gubu

ambient wrote: http://soundcloud.com/mysoapbox/waiting-first-mix I haven't really bothered trying to 'master' it, but I did put a limiter on the whole thing to catch the higher peaks and bring up the level. I doubt it's as loud as most commercial mixes though. This is the first track I've every mixed and also my first time singing/writing lyrics. Any suggestions on improving the mix? Love the song. The mix is pretty good - there are some things that jump out at me. Bear in mind that these critiques are usually wholly subjective! I'd like to hear it without the limiter on there. It seems like the louder sections, the choruses and the outro, are too intense, too much limiting. Everything is a little bit too upfront during these sections. Did you EQ some of the mix elements for presence, boosting the mids/hi-mids? That could also be the reason why it sounds a little too upfront. If you did EQ for maximum presence, I'd kill those EQ boosts on the guitars for starters. Apart from that, the sonics are really good. Crystal clear, and importantly, a good song and a musical mix. My only other suggestion might be to cut somewhere around 1000-1600Hz on the 'loudhailer' style vocals, and take them back 2 or 3 dB overall, perhaps turning up the reverb send as well (especially on the 'come back' lines @~2:20). That type of vocal effect needs quite a bit of reverb to sit in a mix properly, and probably some compression.

Hi Studio Trenches, http://soundcloud.com/gubu/times-blossom3 Hoping to get some feedback on this mix and the overall sound. I just did a K-System calibration and this is the first mix off the machine. All comments appreciated, like how hard is it to mix a vocal and guitar haha! ad

Beautiful guitar. Stolen from Cahir, Co. Tipperary, Ireland in 1996. No reward offered, just give it back please, it was a 1st wedding anniversary present

Where did I ever argue this???? Do you even read my posts? I'm arguing for use of a higher sample rate than 44.1kHz when recording, that's all. GOOD LUCK

Another misinterpretation of the Larvy paper. Well this is no misinterpretation, 48kHz will yield better results than 44.1kHz if the optimal rate according to Lavry is 64kHz. That would lead me to move up to 88kHz at my next upgrade to ensure accurate sampling and more importantly processing of waveforms that I can hear.

No, it is a provable misinterpretation. It's not an opinion, it is a provable fact. He clearly states in several places the exact opposite of what you were claiming. Well you can't argue with his assertion that 192kHz is 3 times faster than the optimal rate. That would make the optimal rate 64kHz. You are welcome to record away in 44.1kHz and the very best of luck to you....

Here's a list of your previous quotes, all of which are based on misinterpretations: All of Post 107 All of post 107?????? Are you crazy or something, how is a direct quote a misinterpretation?????

Here's a list of your previous quotes, all of which are based on misinterpretations: All of Post 107 It's a different interpretation to yours, not a misinterpretation. You can use whatever sample rate you want, I'll use the highest one I can afford, which happens to be 48kHz at the moment. You can argue about interpretations all you want but I've made plenty of points here that no one has argued with. You should work at the highest possible time domain resolution that you can afford, a square wave is a square wave except when it's at the limit of human hearing and sampled at 44.1kHz, when it becomes a sinusoid. That is distortion and if that's good enough for you then good luck to you

And that in a nutshell is exactly why this has been going round and round. You've made your decision and won't listen to anyone saying otherwise. Fine. Just don't expect us to act like the science backs you up. It does back me up tho. Your high frequency signals are distorted at the convertor input by the filter needed to sample at 44.1kHz and that is indisputable. Square wave in->sinusoid out is distortion whether you claim you can hear it or not. I'll bet that quite a few of you that have been arguing with me still track at 48kHz or higher and if you believed your own arguments, you should be tracking at 44.1kHz otherwise why bother...

I find it slightly strange that you continue to make claims about things being "audible to DSP" when you have clearly demonstrated a misunderstanding of basic DSP principles and have misinterpreted statements from the Larvy article. I think you are saying that these inaudible frequencies are still processed by the mixing engine. This is true, but they are inaudible frequencies! You continue to claim that inaudible frequencies have an audible effect, but you have yet to give any explanation for this. Do you have any evidence to support this? Hopefully, something more technical than "it's audible to DSP"? What exactly have I misinterpreted from Lavry? That the optimum sampling rate for digital conversion using current technologies is 64kHz. He's pretty clear about that one. Look, if you have a square wave at the microphone and op amp going in, it's better to process that square wave 'in the box' and not it's sinusoidal 1st cousin generated by your convertor system at 44.1kHz. If you can't see the truth of that, I'm wasting my time arguing with you. The ABX test quoted earlier in the thread is just not a valid argument to me regarding recording of audio as it deals solely with playback. Much more to the point regarding recording would be to set up a scenario such as my friend and I did, feeding the same mic signal to 2 identical systems running different sample rates and do a double blind ABX test from there. I've no problem admitting that our test is not scientifically valid due to it not being double blind but I know what I heard and I know that I struggle a little less when processing my higher frequency signals in 48kHz than in 44.1kHz. That's my experience, take it or leave it, I couldn't care less at this stage, it all ends up on iPods anyway..

