it here and creating this new thread.

As any involved in the recent WUMP will know, there were quite a few digital overs in the source file, which was 16 bit/44.1kHz. My question was basically, "What do you do, if anything, in this situation?". It happens to me often. What I usually do is convert the file to 32 bit floating point, and use a lookahead limiter with very subtle settings in order to get the peaks just under 0dBfs. It's nothing to do with fixing the inherent problem (that would be a whole other thread talking about "peak restoration" software etc.), but more out of a habit of worrying about sending files with digital overs out of my DAC and into my analogue chain.

I had a lot of great replies both on GS and here. What it basically came down, as always, was "Listen, and can you hear a problem?", which is of course how it should be. Some of the other options were: "Just drop the level 6dB in your DAW", "Use a look ahead peak limiter", "Don't worry about it", etc. The thread in GS kind of devolved into a pissing match about ISP's, what was "the best" way, etc., as they often do! I basically need to sit down and do a load of listening tests for all of the options when I have some down time.

I was really interested to see the "Drop it by exactly -6.020599913279624 dB" answer posted here, because I had never heard of that before. Did Mr. Nichols ever write an article about this?

Thanks for all the great answers, and humour too, I appreciate it all, and just wish some of the better posters on GS would also make it over here and to DC's great forum as well!

[Jerry Tubb]

I was really interested to see the "Drop it by exactly -6.020599913279624 dB" answer posted here, because I had never heard of that before. Did Mr. Nichols ever write an article about this?

Thanks for all the great answers, and humour too, I appreciate it all, and just wish some of the better posters on GS would also make it over here and to DC's great forum as well!

Remember seeing it mentioned in Roger's column in EQ magazine back in the early 90s, didn't go into any great detail. Roger was "hanging out" here in Austin in October 97' after a Yamaha mixer seminar, with a group of local engineers, all having waaay too much fun to obsess on technical details.

Check Ken Pohlmann's book for the specifics.

Maybe Nick will comment, he's well schooled in the math of it all.

Herbeck, perhaps you're taking this all too seriously.

Cheers, JT

[Hermetech Mastering]

Thanks Jerry, will hunt down that book! Would that be his "Principles of Digital Audio"?

Took me a year of bed time reading to get through the Curtis Roads' "The Computer Music Tutorial", so hopefully it's not quite that big!

[aleatoric]

Interesting thread.  I just did some tests in Reaper.  It does appear that a gain reduction of -6.020599913279624 dB keeps the "bit order" in tact.  Using the Stillwell Audio Bitter bitscope plugin it seems that it creates a 17 bit signal.  If I'm understanding things correctly the top bit would be the 6.020599913279624 dB of silence and the rest of the bits simply shift down.  Interesting that a gain change of just -6.02 dB does not have the same effect.  In my test -6.0205999 dB was the closest rounded number that worked.   

I have yet to do any real listening tests but from a hypothetical mathematical perspective it does seem that this may be a better way to make a gain change prior to hitting your analog chains DAC. 

I just emailed FabFilter about a very simple stereo and M/S utility plugin I'd like to see and asked for this to be a feature.  Essentially I'd like a plugin that can do all the common things you'd find in a full featured analog mastering console in one plugin.  L solo, R solo, independent L and R (or linked) +/- gain adjustments with a "bit transparent" option to move in -6.020599913279624 steps, L/R channel swap, L and R phase invert, mono button, M solo, S solo and a width knob/slider (gaining up or down "S").  I can accomplish all of those things currently but need to use a combination of like 2 or 3 plugins to do all of them.  It's a workflow thing more than anything else.  I think it'd be a cool plugin to have on your pitch track. 

Alright, enough geekin' out for me today. 

[Jerry Tubb]

Thanks Jerry, will hunt down that book! Would that be his "Principles of Digital Audio"?

Took me a year of bed time reading to get through the Curtis Roads' "The Computer Music Tutorial", so hopefully it's not quite that big!

Yes Gregg, it's its huge, & packed with information, almost a digital audio bible.

I've highlighted many sections over the years.

Continue to go back to it as a resource.

Enjoy, best regards, JT

p.s. I must add a comment, that when I'm making level changes in digital audio, I admire the purity of going "by the numbers", but the ear, as always,  is really the judge!