....Is it possible to use a spectrum analyser to effectivley grab a cross section of a tunes frequencies (for example your favourite tune) then use this as a guide to tweak your own track??? would this help achieve a more rounded mix???
Printable View
....Is it possible to use a spectrum analyser to effectivley grab a cross section of a tunes frequencies (for example your favourite tune) then use this as a guide to tweak your own track??? would this help achieve a more rounded mix???
yeah good method to use for comparison. should be a spectrum analyser on your sample editor.
What ya what is steinbergs Freefilter... it will analise yer reference track in 32 dif. freq. do an average of the level of each... then u'll play yer track analise again, then use the morph slider, presto!! b very carefull cause u need 2 have a track with very similar components 4 this 2 work correctly! ;)
There are other plug ins but I can't remeber the name really... I think there was one called Voxego EQ aswell.
Z
thanks guys im on it :)
What do you recommend should look proper on the spectrum analysis? This is what my tracks usually like when I run them. When I compare them to others I think sound good they more or less sound the same. Anything I am doing wrong here?
http://www.abiotic.net/images/spectrum.jpg
well past 20KHz(if not more) I would roll off those freq. and same as the lows, roll off unused freq.
Not really m8, I can't tell 4 sure just by looking at a graph, don't forget that having a big break with a heavy bass (just as an example) can make the graph look alright in terms of bass equilibrium, where in truth u might have a half dead BD...
there r 2 main schools of though :
Linear/ wall of sound... all freq. should get covered, no holes and the curve should b flat, meaning that if u try 2 find the tone 2 the all track it will come out as nothing.... this is the older way of going about it!
Exciting/ V shape curve ... this one assumes we pay much more attention 2 the sounds in the center so we need 2 bring out the extrems 2 make the final product sound sweeter even if in some way it won't sound as loud (as the other one... exactly bcause we are more sensitive 2 mid-range)! .... Hip Hop is generally a good example!
I try 2 get somewhere in between, with maybe a litlle heavuer on the bass range then the high... a slight dip in the mid range....
from an uneducated guess I would say 2 b carefull case u might b a little bass heavy (especially in the sub-bass area u have 2 b carefull so that any system won't struggle 2 play the tracks) there with maybe a tad 2 much mid-range....
But this depends loads on the program material... this might b perfect 4 certain tracks and it might me total cra* 4 others!
Z
words of wisdom from Ze MigL!! respect m8 :)
hey dustin what are you using to take that analysis?
hey dustin what are you using to take that analysis?
hey dustin what are you using to take that analysis?
Sound Forge 6
Sound Forge 6
Sound Forge 6
:)
doh!
doh!
what the ****??
what the ****??
One fact is in my mind while reading the posts above. Man has to have exact file format as a reference track. Decoded file with the same bitrate and all those things..
Yep Make sure Bit rate, Sampling rate Match... nothing that has been mp3ed is acceptable if u r seriously goinfg about yer mixes! ;)
Z
I know. But tell me how you'll get wav of your fav tune,., I'd like to know,. do you mean grab from EP?
Can anyone recommend a decent VST Equalizer (with quite a few points of EQ, if its parametric) that also has a Spectrum Analyzer with it??? (never understood why there aren't more of these around)
All u need is the waves Transform package... u get analisis tools plug the brillian (and processor hungry) Linear phase EQs.
Yer comment is absolutly right, U can't go crazy with the morphing eq thing. I tryed 2 make noticed that, just like u said, u need 2 choose tracks that have a very similar program material 2 yer own...
If u have a powercore or a UAD-1 Highly recomended u have the most amazing Sony eq!! :clap:
Z
http://www.soundonsound.com/sos/mar0.../harbal1.l.jpg
Harbal
PC Visual Mastering Software
Published in SOS March 2004
Reviews : Software
Harbal enables you to shape the EQ of your mastered mixes to match any reference track, and by compensating for the loudness changes involved in doing so, it provides a truly neutral means of comparing the original and the treated master.
Martin Walker
I first mentioned Harbal back in PC Notes November 2003, having been intrigued by the demo version of its 'visual mastering system'. Harbal is short for 'harmonic balancing', and is developer Paavo Jumppanen's superior take on 'EQ ripping', the technique of matching the frequency response of the target audio file to that of a reference, to hopefully turn an ailing track into one that sounds much more like a world-class commercial product. Of course there's a lot more to creating a perfect track than EQ, but it's still important.