No, it demonstrates that your question was specious as the extra information in a 20kHz square wave is not audible to the human ear. Terry D. Maybe not, it's audible to DSP tho and that will make a difference to your final mix whether you like it or not.

I'm still waiting for an explanation of this statement. Just to be clear, I agree with a lot of what you are saying (I never claimed there was not an audible difference between converters with different sampling rates), but many of your explanations and justifications are wrong. I'm not even trying to argue against your main point (at least I think), but your explanations are spreading misconceptions. I never said anyone was imagining things. The reason nobody is attacking this poster is because he never made incorrect statements about mathematical principles, he simply stated his opinion. That's fair enough but I've had the same people jump down my throat on this thread for expressing my opinion as for attacking mathematical principles. And since I've brushed up on the math, i.e. Lavry's technical paper, I haven't changed my mind, there is a clear advantage to recording in a higher sample rate. No one is going to convince me otherwise. Even if you can't hear the difference, and I reckon you can, even between 44.1 and 48kHz, it's better to process a square wave as a square wave and not as a sinusoid, you will simply get a better recording at the end of the day. Factor in DSP and it's a clear advantage using a higher sample rate. See my quote from Lavry's paper above, his estimation of the optimal sample rate for conversion is 64kHz. Don't tell me that doesn't support my argument for using a higher sample rate. I'm sorry if my points have been poorly made and if I insulted anyone's intelligence earlier in the thread, the science and math of it all was always in the back of my mind, I just needed something like Lavry's paper to refresh it all. I never said 44.1kHz wasn't the optimal or best sample rate for playback, i.e. CD's. CD's sound bloody great and will not become obsolete for probably another 20 or 30 years. My point since the beginning of this thread has been that there is an advantage to recording in a higher sample rate and I have heard, read or seen nothing in this thread to convince me otherwise.

http://mixonline.com/recording/mixing/audio_emperors_new_sampling/ I'm sorry, this article does not say either what the source material was nor how it was recorded, what sample rate, etc. Just because it's playing back on a SACD, doesn't mean it was recorded at the same sample rate. You can convert up as well as down and because of that, this article is inadmissible as scientific proof of anything.. edit:- It would have helped if the OP had linked to a publication of the actual test and not a magazine article about the test

Sorry but you can't keep referring to this test of yours as if it were good evidence for your claim. It is simply anecdotal and is worthless as scientific evidence. If you don't understand why I'm saying that, please google "the scientific method" and "double-blind testing". I was replying to a post asking about my hearing. Well, we're all in agreement that 20Khz square wave in = 20kHz sinusoid wave out in 44.1kHz. That is indisputable. The only double blind scientifically valid test mentioned in this thread showed that it's very difficult to tell the difference between SACD and CD playback and this has nothing to do with my argument which is that it is beneficial to use a higher sample rate than 44.1 when recording. I made this same point to someone else already further back the thread and I'm going back to find the link and have a read of this paper because if the source audio was recorded at a sample rate less than SACD, it invalidates the entire test.. Is that enough scientific method for you?

I'll throw my last two cents in here. The difference I HEAR with a higher sampled rate seems to be with live drums recorded, mainly cymbals and snare. Some other sibulance gets there too but I only get it using full frequency condencers. I dont nessasarily consider those items to generate sine waves in the higher frequencies. Some of it resembles something closer to white noise but the frequencies do reach quite high. To have a good recording you dont have to have those frequencies but they sure sound more realistic to me. I do notice it mostly when I am mastering a mixdown and EQing the complete mix. On a 44 mix there is hardly anything usable over 16K. If I bring up the highest frequencies I hear white noise either from a recorded amplifying device, sound card converters or plugin noise, amp his, whatever. Sounds bad and isnt usable so the completed masters lack a certain amount of air. When the tracks have been sampled at higher rates, and all the normal mixing has been done and the mixdown is being masterd, the overall frequency range the analizer detects is extended above 20K. The high frequencies in the 20K region can be boosted without sounding like noise. The frequencies above 20K arent needed so they can be rolled off. The quality of the 16~20K has less hiss, noise and harshness then it does at lower rates. There is alot of debate peoples opinions here and I think its great. I believe the real answer is using your resources to get a superior sounding final product. For me, that littel extra high frequency has been like a godsend getting live drums to sound better and compete better with the other instruments I normally had to carve up to match the drums. In my case getting the high end of the cymbals in the clouds above the vocal range had been the biggest battel. It was trickey work working with 16/44 on many many thousands of recordings over a 10 year period and trying to get super pro results. Not alot of room for error. For me the job was made a bit easier. I don't see y'all tellin this guy he's only imagining things!! :lol:

