 |
 |
 |
 |
 |
 |
 |
Forums - Mixing & FX
Subject: brittle mixes
Original Message Date: 18-Oct-02 @ 11:46 PM - brittle mixes
argh, I've got digital bite. A term I've heard about for harsh brittle mixes done a digital recording setup.
I've been noticing that a lot of my recordings are really brittle. I notice that the uppermids and highs mix together in this weird way causing a digital harshness that irritates the hell out of me. On a low volume everything sounds good--nice and smooth. But once I turn the volume up, ouch, my ears begin to bleed.
I think my problem is that I dont give enough headroom or make enough headroom by cutting off and rolling off frequencies. But this all makes me wonder... I use a digi 001 and I use it's mic pre's and inputs. Could there be any link in the 001 casuing a harsh sound? Or is it that I'm not knowledgable (sp?) enough in getting a good sound that mixes good together.
I hate that feeling when I turn up some music (on the same monitors) and get absolutely no harshness or bite. bums me out.
Anyone have any ideas on what I could do or what to improve with?
I've been noticing that a lot of my recordings are really brittle. I notice that the uppermids and highs mix together in this weird way causing a digital harshness that irritates the hell out of me. On a low volume everything sounds good--nice and smooth. But once I turn the volume up, ouch, my ears begin to bleed.
I think my problem is that I dont give enough headroom or make enough headroom by cutting off and rolling off frequencies. But this all makes me wonder... I use a digi 001 and I use it's mic pre's and inputs. Could there be any link in the 001 casuing a harsh sound? Or is it that I'm not knowledgable (sp?) enough in getting a good sound that mixes good together.
I hate that feeling when I turn up some music (on the same monitors) and get absolutely no harshness or bite. bums me out.
Anyone have any ideas on what I could do or what to improve with?
Message 21/38 23-Oct-02 @ 11:08 PM - RE: brittle mixes
jesus you guys are pedantic. I love that word
panama..you gotta do BOTH, man. crank it and listen to it very quietly. and...you do this right..solo sounds and see where they sit?
panama..you gotta do BOTH, man. crank it and listen to it very quietly. and...you do this right..solo sounds and see where they sit?
Message 22/38 23-Oct-02 @ 11:31 PM - RE: brittle mixes
nomad, a question: are you sure that hi samplerate converters dont remove stuff beyond 20k prior to sample and hold stage with the anti aliasing filter? i thought that in a "normal" converter the filter is set before the conversion stage, so the sampling frequency has no influence on the harmonic content beyond that.
i´m really curious about this since i´ve never actually inquired about high samplerate converters but have always presumed that they simply offer higher resolution within audible frequency range.
i´m really curious about this since i´ve never actually inquired about high samplerate converters but have always presumed that they simply offer higher resolution within audible frequency range.
Message 23/38 24-Oct-02 @ 12:47 AM - RE: brittle mixes
this is true. ADC's use a brickwall lowpass filter at 20k before conversion to eliminate an effect called "foldover." This is the same effect as a car's wheels looking like they're going backwards when it goes real fast. It happens when a system analyzes something at regular intervals that are slower than the intervals of what's being observed. The filter knocks out frequencies that could be interpreted wrong by the convertor.
So one of the selling points of higher sampling rates is that you don't need that filter anymore, or at least can move it way up beyond our hearing range.
psylichon
So one of the selling points of higher sampling rates is that you don't need that filter anymore, or at least can move it way up beyond our hearing range.
psylichon
Message 24/38 24-Oct-02 @ 03:00 AM - RE: brittle mixes
there's nothing magical about 20K; 44.1K was chosen because humans can hear up to about 20K, not the other way around. a higher-frequency ADC should have the filter(s) higher, but don't have to; a 96K ADC i would expect to have the filter set at around 40K.
you always have
audio->anti-aliasing filter->ADC->digital stream
then
digital stream->DAC->reconstruction filter->audio
both filters are VERY important to the full sampling process, but both filters are only required to filter anything about nyquist (the first to prevent aliasing or 'foldback' as mentioned above, and the second is what turns xoxos' two-point squarewave back in to the original nice sinewave). but as long as these are below the nyquist (half the sampling frequency) it's fine; they can be more than halfway down without killing anything.
in the 'old days' a brickwall filter could cause audible effects (phase shift etc.), but this has long since ceased to be an issue... modern ADCs are typically sigma/delta type, which work internally at much much higher frequencies and a brickwall filter there can work without any audible effect on the signal.
