DIYMobileAudio.com Car Stereo Forum banner

Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

133K views 1K replies 70 participants last post by  cycfari 
#1 · (Edited)
External link: APL1

Want to start off this review by saying that this is without doubt one of the better products I've come across for a long time. Many thanks to "Raimonds" here on the forum, the designer of this product - who introduced and let me demo this product.

APL1, basically an advanced equalizer which utilizes FIR (Finite Impulse Response). It can form any curve you desire with very few restrictions, in addition to that it also corrects phase deviations in all minimum phase regions.

Instead of reciting the entire spec I'll just copy/paste it from the official website.

APL1 – Unbeatable Quality in Pocket Size

APL1 is a FIR based 2 channel equalizer with 4096 coefficient FIR filters per channel allowing curve resolution of up to 6Hz. It is capable of creating any arbitary Amplitude Frequency Respone (AFR), Phase Frequency Response (PFR) or Delay Frequency Response (DFR) suitable for very detailed and accurate corrections and equalization in critical applications such as studio monitors, headphones and car audio systems.

USB connection allows the upload of correction filter files to APL1 unit even in the middle of a live performance without any disturbdances in sound. The unit can optionally have a 16 preset memory accessible trough an external, freely placeable rotary switch, excellent for car audio installations.
APL1 is available in three different configurations:

1. RCA model for car usage – 4 RCA connectors for analog in/out, 2 – for SPDIF digital coaxial in/out, 2 TOSLINK digital in/out

2. XLR model – 4 XLR connectors for analog in/out, 2 RCA– for SPDIF digital coaxial in/out, 2 TOSLINK digital in/out

3. XLR model with optional AES/EBU digital in/out - 4 XLR connectors for analog in/out, 2 RCA– for SPDIF digital coaxial in/out, 2 TOSLINK digital in/out; 2 XLR for AES/EBU digital in/out

Specifications:

*Number of processing channels: 2
*Resolution: 48kHz/24bits
*Frequency resolution of correction curve: 6 Hz, 4096 coefficients at 24 kHz bandwidth

*Inputs: 2 analog balanced, digital - SPDIF coaxial and TOSLINK optical, AES/EBU on order

*Outputs - 2 analog (XLR balanced), digital - SPDIF coaxial and TOSLINK optical, AES/EBU on order

*Input/output analog signal max levels: +10dBu (can be changed on order)
*Analog input/output connection types: RCA or XLR
*Analog input impedance: 10 kOhm
*Analog output impedance: 150 Ohm
*Common mode signal rejection (CMSR) of analog input – 90 dB
*Dynamic range (analog input/output):104 dBA.
*THD (analog input/output): 0,07%
*Signal delay: 1.6 ms (analog input/output), 2.4 ms (digital input/output)
*Control and upload interface: USB
*Interface computer system: Microsoft Windows
*Power: 6.5 to 16 V DC, 3.5 W.
*Dimensions 15х20х5.3 cm.
Let's begin with some pictures:

Good packaging...















Let's take a look inside:

















A little summary of the pictures. The components are high-grade as far as I can tell, ICs from Burr-Brown and ST. The circuit board is sturdy and it looks overall well engineered. I really like the USB connector, lots better than those small mini/micro USB connectors. Connections are gold-plated and feels sturdy, you don't break the board when you pull of those tight RCAs. Removable power connector for quick install/removal.

The impressive part is yet to come though!

Here's the software...



Does it look complicated or perhaps different? Let's "dumb it down" a bit.



A few notes here:

1. APL measures and applies a curve which YOU CHOOSE.

2. Any curve can be modified or created saved as a simple text file. APL will interpolate adjacent frequencies.

3. APL uses multiple measurement points to find out the true sound power response (direct sound + reflections) of the listening space. This is feature different from using pink noise and an RTA.


You can choose the "window" of the impulse response which basically determines the low-end accuracy. For car-audio use, Raimonds recommended 200ms length. You only need to choose the appropriate file and leave it at that. Nothing complicated.



Here's some "real life" use. Those who have read all my measurement guides and various threads involving DSPs should be used to my frequent use of graphs to display objective data.. Sooo - here's a few for you ;)

Initial setup - Simple 2-way passive + sub - No EQ (Crossovers at 100Hz 24dB/oct acoustic both ways). The null at 70Hz is a modal cancellation. Left/Right power response shown.



