Har-Bal. Nonsense or what?

[bob katz]

This is the second time I've gotten a message at our website to the following effect. Here is the message and my reply. Discussions, please!

Viewing Issue #2701 Issue Type Guest Book Submitted By Anonymous Time Submitted 07/14/2006 (11:41:02am) Status open User-Supplied Information Name D.G. Email Subject Har-Bal Url Comments Bob Katz, what's your take Har-Bal? Can you create some proper .anl or .mfl files for the different types of music genres please? You can put them on www.digido.com for dowload.


Dear Doug: I'm just not a Har-Bal type of guy. I listen, I make adjustments and I get results. I do not believe in "genre eq's" and trying to make YOUR PIECE OF ROCK match Green Day or Radiohead by simply using a convolution approach is not only wrong philosophically but also doesn't work. There's far more than just EQ and ultimately the sound you have to get has to be done by listening to the source music, getting its feel, and determining if this music needs to be "pushed or pulled or altered" at all into any other direction. There's no software that can or SHOULD do that, in my opinion. Sorry.

Someday I'll play with the product just to see if it has any use in some particular circumstances but in the past I've been underwhelmed by the sound quality of every manufacturer's convolution eqs (including a very particular famous piece of expensive hardware that includes eqs, compressors, etc.), emulators, and devices that purport to match certain curves by measurement. In the end it's not close enough for me.

However, once I ran into a situation where the identical musical material had been recorded through two different media with radically different eqs and I had to splice in a section from one to replace a dropout in the other, in a major release. It was quite a bit of work to match the EQ manually.

Remember, this was IDENTICAL material recorded two different ways that I had to match. One way was a CD made 25 years ago from the analog tape when it was new but very poorly EQed, and the other was my own modern-day transfer from the same analog tape but 25 years older. I could see Har-Bal as a useful aid for the types of curves and EQs that would be necessary to EXACTLY match this EXACT source material which had been recorded through two different media. I might not use Har-Bal at the end, but rather take notes on its EQ and put it through a different equalizer which might have more resolution, or perhaps I'd be pleased with the Har-Bal result and use that, or tweak it manually to take it the rest of the way. The ear is the ultimate judge.


Bob

[cerberus]

i think there are three issues in this post: "matching eq", transferring settings from one device to another, and convolution.

i don't think i understand filter design well enough to know exactly why matching eq probably will never work, but i feel that the claims made by the makers of these devices have always been unfounded.

also it's obvious to me that a particular setting on a pultec will not transfer well to a weiss eq...

but convolution has become an essential tool for me in mastering. i do not pretend i own analog devices, but i do have some facsimiles of some of their dna. this is better than nothing. it is also stuff that cannot be done in the analog domain...if digital didn't have such attributes, i never would have embraced it.

perhaps we can draw a comment someone who owns the ten thousand dollar dynamic convolution device you referred to? i know k.k. has one and loves it. i hope that if you breach the subject in your second edition, that you balance your opinion with one that considers a relatively underdeveloped class of tools that are not 100% perfect, but are still more useful than a blank panel...and impossible in the analog domain.

convolution is a literal miracle to me. i am so surprised that you feel otherwise. i agree that it isn't elegant to run millions of instructions per second to try and emulate a device that has less than ten active components...but ya gotta start somewhere.

jeff dinces

[bob katz]

Originally Posted by cerberus

perhaps we can draw a comment someone who owns the ten thousand dollar dynamic convolution device you referred to? i know k.k. has one and loves it. i hope that if you breach the subject in your second edition, that you balance your opinion with one that considers a relatively underdeveloped class of tools that are not 100% perfect, but are still more useful than a blank panel...and impossible in the analog domain.

Never fear, I always temper opinions when it goes into the book.