you always have
audio->anti-aliasing filter->ADC->digital stream
then
digital stream->DAC->reconstruction filter->audio
both filters are VERY important to the full sampling process, but both filters are only required to filter anything about nyquist (the first to prevent aliasing or 'foldback' as mentioned above, and the second is what turns xoxos' two-point squarewave back in to the original nice sinewave). but as long as these are below the nyquist (half the sampling frequency) it's fine; they can be more than halfway down without killing anything.
in the 'old days' a brickwall filter could cause audible effects (phase shift etc.), but this has long since ceased to be an issue... modern ADCs are typically sigma/delta type, which work internally at much much higher frequencies and a brickwall filter there can work without any audible effect on the signal.
Message 25/38 24-Oct-02 @ 03:06 AM - RE: brittle mixes
btw, i'm guessing this isn't it based on what you described, but you couldn't be hitting distortion somewhere (on one of the few songs i've actually finished, i did a mix that sounded OK but the mix wasn't that great and it was too low volume; the next mix i did came out sound brittle, and it was because a couple of channels i had turned up, and they were starting to clip).
note that EQ and resonant filters are often guilty of this; if you take a wave file that's nearly full scale, and give it a +4db EQ boost somewhere, you could potentially push it into distortion.
there are some crappy A/D converters (and curiously enough a lot of old digidesign ones are said to be in this category
but i doubt yours is bad enough to do what you suggested.
also (maybe doesn't apply to you, but can on analogs), often a mixer will have several gain stages, but the 'clipping' light only measures one of them... so in some mixers you can overdrive the trim, then pull the levels down on the fader enough, and you can hear the clipping but the light won't come on. same for EQ boost. those kinds of distortions almost always sound brittle to my ears.
note that EQ and resonant filters are often guilty of this; if you take a wave file that's nearly full scale, and give it a +4db EQ boost somewhere, you could potentially push it into distortion.
there are some crappy A/D converters (and curiously enough a lot of old digidesign ones are said to be in this category
but i doubt yours is bad enough to do what you suggested. also (maybe doesn't apply to you, but can on analogs), often a mixer will have several gain stages, but the 'clipping' light only measures one of them... so in some mixers you can overdrive the trim, then pull the levels down on the fader enough, and you can hear the clipping but the light won't come on. same for EQ boost. those kinds of distortions almost always sound brittle to my ears.
Message 26/38 24-Oct-02 @ 09:07 AM - RE: brittle mixes
so you´d guess they´d have the filter set higher? hehe... i was wondering if someone knows the answer to that one. of couse, the real question is what happens when you use 44k resolution on a 96k converter. and sorry about o/t-ness.
Message 27/38 24-Oct-02 @ 02:22 PM - RE: brittle mixes
you can tell where they have the filter set by looking at the frequency range... for instance, if you look at the lucid AD9624, in 88.1/96K modes, its specs say it has a frequency range of 20Hz-40Khz; this means by definition it can't be filtering at less than 40Khz. however, in 44.1/48Khz mode, it only has a frequency range of 20Hz-20Khz. that means in 96K mode it filters out things above 40Khz, and in 48K mode it filters things out above 20KHz. there might be a few badly-designed 96Khz converters that always filter above 20Khz, but then you're not getting much benefit from the higher rate. they just switch which filter they are using based on the sampling rate, it's not that difficult actually.
Message 28/38 24-Oct-02 @ 02:38 PM - RE: brittle mixes
interesting. thanks for the info, you know your stuff. obviously the converter technology has advanced since i learnt about it.
cheers, m.
cheers, m.
Message 29/38 29-Oct-02 @ 12:28 PM - RE: brittle mixes
i had that problem for ages mate could be to do with ur mixer but more than likely its just that u need to spend more time on ur indiviual parts making sure they sound good at high volumes b4 puttin them in ur track
Message 30/38 29-Oct-02 @ 07:40 PM - RE: brittle mixes
stop trying to discredit my reasoning! it's exact! i never said it would be square! you only thought square wave! the square wave is your mind, not mine! i can't believe it! you lead the reader into misinterpreting my statement, then you waffle on about whether a 20k waveform will be discernible which is vacuous. this signal fidelity is an abstract represented by performance over time, not the quality in any given instance, which is much more musical! heh heh
Pages: 1 2 3 4
There are 38 total messages for this topic
Reply to Thread
You need to register/login to use the forum.
Click here to Signup or Login !
[you'll be brought right back to this point after signing up]
Back to Forum
