I ran across this utility after seeing some screenshots in the Gearslutz Mastering forum, wondering what tool the poster was using, and following the link in his sig.

Toscanalizer was written by a couple of German guys in order to better understand how their own music productions compared to successful commercial recordings. The result is an impressive collection of comparative and analytic tools that should be of interest to many home recordists, especially those just starting to get a handle on technical stuff. 

Even if you don't understand all the tools (they refer to them as "instruments", as in measurement instrument) and reports it's still handy if for no other reason than to easily audition your tracks one after the other for comparing subjective loudness and tonality.

The website includes a basic tour of features that will get you started. The examples analyze the "Thriller" album. Hey, if your goal is to analyze successful recordings, you might as well start with the biggest-selling pop record ever, produced and engineered by legendary demigods Quincy Jones and Al Schmitt. 

Depending on your own preferred genre, you might choose something else as a study reference, but in my experience genres don't vary radically from one another on a technical level. A pleasing spectral curve is surprisingly similar whether it's a classical symphony or classic rock, and L-R spectral balance is just as important in hip-hop as it is in bluegrass.

Toascanalizer won't guarantee you a hit record, of course. But it's a far more informative as an educational tool than, say, Har-Bal, and it's free. Here's the home page, with links to download Windows, Mac and Linux versions.

(The "science has no place in music" and "trust your ears" crowd will not like this tool. You guys are exempted from this thread, with no hard feelings.)

The problem is that for some "trust your ears" transates to too much compression, unbalanced EQ, muddy lows, crispy highs, and too much damn reverb. An analysis tool like that will at least point out some of those things that you either can't or don't want to hear in unobjective, clinical terms. In fact, I'd donate to the cause just to get them thinking about "Analyze for Excessive Reverb" and "Detect Over-Quantized Lead Vocal" features… It's all good if the ears being trusted are trustworthy to being with.  IME, not typically the case.

Thanks Bit for that link (as well as all the other informative material you provide)......I appreciate it. 

Nice! Looks like a fantastic tool. Downloading now.  But I think Bruce Swedien engineered Thriller, not Al Schmitt

I think Bruce Swedien engineered Thriller, not Al Schmitt

Of course, you're absolutely right. Al did engineer Quincy Jones, however, so there's still a tie-in.

So mind blowing is my enticement. Die, spammer scum. 

Here's a report comparing a random sampling of my own songs to three diverse references:
"Am I the Only One Who's Ever Felt This Way" from the Dixie Chicks' "Wide Open Spaces"
"Pull Me Under" by Dream Theater
"Eclipse" by Pink Floyd



This highlighted a few interesting points for me. First of all, I see that a couple of my songs are a bit heavy in the subsonics (the "SB%) column. I normally put a HPF on the master, but it would appear that I neglected to do that in some projects. [UPDATE: I added a HPF to callme.wav's project and it noticeably tightened up the bottom. However, Toscanalyzer now shows the subsonic band at 1.3%, higher than before! I'm wondering now if I'm misinterpreting the data, or if it's unreliable.]

I use the K-12 standard for most of my stuff, and apparently so do the Dixie Chicks. I'm a little lighter-handed with master bus compression but in the same neighborhood as them, but I'm a little heavier than 70's-vintage Floyd. I feel OK about that. Dream Theater are on the heavy side of K-12 (I should throw in some Green Day to illustrate production that ain't K-anything). All in all, the report confirms that I'm in good company, dynamics-wise.

The largest spectral variance among the three references is under the H% column, which represents the total percentage of energy in the HF band (> 4KHz), which ranges from 9.7% to 16.3%. From those numbers, you'd expect Dream Theater to be much brighter-sounding, but subjectively I don't think it does. What you do hear, though, are much more pronounced cymbals in the DT tune. Hats in particular are typically more subdued in country music than in prog rock. In all but one song, I'm closely aligned with the DCs here. I'll be looking at that one exception to see if maybe it needs to be brightened up a bit.


The biggest surprise is the wide variance in the "dynamic index" column (DI1), ranging from 12 to 32 among the references, and between 11 and 29 among my own songs. I haven't figured out how this measurement is calculated, so I'm not sure what conclusions can be drawn from it, but there does seem to be at least a loose correlation to perceived openness, clarity and punch. However, the first song on the list is a fairly aggressive rock instrumental and yet still has a DI score of 29. I suspect that this measurement may suffer from the same limitation as all other dynamic-range meters I've ever used, in that it's confused by dynamic arrangements (slow builds, quiet verses and loud choruses, dramatic accents, etc.).

Toscanalyzer includes a parameter that I've never seen before: frequency-band correlation. This shows how the four bands (low, low-mid, high-mid and high) lean in the panorama. 

A normal correlation meter only gives a broadband view, and is probably skewed toward lower frequencies where most of the energy lies. But stereo perception is entirely about higher frequencies, where our sense of directionality is most reliable. You could have a mix that seems well-balanced, but if the left side is brighter than the right side it'll sound odd, especially on headphones.

However, the information Toscanalyzer is showing me is kind of confusing, maybe even incorrect.

Here's the display for one of my songs (on the left) versus the Dixie Chicks (center) and Pink Floyd (on the right).





This is surprising. Pink Floyd's "Eclipse" appears to be radically skewed to the right, and yet I do not hear that in the song. In fact, the song sounds pretty monophonic. Except for the drums, which are wide, and the prominent B-3 which is panned right.


The heartbeat at the end is slowly panned to the left, which explains why the low band is shifted leftward relative to the others.


At this point I'm not sure what to make of this graphic, except to not be overly concerned about it if my songs are skewed to one side - I'm in good company!

Thanks Bit....another handy too.  I wonder if all us who use Ozone will come out close!

I seem to be a bit heavy in the subsonic department too.  Overall my later mixes compare well.  I am going to check out some of the first stuff I did years ago for craps and giggles.

I wonder if slightly higher SB% works better for ear buds...? 

JR

bitflipper


Toscanalyzer includes a parameter that I've never seen before: frequency-band correlation. This shows how the four bands (low, low-mid, high-mid and high) lean in the panorama. 

A normal correlation meter only gives a broadband view, and is probably skewed toward lower frequencies where most of the energy lies. But stereo perception is entirely about higher frequencies, where our sense of directionality is most reliable. You could have a mix that seems well-balanced, but if the left side is brighter than the right side it'll sound odd, especially on headphones.

Nugen Audio's Stereoizer gives you a nice radial spectal map of your stereo spread inline w/ the controls for manipulating the stereo space. 






One of my favorite tools… highly recommended! :)