Previous setup - Basic EQ and a few shelf filters applied using a MiniDSP. Left/Right power response shown.



Some additional fixes after measurements with APL...



The APL software then calculates based on the correction and target response to give you this, (left channel shown / "semi-flat" setting):



Before I get into the subjective part. Let's continue with some additional software presentation:



This is a neat little "extra" program that calculates filters, target responses, EQ, corrections etc. It outputs your curve to a text file that APL software can read and apply in different ways. Of course you can make one for each channel.

*Gain 20dB +/-
*Q from 0.1 to 100
*Center freq from 5-48000Hz.
*Low/High shelf filters.
*Low/high pass filters.

Pretty neat. There's a large room for customization here.


Here's the actual software to hardware interface:



You can store 16 presets in the unit and it only takes a second to shift between different settings. You just choose the *.fir file generated by the APL software and import it into this interface and you're done. Enjoy ;)

As I always do I like to confirm results or at least compare. Before using APL I used RoomEQ to a large extent (it still very useful, don't get me wrong). Again, even if you're not that accustomed to measuring you probably know of some of the procedure involving it, you basically measure around your head and average it together. APL is using a different method. With APL you measure the entire listening space, not just around your head. As a result the "ordinary" method using RoomEQ (or some other measurement program) differs from APL's result. When I first noticed this, I was skeptical because I always found my method quite accurate and correlated well with what I heard...

The thing I first noticed was that APL's version of flat didn't at all sound "flat", as in "RTA-flat". This made me curious to see how APL derived and displayed the curve. Below are results from RoomEQ using spatial average measurement method (headspace), using RTA and Noise with infinite averaging method.







Well, this is interesting indeed. "APL Flat" isn't actually "RTA-Pink Noise flat". If you were to use white noise as excitation signal, the results are fairly consistent with "flat". Using pink noise, it does create a downwards tilted curve explaining how the difference between "flat" and "flat". Another interesting thing is that the right side (passenger side) is lower in amplitude than the left side. This is consistent with the listening experiments I did when I derived my "house curve/target response" in RoomEQ.

And... thirdly, look how good the curves sum. A testament that APL indeed performs proper phase correction between left and right.

What does this mean? It means that APL does indeed derives data and compensating correctly and it correlates exactly with what we hear.

Conclusion & Subjective analysis

*This is a very powerful unit, very different from the EQ available in the DSPs available for car use. Different in a good way, it can do things any "normal" DSP can't and it's easy and straightforward to use.

*T/A and L/R EQ issues are gone with this unit. The acoustic center is perfect, even a simple 2-way passive system like mine sounded like a competition grade car. Depth is improved subjectively. Everything sounds clearer, more transparent and so easy to listen at. No annoying frequencies that stand out, even crappy recordings that normally is un-listenable suddenly becomes tolerable. The center focus is insane, the up-front bass is improved by an enormous amount (in this car). With a crossover point at 100Hz/24dB both ways, the sub was impossible to localize with any material (any amount of processing won't fix resonances in the car or non-linear distortion so take that into account) . The image stayed up-front and kept staying there.

(I am aware that the result varies, if you already got an audiophile grade competition car the improvement is less than with a "normal" install but in my opinion it's still worth every penny. It's that good.)

*My noise-free preamp is still broken so I couldn't run I/O tests but I'll tell you, it's transparent as far I can hear. No noise whatsoever, I even maxed volume with a 0 bit track but nothing. Couldn't hear any coloration with the unit in "raw-passthrough".

*This unit does NOT replace an ordinary DSP, in fact you need an ordinary DSP as well. You place the APL1 unit before the DSP - 2RCA's in - 2 out. Not more complicated that that. If you don't own a DSP and is interested in this product, I recommend using it together with a MiniDSP (among the cheapest and best on the market).

*Support and feedback is great! The value in this is enormous with these types of products.

Overall a 10/10 score from me. Will use it in my main competition build...

Link to Raimonds profile; Car Audio | DiyMobileAudio.com | Car Stereo Forum - View Profile: Raimonds
 
See less See more
27
#43 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

The more I read about the details, the more I realize that I don't quite understand. I am going to download the plugin and see what I can do with it.

I have a laptop with foobar, a Dayton Audio iMM-6 Calibrated microphone, a head unit that has a line level input and the soon to be downloaded plugin. It there anything else I need?
 