My current objections to the state of the art of this particular class of tools (of which only one major brand comes to mind) are:

1) The compressor implementations tend to be dynamically rather sluggish and to my ears do not adequately simulate the action of the real analog compressor at all input levels. It's currently in the primitive kind of state that multisamplers were before they started to realize that you could not sample just one note. In the case of convolution of a complex device like a compressor, you have to very complexly sample each dynamic level, and this requires far more processing power than perhaps even 4 sharcs can do well.

2) Digital compressors should be oversampled or at least run at 88.2 kHz to reduce harsh intermodulation artifacts (inharmonic distortion). I hear and have measured these harsh artifacts from the box in question. It just doesn't sound good to me.

However, I've heard some of their simulations of more linear devices such as equalizers and I'm pretty pleased.

In short, I think we're a long way from the realization of this "promise", despite the hype. It's still too "liquid" to me! I prefer a good simulation to a convolution, right now.

However (there's always a however) convolution-based reverbs are doing pretty well right now. In my opinion, only the VERY best artificial simulators now beat the convolution units in the area of depth and early reflections. But I hear in one of the major brands of plugin reverbs a truncation of the ultimate tail... you have to apply a lot more CPU power to get that low level tail down to past -60 dB and at least one manufacturer has decided to compromise so they won't eat up that much CPU power.

BK

[UnderTow]

Not much to discuss really as I agree with your view on stuff like Har-Bal. To me it is a gimmick. The one good use I can see is as an ear training device. It can teach new-comers to find what the EQ/frequency differences are between their material and whatever reference material they are matching. It can work as a guide for people but that is about it. (Except for special circomstances as you describe).

My only comment is that IMO Voxengo CurveEQ is a much better tool than Har-Bal.

Alistair

[cerberus]

thanks bk! i didn't realize that the state of the art dynamic convolution was so primative, people are probably still amazed that it comes close, perhaps accounting for some of the glowing reviews i've read. but if it were not sampling frequently enough, i wouldn't want to use it either. another company has announced a product using sharc, i've noticed that the chip might not always deliver the power its reputation implies. i suppose i should put my "dream" for this technology on the shelf for a few years and revisit it then...

i am somewhat astounded with my own results using static convolution [q-clone] on every mastering job for the past year or so. i begain to get very frustrated with my eq...what had been said about itb eq became true for me too. so i needed to find some way to do eq without...analog, because i can't afford it. i've restricted the type of i.r.s i use to what i'd like to call "all pass filters" but i hardly understand what an all pass filter really is. for now i call them "transfer functions"[see attached image]...it seems a safe description, but a rather bland one considering their effect. i have more control over nuance of "color" in mastering this way than i ever did before. it's like a drug i want more of. when i learned on this thread that har-bal uses convolution, it made me want to like it!

on the reverb tail truncation issue (only used once so far in mastering , but used tons of it on mixes): yes, it's a shame...

===

i think what alistair is getting at is that har-bal makes lots of eye candy; so for the money perhaps it's not a bad analyser.
[attached image is not from har-bal]

jeff dinces

[Alécio Costa]

I think it all started circa 2003. Their messages entered Brad´s PSW room , RO and GS forums.
Imagine all those top mstering engineers receiving these "news"

[UnderTow]

Originally Posted by cerberus

thanks bk! i didn't realize that the state of the art dynamic convolution was so primative, people are probably still amazed that it comes close, perhaps accounting for some of the glowing reviews i've read. but if it were not sampling frequently enough, i wouldn't want to use it either.

My understanding of the Sintefex, Liquid Channel and Liquid Mix dynamic convolution process is that it takes 128 (or less) samplings of the tone and non-linearities of a device at the different levels but it doesn't actually sample the compression characteristics of the device. The compressor in these machines is a very normal digital compressor (modeled with traditional methods) with the added tone of the device for any particular level of gain reduction.

I got quite a bit of criticism last time I wrote this on another forum. Basicly people were telling me that I didn't understand dynamic convolution. These included people that have reviewed it for well known mags. After I quoted stuff from the Sintefex manual people stopped responding so I guess I was right. If anyone understands the process better than me, please correct me.