Currently available only as a stand-alone PC application that runs under Windows 95, 98, ME, NT, 2000 and XP, and consisting of a spectrum analysis engine and an 8192-point linear-phase digital filter, Harbal is a considerable improvement over other similar products I've reviewed in the past including Steinberg's Freefilter and Voxengo's CurveEQ, because it incorporates A-weighted compensation for perceived loudness changes.
The human ear is most responsive to frequencies around 4kHz, and either side of this, it is less sensitive. Its frequency response also varies with level, according to the well-known Fletcher-Munson loudness contours. This is the reason why budget hi-fis have a 'Loudness' control — to boost both bass and treble extremes for more 'balanced' listening at low volume levels. It also means that whenever you change a hardware or software EQ setting, the loudness of your track will change, and herein lies the problem. For instance, if you apply some boost and think it improves your track, how do you know whether this is due to the change in EQ, or the increased overall loudness? And does a cut at one frequency work because it results in more clarity, or because the overall level has dropped and made the playback sound less bass-heavy? Harbal attempts to solve this problem by automatically compensating for the perceived changes in loudness that occur when you EQ a track, enabling more objective A/B comparisons.
Harbal Tea
Harbal consists of a resizeable window containing a spectrum analysis display with a toolbar above and status bar beneath. It supports WAV, AIFF, NeXT/Sun AU, Amiga IFF/SVX, Sphere NIST, Raw PCM, Paris PAF, ULAW and ALAW audio files, in formats up to 32-bit/96kHz.
When you first open an audio file, Harbal analyses both its average and peak spectrum content, and measures the A-weighted sound power for the original recording. The main spectrum display then plots three traces for peak, average, and the mean between the two responses in yellow, green and red respectively. I found this detailed static analysis of the whole track picked up anomalies on some mixes, particularly at the low end below 50Hz, that I simply hadn't noticed using more conventional spectral analysis plug-ins.
To alter the EQ response you select the gain cursor from the toolbar or Cursor menu, and then click on the desired frequency on the average response curve, whereupon it changes into a triple-superimposed crosshair. Up/down mouse movements separate one cursor to indicate ± gain changes, while the other two remain on the original horizontal base line, and left/right mouse movements control their spacing, and hence the bandwidth or 'Q'. Once the gain cursor is active you can also use the Tab key to jump to the Peak or Mean curves; sometimes a momentary harshness may show up more easily on the peak curve and be more easily controlled from here, although there's only one filter controlled by the three curves. You can zoom in or out of both axes on the spectrum display for more detailed work, and there's also a 'pan cursor' in place of scroll bars to let you move around this zoomed area, a Zoom All function to reset the display, and multi-level Undo/Redo functions to help you retrace your steps if necessary.
Other useful features include a log-scale saturation/overload indicator whose length indicates the percentage file-length duration of any overload, plus Average and Peak loudness figures for the entire file — useful when tweaking track levels to suit an album — plus a drop-down gain box that provides up to ±15dB of overall filter gain in 0.5dB steps if you want to lift your track to more closely reach 0dBFS, or to fit it better within the context of an album.
To hear the results there's a Play button with associated track position slider so you can commence at any point in the file, plus an EQ filter toggle button that switches between the original and the EQ'ed versions, with equalised overall loudness. This works extremely well, exposing the changes in spectral content beautifully.
This is a very flexible and easy-to-use system, and with a little practice I found it easy to call up every change from the narrowest of notches to the gentlest of broad lifts. However many of these individual 'parametric filters' you add to your curve, overall sound quality isn't degraded, since only the final EQ changes are used to design the overall linear-phase filter response.
Filter Tips
As of version 1.02, you hear all the changes in 'real time' during playback as you tweak the spectral response, and the filter itself sounded fairly transparent to my ears. The only frustrating aspect for me was a disconcerting delay after an Undo/Redo or EQ on/off comparison before you hear the audio change. The developers confirmed that this was mainly due to filter recalculations rather than the latency of the MME drivers (there's a 8192-sample delay for the 8192-point filter), but it does make subtle A/B comparisons far more difficult.
At any time you can save the EQ'ed version of your audio by clicking on the Record button — which, rather than overwriting the original, sensibly produces a new file with an '_eq' appended to the file name — or using the EQ Output menu option with your choice of new file name. You can also save the current filter response for future reference, which might be a useful approach when pre-mastering an album, since there's a good chance that the anomalies on further tracks wil benefit from a similar starting point, always assuming that they are of a similar genre. If you want to start afresh, you can at any time use the New Filter option to revert to the original sound of your track.
