Alonzo's frequency analyses
|
Bronze Member Username: Thegr33nninjaCA United States Post Number: 16 Registered: Oct-11 | Okay, for lack of a general discussion forum I am left with posting this here. Alonzoub has been kind enough to let us give him songs to test the frequency on to let us see exactly what frequency the song is hitting! Both to give him practice, and us cool information.(plus maybe new music to bump to! haha) Just list up to 3 songs at a time (don't want to over burden him haha) and a 30 second time frame, or around there, and hope to hear back from him! (If this thread belongs somewhere else let me know, or ask me to take it down) I'll start: Datsik & Bassnectar - Elevate (0:55-1:25) Niraya - Earth Colonized (1:08-1:38) Omnitica - Dubwoofer Substep (1:00-1:30) Hey Alonzo, if you're ever done with doing this let us know and we'll stop asking you...as long as you provide us with a good source for some software hahah I got time to kill before the weekend and I plan on learning about some songs |
|
Silver Member Username: MakinblakMonroe/Monticello, La/Ar Post Number: 775 Registered: Oct-10 | 1) Soulja boy ft. Lil john-G Walk screwed and chopped :30-:60 2) Hank Williams jr-outlaw women :00-:30 3) Dubstep- MWanted :60-:90 |
|
Silver Member Username: Joe1234Post Number: 801 Registered: May-09 |  Kafani - Feelin Like |
|
Silver Member Username: AlonzoubPost Number: 274 Registered: Apr-10 | @Bo Sorry I couldn't seem to find "Niraya - Earth Colonized" anywhere where it was readily available for download... if you have another method I can aquire the song then I can analyze it for you. Heres the other two:   @SlackinMack I couldn't find "Dubstep- MWanted"   |
|
Bronze Member Username: Thegr33nninjaCA United States Post Number: 17 Registered: Oct-11 | Dude Joe, that song is sick just for the bass haha Only rap song i'll listen to at the moment lol Wonder how that will play on my future sub... And nice work Alonzo! I'll be keeping em coming haha About that Niraya song though, I forgot I couldn't even find it lol...good song though Dude....can't wait to bump this one on the sub: Roksonix-Madness(0:27-0:57) [forum is filtering this song's title...IV = N]Borgore-IVympho(0:07-0:15) wobble bass! And I want to see if the bass changes once it gets filthy...(1:02-1:32) And if you can make an exception haha (it's only 3 songs..never said more than 3 charts! lol) This song is sick: Kanji Kinetic-New Era (1:07-1:37) |
|
Silver Member Username: Joe1234Post Number: 813 Registered: May-09 | Put On - Young Jeezy |
|
Silver Member Username: AlonzoubPost Number: 277 Registered: Apr-10 | OMG NO MORE THAN THREE CHARTS!!! just kiddin... heres what I got: @Bo     @Joe You didn't specify a time frame so I winged it and picked 0:15 - 0:45  |
|
Platinum Member Username: Rovin1 15 = 153 DBs ...Trinidad & T... Post Number: 17080 Registered: Jul-05 | nice thread always wondered what that kafani song dropped too , that is quite low ... |
|
Platinum Member Username: Rovin1 15 = 153 DBs ...Trinidad & T... Post Number: 17081 Registered: Jul-05 | that shud be "dropped to" not 'dropped too' , bad english ftl lol ... |
|
Silver Member Username: Joe1234Post Number: 823 Registered: May-09 | Alonzo, thanks, you did hit the lows! |
|
Platinum Member Username: Rovin1 15 = 153 DBs ...Trinidad & T... Post Number: 17083 Registered: Jul-05 | i kindly request u do these 3 .......................for now - lol white girl - jeezy solja boy - pretty boy swagger how to love - lil wayne ... |
|
Bronze Member Username: Thegr33nninjaCA United States Post Number: 18 Registered: Oct-11 | Alright Alonzo, I probably should of asked this sooner...but what does the left column represent? The intensity of the bass? Don't know what Y(f) is lol The bottom row is obvious haha and I might have a couple songs: Mr Postman-Cragga(0:18-0:48) Excision-X(Original Mix) (0:40-1:10) heres a link to it just because it may be a little tricky to find.. [http://]tinyurl[.]com/6w6kvmb (no brackets) InDisco-Bar9 Remix (1:33-2:03) And Rovin, I do think this thread is pretty sweet haha I just noticed you didn't specify a time frame though...let's hope he picks a good spot in the songs you want! haha And Joe, nice recommendation, that song hits quite low...might bump it one day lol maybe... |
|
Silver Member Username: AlonzoubPost Number: 278 Registered: Apr-10 | K, I just wanna say, you can interpret the graph however you like, the data points that I mark are simply chosen because of the peaks that I see in the graph, a couple of these looked a bit messy so I wasn't quite sure where to mark them. Oh and btw, that is not the original song by Kafani in the screenshot, it is the edited (slowed down) version by DECAF. @ Rovin    @ Bo    |