Alistair

[UnderTow]

Originally Posted by cerberus

when i learned on this thread that har-bal uses convolution, it made me want to like it!

Oh I missed this bit when I first read your post. As far as I know, Har-Bal doesn't use convolution. It just creates all these EQ points to match the frequency spectrum of one file to another. (Or uses saved sets of EQ points when you choose the "styles").

It might be different in Har-Bal 2 but I lost interest after playing with version 1 for about 15 minutes so I havn't checked.

i think what alistair is getting at is that har-bal makes lots of eye candy; so for the money perhaps it's not a bad analyser.
[attached image is not from har-bal]

jeff dinces

Lol. Not really. I think CurveEQ is a better tool with a better FFT analyzer than Har-Bal (1) and it also has convolution[1] built in (You can match the tone of a list of gear) and analogue saturation and some other niceties. More importantly, it is a powerfull, very flexible spline EQ in it's own right reguardless of any spectrum matching. And it is VST unlike Har-Bal (1) which is standalone.

Alistair

PS: Those transfer function graphs look very similar to the gear-matching convolution[1] that CurveEQ does.

[1] I'm quite sure it uses convolution but I would have to check the manual to be 100% sure.

[cerberus]

spline eq? wtf?

Originally Posted by UnderTow

PS: Those transfer function graphs look very similar to the gear-matching convolution[1] that CurveEQ does..

it's comforting to know i'm not the only engineer on earth using the strategy in mastering...


jeff dinces

[UnderTow]

Originally Posted by cerberus

spline eq? wtf?!

jeff dinces

It doesn't have set bands like most EQs. You just double click somewhere on the graph and it adds an EQ point. You can add/remove/drag/etc these EQ points to wherever you want. The curves between the EQ points are calculated by spline maths. A spline EQ.

I forgot to mention that it is linear phase or mininal phase and also works in M/S mode. Now I'm sounding like a marketing guy so just check out this page and/or download the demo: http://www.voxengo.com/product/curveeq/

Alistair

[cerberus]

spline maths are well suited for depicting the curves of the typical filters that are used in eq design. unless this eq can make filter shapes that couldn't be done by combining conventional filters, there would be no reason to assume its algorithm does not employ conventional filters like the other eqs that use splines to draw the u.i. (e.g. metric halo channelstrip). i see the uniqueness in being able to run as many individual filters as the user desires..most digital eqs have a fixed number of bands. but the use of spline maths in digital eq is not unique; not re-inventing the wheel like the name would imply... very good marketing name; but i suspect it is actually a "parametric eq".

jeff dinces

[not_so_new]

Originally Posted by UnderTow

Not much to discuss really as I agree with your view on stuff like Har-Bal. To me it is a gimmick. The one good use I can see is as an ear training device. It can teach new-comers to find what the EQ/frequency differences are between their material and whatever reference material they are matching. It can work as a guide for people but that is about it. (Except for special circomstances as you describe).

My only comment is that IMO Voxengo CurveEQ is a much better tool than Har-Bal.

Alistair

As a non-ME on this thread I would like to say that I agree with this, it would be a great ear training tool but don't think I would want my ME using it on my projects.

BUT

I would add that Har-Bal would be a great great way to see the deficiencies in your room pretty quickly. I mean right now I use my ears when I am in a new room and that has been working for me for years but to have a little backup from what I am hearing and what something like Har-Bal is seeing might not be a bad thing… right?

I have been thinking of picking up a copy of CurveEQ for just this reason (I have not researched it much so maybe this is not the right tool for me??). Put it on a mix when I am almost finished with it and see how out of wack my speakers / ears / mix is (hopefully not very) then make some final adjustments with a program EQ or start looking at my channel EQ's for some help… then take CurveEQ out of the path and print. *shrug* may work may not but I am interested in the idea.