#46 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

The more I read about the details, the more I realize that I don't quite understand. I am going to download the plugin and see what I can do with it.

I have a laptop with foobar, a Dayton Audio iMM-6 Calibrated microphone, a head unit that has a line level input and the soon to be downloaded plugin. It there anything else I need?
Except the APL software... that's it.
 
#44 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

Hanatsu,
Both yes and no. If 1 ms at normal atmospheric pressure corresponds to appr. 34 cm (little bit more, indeed), so you are right. You can backup fir-flter with the max delay of 4,797 ms.
It will be a new filter: ***-fir_4,797ms. Then you can use this new file to delay it further. New file name will be: ***fir_4,797ms_4,797ms. I mean that the last one has a delay of 4,797 x 2, which is about 350 cm. I can not see any restrictions for this process. Going to ask Raimonds about this issue.

I'm now trying to tune my front again as some things have been changed (HPF, LPF, crossover frequencies and the most important - frequency response of the both fronts). If I would still use old APL-corrections made onr year ago, the sound will be more bright than necessary. I will for sure make some tests of the delayed files with TDA software.

Regarding time alignment, phase control etc. - to my opinion the majority of people underestimate its importance as they think it is not audibe. I could easily hear the stage center moving by just applying different delays between left and right speakers. It is kind of rotating image, which has its own optimal position. Do not forget about midbass and sub. For some complicated subs it is required quite siginificant optimization ion the time domain (much more than 5 ms). Try to spend your time analysing results of tests with TDA - sure you will be surprised. By the way, the latest version of this soft: TDA 9.0 gives you even more information - distortion analysis.

When you open graffs in TDA, it is possible to use a cursor for determination of exact values needed to be corrected. APL's accuracy is rather high (compared to fx. min. delay steps in my HU - 0,1mc (3,4 cm)).
 
#45 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

Hanatsu,
Both yes and no. If 1 ms at normal atmospheric pressure corresponds to appr. 34 cm (little bit more, indeed), so you are right. You can backup fir-flter with the max delay of 4,797 ms.
It will be a new filter: ***-fir_4,797ms. Then you can use this new file to delay it further. New file name will be: ***fir_4,797ms_4,797ms. I mean that the last one has a delay of 4,797 x 2, which is about 350 cm. I can not see any restrictions for this process. Going to ask Raimonds about this issue.

I'm now trying to tune my front again as some things have been changed (HPF, LPF, crossover frequencies and the most important - frequency response of the both fronts). If I would still use old APL-corrections made onr year ago, the sound will be more bright than necessary. I will for sure make some tests of the delayed files with TDA software.

Regarding time alignment, phase control etc. - to my opinion the majority of people underestimate its importance as they think it is not audibe. I could easily hear the stage center moving by just applying different delays between left and right speakers. It is kind of rotating image, which has its own optimal position. Do not forget about midbass and sub. For some complicated subs it is required quite siginificant optimization ion the time domain (much more than 5 ms). Try to spend your time analysing results of tests with TDA - sure you will be surprised. By the way, the latest version of this soft: TDA 9.0 gives you even more information - distortion analysis.

When you open graffs in TDA, it is possible to use a cursor for determination of exact values needed to be corrected. APL's accuracy is rather high (compared to fx. min. delay steps in my HU - 0,1mc (3,4 cm)).
I agree...

If correcting for group delay in the subwoofer frequencies, we either require channel independent delay (not possible with 2ch), which means more than 1 board of APL.

OR

Can it do frequency dependent delay to correct a specific frequency range? If so, how's that done manually? In TDA? I'm hopefully trying out TDA this week ;)
 
#48 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

Raimonds,
Does it mean that if there is a dip of -10dB, correction for avoiding it will be -20 dB for the best results?
 
#50 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

An issue about use of two soundcards ( one for playback, another for recording) become actual.
Two different clocks are working in such case. And we will have time compression or expansion in our measurement data that depends of which clock is faster. Therefore the actual set of soundcards must be tested for that issue. APL Workshop Impulse Response tab is drawing all IRs in one graph. All of IRs must be on 5 ms with they max spike.



In case of time compression or expansion we will see IRs covering some time interval. It is ok if this interval goes under window (green curve).



If no, we have a problem.