|
Silver Member Username: AlonzoubPost Number: 279 Registered: Apr-10 | As far as your question goes Bo, basically yes you COULD say it is the intensity of the bass, but its all relative. I will explain how my code works just incase it may answer some un-asked questions or just help someone out. I start by converting the song file I download from whatever format to an uncompressed WAVE file. The way a wave file works is it is simply a LOOONG list of datapoints that create the waveform for the sound being played, actually for STEREO files it is two lists, on for each channel. The datapoints range from 1 to -1 as far as amplitude goes. When a song is sampled at... say... 44kHz... that means every SECOND there are 44000 data points that define the wave form PER channel. You may be able to see now why I would want to cut the analyzed portion of the song down to 30 seconds, 30 seconds = 2.64 million data points. In my code I decided instead of analyzing each channel separately, I would divide the amplitude of the signal of each channel by 2 and then sum the channels to make a mono signal. This screenshot below is about 1/10th of a second of a waveform extracted from a song. This small section is comprised of about 4400 datapoints to create the wave looking function.  Obivously I can not tell directly what kind of frequencies are being represented here, the signal is just too complex. This is where the Fast Fourier Transform comes into play. If any of you have taken engineering courses you should be familiar with this... if not I will explain what this does. The Fast Fourier Transform basically takes an input wave form in the form of sampled datapoints (like an uncompressed WAV file) and runs it through an algorithm that deconstructs the waveform into frequency vs amplitude. This was made so that engineers can record some sort of electrical/mechanical/vibrational data in the form of a wavefrom and analyze the data to find how a system is responding to something in terms of frequency and amplitude, for example if you were to attach an accelerometer to something like a tuning fork or a bell and record data of the vibration of the object once it has been struck, you would end up with a waveform that (once run through a fourier transform) would define the natural/resonant frequencies of the object. Im sure there may be a simpler method of doing this that was actually designed and coded for the AUDIO world... but for now this works for me. Since I use this program for school (program is called Matlab btw) I thought I might aswell get some practice with it. Hell maybe having a-little APPLIED experience will help me land a job one day.... lol. ....lemme know if you have any questions. Basically the left side of the graph is Amplitude of the signal but in reference to the rest of the signal. The numbers should mean nothing more than a guide to reference the amplitude of one specific frequency with respect to another. If one song has 60Hz at an amplitude of 0.2 and 50 Hz at an amplitude of 0.1... I would find it FAIR to ASSUME (im not 100% on this) that the POWER at 60Hz may be double than the POWER of the 50Hz section. |
|
Bronze Member Username: Thegr33nninjaCA United States Post Number: 19 Registered: Oct-11 | Dang Alonzo...that was a deep explanation haha I got what I wanted out of the first two sentences though lol Thanks for that though and for checking all these songs out! What have you been doing with them anyways? I know you download them, but I wonder if you just delete them after or actually listen to them haha Anyways...I just found the name of a song where i've seen sooo many people testing their subs to it: Young Jeezy-Hypnotize (0:26-0:56) bass line is pretty sweet actually, that's about it though lol imo.. |
|
Silver Member Username: AlonzoubPost Number: 280 Registered: Apr-10 | Eh.. im thinking one of the sketchy places I got a song from gave me a virus. I got a crap ton of errors and then windows decided to not want to boot... and none of my partitions showed up in the windows repair utility, just a blank list. Just got done with a clean format/re-install... Good thing I already have Jeezy Hypnotize from back when I finished MY box and needed some test music .  For the most part I am keeping all of the music, but for no real reason, I haven't really listened to much of it. I might start deleting tho, due to 30 seconds of uncompressed WAV files being larger than the ENTIRE song in mp3 format... my "frequency analysis" folder has gotten up to half a gig. |
|
Silver Member Username: Joe1234Post Number: 840 Registered: May-09 | I recommend using a virtual machine for those nice p2p softwares or the likes, specially if you are doing worthwhile stuff on your computer. |