[RainbowStorm]

I have not used Har-Bal, I've read about it though. In my opinion you can't create effects that accurately emulate deep mastering processes. First of all, such processes need to be 100% dynamic on the audio material, so you can't really expect a static EQ+comp preset to do anything in particular with the audio material, it all depends on the input. So generally, these kind of effects are extremely complex. While it's possible to create effects like these I think there is still a huge gap left between the real very detailed mastering job and the implementation accuracy of the digital emulation processes. I think there is a deep misunderstanding going on in the recording world, that is that there is a certain magic formula for beautiful sound. Engineers target things systematically and expect that to be the ultimate way of dealing with audio most efficiently, when it's the other way around. So a unit like Har-Bal can't successfully reproduce the result of a good mastering environment, good mastering ears and good decision making, it's a too complex real-world process. You can try to create signature settings, but these will perform very poorly. Bob, I totally agree with your post.

[Sammy]

Let me chime in as a casual Har-Bal user (it's been a while though..). I find it extremely useful as a sanity-check. The provided reference curves are pretty useful.
Har-Bal doesn't match eq to a preset curve as far as I know: it only gives the actual curve and the reference curve visually, and you can modulate the actual curve to get it closer to the ref. It's still a man-driven process though. Maybe the newest version has an automatic matching function, I don't know.

Anyway, the EQ filters sound more than decent on both ends of the spectrum, but not so good in the midrange.

Indeed, it can also be used as a nice visualizer for your room-response: just record 30 secs of noise through your monitors and let Har-Bal draw the curve. Pretty useful IMO (though not a perfect approach ofcourse).

[UnderTow]

Originally Posted by not_so_new

I would add that Har-Bal would be a great great way to see the deficiencies in your room pretty quickly. I mean right now I use my ears when I am in a new room and that has been working for me for years but to have a little backup from what I am hearing and what something like Har-Bal is seeing might not be a bad thing… right?

I'm not quite sure how you intend to use spectrum matching to find the room deficiencies. Care to explain?

I have been thinking of picking up a copy of CurveEQ for just this reason (I have not researched it much so maybe this is not the right tool for me??).

I don't know. You can download the demo and try it.

Put it on a mix when I am almost finished with it and see how out of wack my speakers / ears / mix is (hopefully not very) then make some final adjustments with a program EQ or start looking at my channel EQ's for some help… then take CurveEQ out of the path and print. *shrug* may work may not but I am interested in the idea.

What reference would you use to match to? And how would you judge the room response based on the resulting EQ curve?


Alistair

[UnderTow]

Originally Posted by Sammy

Let me chime in as a casual Har-Bal user (it's been a while though..). I find it extremely useful as a sanity-check. The provided reference curves are pretty useful.
Har-Bal doesn't match eq to a preset curve as far as I know: it only gives the actual curve and the reference curve visually, and you can modulate the actual curve to get it closer to the ref. It's still a man-driven process though. Maybe the newest version has an automatic matching function, I don't know.

From memory, Har-Bal does match the curve. Then you can tweak it to better fit the actual material.

Anyway, the EQ filters sound more than decent on both ends of the spectrum, but not so good in the midrange.

Indeed, it can also be used as a nice visualizer for your room-response: just record 30 secs of noise through your monitors and let Har-Bal draw the curve. Pretty useful IMO (though not a perfect approach ofcourse).

I think the most relevant test for room response is a waterfall plot. A matching frequency plot (or a plot that shows the deviation from the original noise source) doesn't tell you anything about the time-domain. Of course this is the problem inherent with any EQ "solution" to room problems.

I think doing quick bursts of pink noise into the room and watching a FFT analyzer realtime to see which frequencies stay the longest (the poor man's waterfall plot ) give more info than watching any static frequency plot.

But I am no accoustician so I could be wrong. Anyone with more knowledge/experience on this feel free to join in.