Again, this means that you claim to be able to hear frequencies above 22 kHz, or at least ~20 kHz when factoring in filtering. Is that what you are saying? No, but how would I know anyway when my convertors only go up to 48kHz? The very least I'd like to do is keep the waveforms intact when mixing and this is not possible in 44.1kHz. Now I'll refer back to the AB test I carried out with a friend a couple of years back and outlined repeatedly earlier in the thread. The recordings in 96kHz had more depth, spaciousness and definition on playback than the ones in 48kHz, pure and simple. We didn't imagine it, the difference in detail alone was remarkable, and this is from a single mic on an acoustic guitar. It doesn't matter to me at this stage if it was more to do with filtering than sample rate, if you can hear your sources more clearly when mixing, that is a GOOD thing. Factor in DSP and multiple tracks and you are going to get a better sounding mix if you work in a higher sample rate.

Ok, now I'm completely lost. Are you claiming you can hear the difference between a 22 kHz sine wave and a 22 kHz square wave? I don't know, I've never done a test on that and I can't either because the convertors in my equipment can only handle 48kHz See what I mean. I'm getting tired too. I'll leave it this:- I would much rather record and mix in a format where the harmonic content that affects waveforms at frequencies I can hear is preserved and that means using a higher sample rate. When I can afford it, I probably will move up to 96kHz, the rest of you can do what you like :D:D

Agreed, but again you were attacking mathematical principles, not implementations based on those principles. Can we agree that the "optimal" sampling rate for recording is the rate that can capture all audible frequencies such that the filters have no effect on audible frequencies? Please! I'm getting tired. No I can't agree to that. The inaudible harmonics have an affect on the audible waveform

Ok, to completely avoid confusion, please define what you mean by an "accurate reproduction". If you mean "mathematically identical signal", then you are correct (obviously a sine save is different than a square wave). If you mean "audibly identical signal", then you are wrong. Please elaborate. I think we're venturing into human biology and I'll go there if we have to

I don't think anyone disagrees with this, but the reason is due to the practical consideration of filters (which is discussed in the article). You have claimed that a "sampling rate" cannot reproduce those frequencies. I think you mean that "converters" with a particular sampling rate cannot reproduce those frequencies. Right? OK, this is just semantics, we have to use 'convertors' to apply the sampling rate in the real world, otherwise we're still drawing graphs and recording nothing. And if you're not disagreeing with what you've quoted from me above, then you can't be disagreeing with my original argument which was that a piece of equipment running at a higher sampling rate will yield a more accurate reproduction of the waveform

This is getting off topic, but he doesn't claim that 192 isn't optimal *because* machine speeds aren't sufficient, he is claiming that 192 isn't needed because of the reasons we've stated *and* machine speeds aren't sufficient. Again, he makes the following statement: "It is important to realize that the end result yields a waveform where the values are correct, not just at sample times but at all times. You DO NOT need more dots. There is NO ADDITIONAL INFORMATION in higher sampling rates. As pointed out by the VERY FUNDUMENTAL Nyquist theory, we need to sample at above twice the audio bandwidth to contain ALL the information." That is all well and good but if you input a square 20kHz wave and output a sinusoid 20kHz wave, that is NOT an accurate reproduction of the waveform and it DOES make a difference, particularly when DSP is applied and there is no getting away from DSP these days.

I think we are all actually saying the same thing now, but this is clearly not what you were originally arguing. Well I think my original argument was that a higher sample rate will give you a more accurate map of the waveform, and now you're agreeing with me. I apologise that it's taken me so long to actually express this in proper technical language but my book learning was way back in my grey matter at the start of this argument and it's taken 2 days for me to retrieve it.

He says it's 3 times faster than the optimal rate, AND compromises accuracy. How does this back up your claim that it will provide a better waveform? 192/3=64 This lays to rest everybody elses' claim on here that 44.1kHz is just as accurate as higher sample rates. I've been arguing for higher sample rates since the beginning of this thread. Since reading Lavry, I'm no longer claiming that 192 will provide a better waveform as he states that machine speeds are not yet fast enough to encode enough data per sample at that speed and I'm quite happy to accept that. Quite simply, the more harmonic content you can sample for waveforms approaching the upper limit of human hearing, the more accurate and faithful a representation you will have of those waveforms. That means raising the filter frequency and therefore raising the sample rate.

On what page does he make this claim? ^^^ There's the quote in previous post, it's on page 27 of his paper.

'There is an inescapable tradeoff between faster sampling on one hand and a loss of accuracy, increased data size and much additional processing requirement on the other hand. AD converter designers can not generate 20 bits at MHz speeds, yet they often utilize a circuit yielding a few bits at MHz speeds as a step towards making many bits at lower speeds. The compromise between speed and accuracy is a permanent engineering and scientific reality. Sampling audio signals at 192KHz is about 3 times faster than the optimal rate. It compromises the accuracy which ends up as audio distortions.' That's a copy and paste from Lavry's paper, my italics.