|
Silver Member Username: AlonzoubPost Number: 281 Registered: Apr-10 | I was too lazy to even do the p2p... i simply would google the name of the song along with "mediafire" or "hulkshare" and 99% of the time I would find the song. Its ok though because I still had another harddrive with Vista installed and I was still able to backup all of my important things before formatting. A bit of a pain re-installing all of my beefy programs but honestly it needed a nice cleaning anyway. Ill keep the virtual machine thing in mind though. |
|
Silver Member Username: Joe1234Post Number: 842 Registered: May-09 | Hmm ok a virtual machine is hard to setup so probably I'll advice a sandbox, it set itself up in 20 sec and wastes no harddrive space: http://www.sandboxie.com/ When I have the time I will evaluate a free spectral analysis software I have downloaded so everyone here get the chance of blowing themselves up! |
|
Silver Member Username: Joe1234Post Number: 844 Registered: May-09 | Well here it is: Kafani - Feelin Like As drawn by WaveSpectra 1.40 (default settings)  To start it draws on log-log axes and to 20Khz, the curve represent the peak level that the song reached at every frequency in the full length of the song, if anyone is interested download the program.. here: WS140.ZIP *scan it well* before using. |
|
Silver Member Username: AlonzoubPost Number: 282 Registered: Apr-10 | For anyone else that is interested in downloading the program: Scan finished. 0 out of 20 scanners reported malware. Scan taken on: Wed 9 Nov 2011 10:16:55 (CET) File Virus scan courtesy of: http://virusscan.jotti.org/en Looks clean. VERY COOL program Joe, where did you find this thing? I Like it alot because it shows the spectrum in real-time as the song plays. I played with the settings. In order to get the low spectrum to be a bit more defined I went to FFT (see I had the right idea lol... FFT = Fast Fourier Transform ) and changed the sample points to 32768 and the "window function" to either Hammin or None(rect), this makes the spectrum graph a bit slower because its averaging over a larger some of data but it yields accurate results, you can pause the song during bassy parts and mouse over the peaks to check the frequency. Unfortunately since it needs to take the average across such a large number of samples to gain detail in the low frequencies, it really only works for long basslines. Very neat program though. EDIT: Ah, program got alittle cooler, just noticed if you click on the "Measure Mode" button on the top (looks like a ruler with Hz over it and dB under it) you can then click on MAIN and PEAK on the bottom left to view a Peak Hold graph along with the real-time, and you can also view a moving average graph if you set the AVG pull-down list. It also has a little display for the Realtime "MAX" frequency while the song plays along with RMS and THD+N. SWEET program Joe. |
|
Silver Member Username: Joe1234Post Number: 851 Registered: May-09 | Glad you reminded me where found it, it's from a Japanese guy, and he has a beta 2 version which is much faster (thanks I now tested the smoothing and the sample size), just google "WaveSpectra" pick the first result and ask for a translation from google. Yes the plot above for "Fellin Like" was obtained turning on the "Peak" and off the "Main" and letting the song run to the end, it also has a N samples average. Hats down to this guy for releasing as freeware. |
|
Silver Member Username: AlonzoubPost Number: 285 Registered: Apr-10 | Yea no doubt... ive seen crap spectrum analyzers that aren't freeware, this one is gold. |
|
Silver Member Username: Joe1234Post Number: 907 Registered: May-09 | USDA - White Girl  WOW! |
|
Silver Member Username: MakinblakMonroe/Monticello, La/Ar Post Number: 786 Registered: Oct-10 | Don't know if you are still up to this. But if u r.... Roy Jones Jr-can't be touched paul wall-break em off wocka flocka-this is what I do They are in descending order of importance in case u don't wanna do all 3. Thanks in advance. |
|
Silver Member Username: AlonzoubPost Number: 326 Registered: Apr-10 | Here ya go, you didn't put times, so I guessed. +++ png +++ 699196 +++ Upload +++   |
|
Silver Member Username: AlonzoubPost Number: 327 Registered: Apr-10 |  |
|
Gold Member Username: Joe1234Post Number: 1016 Registered: May-09 | Paul Wall - Break Em Off.mp3  Peak on Red, random 800 sample average on blue. This plot shows that the song does hit but I GUESS that since the hits are so brief and spaced among themselves end up averaging down over a large sample as shown in the previous plot. |
|