Alistair

[bob katz]

Originally Posted by cerberus

thanks bk! i didn't realize that the state of the art dynamic convolution was so primative, people are probably still amazed that it comes close, perhaps accounting for some of the glowing reviews i've read. but if it were not sampling frequently enough, i wouldn't want to use it either. another company has announced a product using sharc, i've noticed that the chip might not always deliver the power its reputation implies. i suppose i should put my "dream" for this technology on the shelf for a few years and revisit it then...

Not necessarily. You also have to go by your own ears! I don't like the sound of the HEDD's digital processing at 44.1 kHz either... it gets harsh and I have the measurements to show why. It's possible that working at double sample rate in that convolution-box-that-I-shall-keep-nameless makes its aliasing artifacts go away sufficiently so that is no longer an objection. It's also possible that K.K. doesn't notice the problem, or uses the box in a fashion where it is not a liability for her, or any number of scenarios, including "one man's meat is another man's poison."

[bob katz]

Originally Posted by RainbowStorm

I have not used Har-Bal, I've read about it though.

SNIP

You can try to create signature settings, but these will perform very poorly. Bob, I totally agree with your post.

And what about the incredibly strange philosophical idea that you can apply a given eq curve to a completely different kind of music in a misguided attempt to achieve the "effect" of another kind of music? Furthermore, for any change in compression or other dynamics that I may do when processing, this interactively affects the EQ... and if I'm 1/2 a dB off from some magical "preset" I might as well be 10 dB off to start without a preset....

BK

[bob katz]

Interesting pictures. So, all I have to do is model the static sine-wave-derived frequency response of an Ampex 351 line amp and I'll have the sound of it?

Not....

[UnderTow]

Originally Posted by cerberus

spline eq? wtf?
it's comforting to know i'm not the only engineer on earth using the strategy in mastering...


jeff dinces

Just a quick note to say that I don't use the gear-match (or spectrum match functions) in CurveEQ. I just mentioned that they are there and look similar to your graphs.

Alistair

[UnderTow]

Originally Posted by bob katz

And what about the incredibly strange philosophical idea that you can apply a given eq curve to a completely different kind of music in a misguided attempt to achieve the "effect" of another kind of music?

BK

As this is just a gimmick, a toy, I would say that asking wether you should have a device with a loudness button in your mastering chain is about as relevant as your question above.

Or are you looking for a succinct yet clear and evocative piece of text to explain why these "tools" are not approriate for, or replacements of, proper mastering techniques?

Alistair

[cerberus]

Originally Posted by bob katz

Interesting pictures. So, all I have to do is model the static sine-wave-derived frequency response of an Ampex 351 line amp and I'll have the sound of it?

Not....

you may also notice that i edited the screenshot with a real razor blade! and if you look very carefully at the "ampex", you can see that the ones and zeroes are hand wound... and ampex is a nicer name than scully, imo.

since i run it as a virtual linear phase filter, i wouldn't be attempting to fake a real tape deck with it at the moment. also i don't look too hard for authentic ampex noise sources. however, the ampex transfer function does kinda tend to smell nice after it warms up a bit, which i can't explain at all.

jeff dinces

[not_so_new]

Originally Posted by UnderTow

I'm not quite sure how you intend to use spectrum matching to find the room deficiencies. Care to explain?

Pretty easy.... Do a mix and make it sound as good as you can in the room. Then you might be able to break out the Har-Bal and check your mix against another mix you did in a different room or check it against another release. If you find that Har-Bal is cutting out a bunch of low end on your mix between 60 and 70 and it is doing the same thing at 120 and 140 then you know that you have a problem with the lows in your room.

Yes you can probably tell this with your ears using a ref CD but for the most part it takes me some time to get used to a new room even with ref CD's, seems that something like Har-Bal would help speed that process up is all.

I don't know. You can download the demo and try it.

I was not really asking if it was a good tool, I was commenting on the fact that I have not tried it but it might be cool. If and when I want to give it a try I will download the demo for sure.... thanks.

What reference would you use to match to? And how would you judge the room response based on the resulting EQ curve?


Alistair

See above.....