The workaround for this is to accent first or last measurement (put the microphone more closely to the speaker/tweeter) that depends which clock is faster and you must find it by an experiment.
 
#53 ·
I'm curious as to how it removed your 70 Hz "dip". Did it just decrease the output above and below that dip to flatten the response?
Yes as I said, there was in fact large modal peaks above and below the "dip" ;)

Sent from my Samsung Galaxy S5 using Tapatalk
 
#55 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

Great news! Thanks. Even the first model of APL1 was not rather thick after my mind, but to get it smaller makes this unit more portable with the increased flexibility for installation.
 
#57 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

I have a convertible which is a bit challenging because the response changes drastically with the top up and down. Luckily, I drive the car with the top down 80-90% of the time.

For that 10-20% that I have my top up, is it possible to do two separate calibrations and switch on the fly?
 
#59 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

Yes, it is possible to do two separate calibrations (and not only two) and then switch on the fly. The switching takes about 1 second. The switching process is clear – without pops, noise or muting.
The switch is this:
BCKS1001 - LORLIN - SWITCH, ROTARY, HEX, NO STOP, IP65 | Farnell element14
The switch and flat cable of length you need are included in “16 preset” option.
 
#60 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

Yes, it is possible to do two separate calibrations (and not only two) and then switch on the fly. The switching takes about 1 second. The switching process is clear – without pops, noise or muting.
The switch is this:
BCKS1001 - LORLIN - SWITCH, ROTARY, HEX, NO STOP, IP65 | Farnell element14
The switch and flat cable of length you need are included in “16 preset” option.
Where do you connect the cable btw?

There's a 5-pin connector on the board somewhere? Interested in making a switch myself.
 
#61 · (Edited)
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

I've got my 16-position switch included in the package from Acoustic Power Lab along with APL1 unit. In my car it is installed now near the cigaret lighter - so it is easy to switch when driving.

Normally I use no more than 3-4 presets (one with "dispersion" solution for the tweeters). 6 positions are programmed for volume control when the CD-changer is connected to APL1 with the optical cord (from-24dB till -6dB).
 
#62 · (Edited)
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

I've got my 16-position switch included in the package from Acoustic Power Lab along with APL1 unit. In my car it is installed now near the cigaret lighter - so it is easy to switch when driving.

Normally I use no more than 3-4 presets (one with "dispersion" solution for the tweeters). 6 positions are programmed for volume control when the CD-changer is connected to APL1 with the optical cord (from-24dB till -6dB).
Does the same 16 position switch also control the inputs?

Are you using another DSP to control crossover so and TA? ( Is your system active?)
Thanks
 
#63 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

You can't switch inputs with it. 16 position switch is for changing between 16 presets, in other words - different target curves. Apl1 is only two channel device, but could be used together with every dsp. One of examples - aplaudio on Facebook. Here you can see the two diagrammes of the install in the champion car (2 first places in Eurofinals in 2014).

In my present car I am running passive front with TA from the HU. Almost finished a new passive 3way crossover (my own design) for new install.
 
#64 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

Some good questions:
- Can I use digital in and provide an analog (RCA) output? In other words, are the inputs and outputs assignable?
- If I run an active setup, what should I look for in a processor? Is time alignment and level adjustment sufficient or would more EQ help? Or should I just leave it flat and let the APL1 do its thing?
- What is involved for the measurements for the driver seat position? In other words, how many measurements?
1. All three outputs (analog, SPDIF, TOSLINK) are working in parallel and simultaneously.
The active input depends on fact is digital carrier present on digital input or not.
If yes - digital input is working. If not - analog input is working.
You can use APL1 as commutation unit between two sources - analog one and digital one such way.

2. You should tune up your processor as accurate as possible. The TDA software is good helper for this. http://www.diymobileaudio.com/forum...ain-analysis-measurement-software-review.html
You must critically evaluate FR curve measured by APL Workshop also. Are there possibilities to improve it by more accurate processor tuning?
Please keep in mind that it is not good practise to use EQ to correct processor tuning mistakes. Therefore the TDA was developed.

You should leave flat processor's EQ part and let the APL1 doing all eqing tasks.
The switching on even one of processor's parametric dramatically degrades quality for most of ordinary processors ...