Gold Member Username: Joe1234Post Number: 1019 Registered: May-09 | After some tweaking, here is a second look at: Paul Wall - Break Em Off 1) Linear amplitude. 2) Log amplitude.   Well it was just a matter of drawing on a log amplitude scale and one can see that the song does hit pretty hard on the 30's too as I expected! For those who care humans hear more like in the second plot. |
|
Silver Member Username: MakinblakMonroe/Monticello, La/Ar Post Number: 793 Registered: Oct-10 | That's sick man, thanks. I would have guessed that the low note on 'cant be touched' would b lower... but I guess not. It seems like u may have touched on this before, but what does the "Y" axis represent? I notice that 'break em off' tops off at .11 while 'this is what I do' tops out at .34... does this mean that flocka's song has louder, stronger Bass? |
|
Silver Member Username: AlonzoubOrlando, Florida Post Number: 343 Registered: Apr-10 | It does represent amplitude of the frequency... in a way. Short Answer: No you really can't compare the Y-axis value of different "hits" or "notes" to get a good bearing of the loudness of the hit/note. Longish Answer: Its not necessarily a good idea to compare the Y-axis value in different songs (or even same song/different note) to get the output level, this is because the way these programs acquire the frequency plot is through an algorithm called Fast Fourier Transform (FFT). If you have ever taken Calculus 2, you'd be (somewhat) aware that any "signal" or "function" can be represented by a series of Sine Wave Functions of varying amplitude and frequency summed to each other. This is called a Fourier Series. The Fourier Series is an approximation though, the length of the Fourier series can be VERY long. Music is nothing more than a VERY complex function or signal. The FFT algorithm attempts to deconstruct the signal into its fundamental Sine Wave functions each with their own Amplitude and Frequency. Since it is an approximation, and music is not made up of pure Sine Waves, the FFT algorithm will give slightly inaccurate output when deconstructing music. For example if you look at the plot I gave you for "Paul Wall - Break Em Off" you will notice that the "peak" around 36Hz has somewhat of a proportionally wider base, this is most likely because the "note" that hits around 36Hz isn't REALLY a 36hz Sine Wave, it is a complex low frequency signal that encompasses frequencies between about 33Hz to 40Hz. You could theoretically say the TRUE amplitude of that "note" would be the amplitude of all of those points between 33Hz-40Hz all stacked together. Same goes for every other note in every other song, Not alot of artists actually use PURE Sine Waves in the music. Hope that helped a littlebit |
|
Gold Member Username: JtownTexas Post Number: 4016 Registered: Mar-07 | Just curious, but if I'm reading this right in addition to what I've been told by past employers, why do people still insist on tuning at 32hz and below when the potential to get loud is actually more around 37-39hz for most songs? |
|
Gold Member Username: Joe1234Post Number: 1086 Registered: May-09 | Thieves.. there are songs that hit the 20s like: White Girl up above.. But box tuning is pragmatic, in a way like equalization, you set it up in the way you like it best.. it wasn't the intent of their creators but are used that way. |
|
Gold Member Username: JtownTexas Post Number: 4019 Registered: Mar-07 | Joe...I never said there weren't songs that went that low. Most of my previous boxes, which I've posted pictures of on this site, have been 34 hz and below because I listened to a lot of Jeezy and chopped and screwed. I used to love the lows, but got tired of missing the rest of the Hz range. |
|
Silver Member Username: AlonzoubOrlando, Florida Post Number: 346 Registered: Apr-10 | Box tuning also has alot to do with the vehicle you put the box in. Cabin gain is most effective in a smaller car therefore tuning higher in a small car could still yield good lows. In my case (my box is tuned to 30Hz) If I would have tuned higher the box would have been too peaky for my taste. No I don't listen to TONS of music with LOWLOW content, but tuning doesn't JUST effect the low frequency extension, but output throughout the entire subwoofer "band". |
|
Gold Member Username: Joe1234Post Number: 1091 Registered: May-09 | Thieves so I see what's the problem, when you tune low or use higher diameter subs you use midbass drivers to fill in the gap. Also sealed blends in much nicer to standard components than ported but then you loose low end of course. |
Main Forums
Today's Posts- Home Audio Forum
- Home Video Forum
- Home Theater Forum
- Car Audio Forum
- Accessories Forum
- All Forum Topics