And as I said, I am not sure it would work, I have not tried it yet. To me it is just another tool to use. Some tools work all the time, some tools work most of the time and some tools get used once in a while here and there.

[UnderTow]

Originally Posted by not_so_new

Pretty easy.... Do a mix and make it sound as good as you can in the room. Then you might be able to break out the Har-Bal and check your mix against another mix you did in a different room or check it against another release. If you find that Har-Bal is cutting out a bunch of low end on your mix between 60 and 70 and it is doing the same thing at 120 and 140 then you know that you have a problem with the lows in your room.

I think it would be interesting to start with comparing the spectrum of different mixes you made in a room that you are familiar with (and are happy with the mixes of course). I have a feeling this will show you that this method won't really tell you much usefull info about the room.

Yes you can probably tell this with your ears using a ref CD but for the most part it takes me some time to get used to a new room even with ref CD's, seems that something like Har-Bal would help speed that process up is all.

Well you did the mix with your ears so that is part of the equation whatever happens. With this method you are also adding the difference in the actual music you are mixing. I think if anything, you are making it harder to judge the room if anything because you now have two factors to take into account instead of just your ears.

The problem I see with this method and basicly the whole spectrum matching concept is that it starts with the premise that all mixes have exactly the same frequency distribution. I don't think this is true for different songs of the same artist on the same album let alone any material that has no relation to each other beyond (maybe) genre.

And as I said, I am not sure it would work, I have not tried it yet. To me it is just another tool to use. Some tools work all the time, some tools work most of the time and some tools get used once in a while here and there.

Please do try it and tell us if it was usefull or not. Don't let my doubts about the method stop you from experimenting!

Alistair

[whosyourdaddy00]

i gotta admit, i'm a happy Har-bal user.....we're a "one stop shop" that does tracking, mixing, mastering, promo print and dupe in one place for mostly local bands and what not so i'm not dooin' anyting for serious commercial release, and i've got the utmost respect for y'alls opinions.

i use harbal strictly to help my mix's translate a bit better...i'm in a small minimally treated room mixing on 824's or in my bedroom on 20/20's (in a corner! )and i get stuck mastering my clients projects 9 times outta ten although i ALWAYS recomend takin' it elseware to be mastered. i've used har-bal to show my room's weakness by runninn' pink noize and recording that from my mix position (although simple math could give the same results)

when i use it as a 'mix translation' tool, i NEVER match it up to the 'preset' type settings for genra....rather i smoothe out my mix's peaks and dips VERY gently in my mix's 'snapshot', and generally i like the results....this has helped me out alot as i'm getting better with my mixing and am not making very drastic changes as often.......also if someone says "i want it to sound like x" i can load it in to get an idea of where i should be heading....


i don't have the luxury of a nice room with good monitoring, but i'm forced to make the best of what i have....and as far as mastering goes, all i have is Wavelab 4, Harbal, and ozone (waste of $$ ) but for these kids that wanna make a record for under a thousand dollars, well, they get what they pay for. but i do think har-bal and the other types are just another tool in the box (no pun intended).....if some kid says "all i gotta do is match the eq curve with x song and it'll sound like it" well, he'll learn soon enough. am i talking about myself?


anyway, bob, i've got your art and science book and am learning tons! i have little interest in becoming a mastering engineer, but your book has helped me mix better by keepin' my eye on the big picture. thanks man

chris

[jasman]

I use Har-Bal sometimes. I thought my personal list of HAR-BAL DOs and DON'Ts might shed some light on this misunderstood software.

DON'T pay any attention to their marketing claims.
DON'T think that this software is about matching the spectrum of one song to another. It isn't.
DON'T even look at their suggested templates.
DON'T judge the engineering smarts beneath the hood by the cover. (to mix metaphors)
DON'T be too put off by the somewhat awkward GUI.
DON'T spend any time on their support forum. Most of the users there are relatively clueless.