The Reference file is how Harbal provides its 'EQ ripping' feature, but unlike Steinberg's Freefilter, which calculates the difference between the reference and target files and then applies it automatically, Harbal instead overlays the response of the target file as a second set of three spectral responses in paler colours, and leaves you to drag your curves to match as you will. (If you find the sheer number of on-screen curves confusing you can use the Options menu to change their colours to that of the current background and make them temporarily disappear, although a menu option to selectively display the average, peak and mean curves at will would be more useful.)
Although more time-consuming, to my ears this is a much better approach, since Freefilter often tried to force huge amounts of boost at frequency extremes where your recording lacked any significant level, resulting in distortion. Moreover, during my many tests using Harbal I found that differences in key, instrumentation, or even studio acoustics could all result in significant spectral differences at some points in the response that don't necessarily need eradicating when compared to a good reference. The art in pre-mastering EQ is in recognising and dealing with any main problem areas caused by poor acoustics in the original recordings, plus badly chosen mix EQ settings, but not necessarily by trying to eradicate narrow spectral peaks or troughs that may simply be caused by individual notes in the mix, along with their harmonics and resonances. Even with a tool like Harbal, you still need experience and a pair of good ears to get the best results.
Analyse This...
As well as the ability to load in your choice of audio files as a Reference (with a Harbal filter response applied if you like), nine Analysis files for differing genres of music are also supplied. These have apparently been compiled from hundreds of songs to give you a useful set of starting points, and may instantly explain why your tracks don't sound 'right'. They comprise classical, contemporary, folk, hip-hop, jazz blues, jazz fusion, rock, rhythm & blues and techno. Judging by the wide selection of high-quality commercial tracks I loaded in for comparison, these analyses have been carefully created, and should prove very useful, particularly if you're ever asked to produce music in an unfamiliar style.
Using Harbal it's also possible to enhance a track in ways that would be far more difficult to do otherwise. You can, for instance, compensate somewhat for a non-perfect monitoring system — even if you can't hear what's going on at the bottom end when using small nearfield monitors, you can have a stab at matching the Reference response of a track in a similar genre that was recorded in a world-class recording studio.
Harbal's help file is comprehensive and readable, with plenty of useful tutorial-style content to help the beginner get the most out of it, and the developers are also extremely active on the Harbal forum, where you'll find plenty of extra advice, hints, and tips.
Final Analysis
The ideal place for Harbal in the pre-mastering chain seems to be after any multi-band compression, but before any final loudness maximising and dithering, so it's unfortunate that it's currently available only as a stand-alone application. However, apparently the developers' first priority by early 2004 is to add plug-in support to Harbal so you can include your own choice of compressor or limiter. This should allow Harbal to be used as a one-stop software pre-mastering solution in a similar way to IK Multimedia's T-Racks or Izotope's Ozone. A plug-in version of Harbal itself may appear in six months or more, but while this will endear it to Wavelab and Sound Forge users, it may have to be slightly cut down in terms of features to work as a plug-in. Mac and Linux versions have also been promised for some point in the future.
However, an SOS review is always based on what's available now, and even at the current Windows version 1.02 I found Harbal an altogether more serious application than any other EQ ripping utility I've reviewed to date. Harbal's main window can be resized all the way up to full-screen for detailed work, and the spectral responses zoomed to work in even closer detail if required, its Analysis curve options can be taken far more seriously than those of its rivals, and its graphic interface is infinitely more usable. You need to work with it to gain experience of its many aspects, but I found the results highly informative, and well up to professional standards. Anyone who uses a PC as part of the pre-mastering process should download the 8-bit only demo version and see (or rather hear) for themselves.
Harbal $95
pros
Loudness compensation makes EQ choices far easier.
Zoomable main display can show up many mix anomalies.
Easy-to-use parametric filter cursor.
Sounds good enough to use on commercial tracks.
Well-chosen Analysis files in different genres supplied for reference.
cons
Delay after using Undo/Redo or EQ on/off function makes A/B comparisons trickier.
Graphic interface won't win any awards.
summary
Obviously designed by a passionate engineer who knows what he's doing, Harbal is an excellent tool for anyone from beginner to professional who wants to improve their mixes.
this guy should be getting paid for all this :cool: :clap: Ze :notworthy:
;)
Nah I really should b off in bed as I got 2 b in tha shop in about 5 hs tops...
but I ain't djing this weekend so I'm in the clear, and sunday i'll have my coach potato day (after some tinkering with some organized noises ;) ).
Z