3. It depends on your driver directivity. I have seen setups where directivity issues was not presenting and was not need to work on driver seat position.
In case of strong directivity you should run additional measurement to deal with that directivity.
You can chose number of measurements 25 ...50 ...100 ...200. But even 200 takes just 1 minute ( 3 measurements per second). The 50 ... 100 usually are sufficient.
 
#65 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

Some good questions:

3. It depends on your driver directivity. I have seen setups where directivity issues was not presenting and was not need to work on driver seat position.
In case of strong directivity you should run additional measurement to deal with that directivity.
You can chose number of measurements 25 ...50 ...100 ...200. But even 200 takes just 1 minute ( 3 measurements per second). The 50 ... 100 usually are sufficient.
For the on-axis measurement, is it being used as additional measurment or as the corrective ?
 
#66 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

On-axis measurement is used as a correction to the total EQ, not as a separate setup
 
#72 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

It would seem to me the main benefit comes from how one ALP1 works for the entire frequency range. Its ability to smooth out crossover points being a very big part of what it does.
I'd think you'd want all your source signal to go through one ALP1, then to your DSP for Time Alignment and crossover.
 
#73 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

Im wanting to run active and would like to use the APL. It does not make sense not to apply it at all frequencies as Fishman also said.

From the APL into my DSP and then on to amplifiers.

Raimond my question to you sir is what product do you offer to apply APL to my build across all Channels that will be controlled in an active manner. Do I need 3 or 4 of the 2 ch devices? Is there a 6 or 8ch option?

I will have one amplifier per driver in my truck. Each mono amplifier on tweeters, midbass and midrange will be fed a signal of L/R accordingly.
 
#75 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

Im wanting to run active and would like to use the APL. It does not make sense not to apply it at all frequencies as Fishman also said.

From the APL into my DSP and then on to amplifiers.

Raimond my question to you sir is what product do you offer to apply APL to my build across all Channels that will be controlled in an active manner. Do I need 3 or 4 of the 2 ch devices? Is there a 6 or 8ch option?

I will have one amplifier per driver in my truck. Each mono amplifier on tweeters, midbass and midrange will be fed a signal of L/R accordingly.
The recommendations earlier in the thread are to let the apl do its work in stereo.....as in directly after the source, either digitally or via rca.....then allow your processor to split the signal with crossovers and allignment. Then you only need one two channel apl unit for the entire system.
 
#78 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

SO what exactly does the APL unit do ?

Is it just a measurement device or does it provide some form of filtering/manipulation ?
As i understand, it provides quick Auto EQ (4096 bands of it), and some minor phase correction for a near perfect tune.
Post 11 & 13 seem to explain it best to me.
 
G
#81 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

So, you enter your desired final EQ curve into APL software.... The processor then creates that curve in real time.... ?

Very nice...
 
#82 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

The good question:
Why is a FIR based EQ better than others ?

Because you can create any curve you like and can do that instantly.
With ordinary parametric EQ you can only approximate to curve you need. And you may spent hours for that.
FIR also allows create phase/delay curves independently from an amplitude curve.
And FIR has much higher processing quality than ordinary IIR.
You can sum any number of curves (parametrics) and run them in one FIR filter.
 
#85 ·
Re: Comprehensive Review: " APL1 " - Advanced DSP / EQ / Phase Correction Unit

Raimonds,

What is the difference between the APL workshop and TDA EQ?
The APL Workshop software is based on Sound Power evaluation and it takes special measures in some „not simple” cases – 1) tweeters have serious directivity issues 2) the complex sub is used and it is introducing very large delay in its band. The solution for first case is published few posts higher. You should use large Time Window – 100 ms, 200 ms (and may be curve editing) for second one.
The APL TDA EQ software is exploiting incredible time resolution of TDA processing and is free from two drawbacks of APL Workshop mentioned. But it is not as stable as Workshop. It was developed as a tool for near field application that is free from very early reflections. The car environment is opposite. It has very serious early reflections that can ruin the result of TDA EQ work. Therefore it is possible to get very exceptional result by use of TDA EQ in car environment and it is possible to go into treble. Some examples for a car environment.
1. A pretty nice result. First image – before eq, second – after EQ





2. A nice result also but we can see an excessive phase added on crossover points caused by an improper initial tuning of crossover.






3. A problematic result. Midrange and HF drivers have strong side lobes of directivity (not clear piston motion, surface waves on driver membrane) that are „ caught” by early reflections and than is „amplified” by TDA EQ.



 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top