DO be open minded about using visual cues in addition to your ears.
DO check out the advantages of having any and every potential EQ adjustment you try automatically loudness-normalized so as to not be fooled by the "louder is better" phenomenon. It uses A-weighting -- not perfect, but pretty good!
DO listen closely to the quality of the EQ. It is linear phase, using FFT based FIR filtering, which can also accurately be called convolution. I mention the quality issue because I still haven't come to a definite conclusion about it. It is certainly usable in some circumstances. I'd like to hear other opinions from those following my list here. I have more often used it as an analysis tool, and then transferred the results to another linear phase EQ that I prefer.
DO forgive them their sin of restricting their demo to 8bit audio!

And their tech support is great.

[Kiwiburger]

I bought Har-bal, it's cheap enough. There are a lot of misunderstandings about what it is and how it's used. I would have thought their web site addresses this, if you read it. You have to over-look the marketing hype and perhaps english isn't their first lanquage. But many great products have crap marketing, just like many crap products have excellent marketing.

It's primarily a spectrum analyser, that gives you a graph for the entire song (or selected parts).

IF you are a beginner, struggling to get your mix into the ballpark,
OR if you have bad room acoustics and have overcompensated your mix,
THEN Har-bal will show you very clearly what your problem frequencies are.

Whether you believe it or not, certain genres of music tend to have very similar spectrum curves. If your music is going to compare professionally with these, you need to ensure that your mix has a very similar curve. Mastering Engineers already know this, because that's a major part of their job. Whether you use ears, or experience, or graphical programmes - the end result is a mix that has a similar spectral balance curve to other popular works in the same genre.

Where I disagree with Har-bal is in the process of using their eq to 'fix' a mix that is out of balance. As digital eq's go - the Har-bal eq is not bad, not brilliant, but not bad.

Where I totally agree with Har-bal is that you can't just grab a curve from one work and apply it to another. Har-bal does not do this. There are other products that do that - and they don't get the same amount of flak as Har-bal, who don't even do what people allege they do.

Har-bal is intended to be used with your ears first and foremost. It is for that reason they put in the loudness compensation. With a normal eq, if you boost a frequency, the whole mix gets louder. People tend to think louder is better, so we tend to be biased in favour of adding eq. Har-bal compensates for this - so your ears are not fooled by loudness increases.

In my view- Har-bal is not a gimmick, and shouldn't be compared to the real gimmicks out there. It's just tool - a good tool for the price. It was never intended to replace ears.

But for most of it's intended market, this is for guys 'mastering' their own mixes. I personally believe that if you have access to the mix, you should fix any identified problems in the multichannel mix, not by eq'ing the 2bus mix.

But that's just me. I found it a useful learning tool for making better mixes that are more like a record. I would still prefer to pay a ME to master my best work.

[tubedude]

Originally Posted by Kiwiburger

I bought Har-bal, it's cheap enough. There are a lot of misunderstandings about what it is and how it's used. I would have thought their web site addresses this, if you read it. You have to over-look the marketing hype and perhaps english isn't their first lanquage. But many great products have crap marketing, just like many crap products have excellent marketing.

It's primarily a spectrum analyser, that gives you a graph for the entire song (or selected parts).

IF you are a beginner, struggling to get your mix into the ballpark,
OR if you have bad room acoustics and have overcompensated your mix,
THEN Har-bal will show you very clearly what your problem frequencies are.

Whether you believe it or not, certain genres of music tend to have very similar spectrum curves. If your music is going to compare professionally with these, you need to ensure that your mix has a very similar curve. Mastering Engineers already know this, because that's a major part of their job. Whether you use ears, or experience, or graphical programmes - the end result is a mix that has a similar spectral balance curve to other popular works in the same genre.

Where I disagree with Har-bal is in the process of using their eq to 'fix' a mix that is out of balance. As digital eq's go - the Har-bal eq is not bad, not brilliant, but not bad.

Where I totally agree with Har-bal is that you can't just grab a curve from one work and apply it to another. Har-bal does not do this. There are other products that do that - and they don't get the same amount of flak as Har-bal, who don't even do what people allege they do.

Har-bal is intended to be used with your ears first and foremost. It is for that reason they put in the loudness compensation. With a normal eq, if you boost a frequency, the whole mix gets louder. People tend to think louder is better, so we tend to be biased in favour of adding eq. Har-bal compensates for this - so your ears are not fooled by loudness increases.

In my view- Har-bal is not a gimmick, and shouldn't be compared to the real gimmicks out there. It's just tool - a good tool for the price. It was never intended to replace ears.

But for most of it's intended market, this is for guys 'mastering' their own mixes. I personally believe that if you have access to the mix, you should fix any identified problems in the multichannel mix, not by eq'ing the 2bus mix.

But that's just me. I found it a useful learning tool for making better mixes that are more like a record. I would still prefer to pay a ME to master my best work.

I'd be interested in the demo if it werent 8 bit, how goofy is that?
It does sound like a lot of people are wanting to harass the software without hearing it, which makes me want to try it even worse. The idea is good, the implemnation MIGHT be good, but you def cant knock it if you havent tried it extensively.
I may buy it shortly and check it out. Voxengo might be a better way to go though, but who really knows.

[masteringhouse]

As an analogy, has anyone tried using something like Photoshop's auto levels or auto contrast to fix an image? Assuming Har-Bal is even that good, how close does any "wizard" compare against an experienced person in the trade?

[everybody's x]

Originally Posted by Kiwiburger

I bought Har-bal, it's cheap enough. There are a lot of misunderstandings about what it is and how it's used. I would have thought their web site addresses this, if you read it. You have to over-look the marketing hype and perhaps english isn't their first lanquage. But many great products have crap marketing, just like many crap products have excellent marketing.

It's primarily a spectrum analyser, that gives you a graph for the entire song (or selected parts).

IF you are a beginner, struggling to get your mix into the ballpark,
OR if you have bad room acoustics and have overcompensated your mix,
THEN Har-bal will show you very clearly what your problem frequencies are.

Whether you believe it or not, certain genres of music tend to have very similar spectrum curves. If your music is going to compare professionally with these, you need to ensure that your mix has a very similar curve. Mastering Engineers already know this, because that's a major part of their job. Whether you use ears, or experience, or graphical programmes - the end result is a mix that has a similar spectral balance curve to other popular works in the same genre.

Where I disagree with Har-bal is in the process of using their eq to 'fix' a mix that is out of balance. As digital eq's go - the Har-bal eq is not bad, not brilliant, but not bad.

Where I totally agree with Har-bal is that you can't just grab a curve from one work and apply it to another. Har-bal does not do this. There are other products that do that - and they don't get the same amount of flak as Har-bal, who don't even do what people allege they do.

Har-bal is intended to be used with your ears first and foremost. It is for that reason they put in the loudness compensation. With a normal eq, if you boost a frequency, the whole mix gets louder. People tend to think louder is better, so we tend to be biased in favour of adding eq. Har-bal compensates for this - so your ears are not fooled by loudness increases.

In my view- Har-bal is not a gimmick, and shouldn't be compared to the real gimmicks out there. It's just tool - a good tool for the price. It was never intended to replace ears.

But for most of it's intended market, this is for guys 'mastering' their own mixes. I personally believe that if you have access to the mix, you should fix any identified problems in the multichannel mix, not by eq'ing the 2bus mix.

But that's just me. I found it a useful learning tool for making better mixes that are more like a record. I would still prefer to pay a ME to master my best work.

I support this post ^

I bought har-bal, I learned a ton from doing so. It doesn't match itself to another file, it does point out very specifically where my weaknesses are when I am trying to get a certain result. It is a useful tool. I haven't mastered anything with it, but I have analyzed alot of individual instruments to understand how to get in the ballpark, and my mixes have improved tremendously because of it. Much like reading every post on GS from helpful slutz