Loopback Method Trying To Get It

[Trdat]

The simplest concepts are often the hardest, I still don't get the loopback method.

But thanks to this random post I finally got half of it.

Using REW with Focusrite Solo and Dayton EMM-6 for speaker design setup questions.

Hello, For the last year I had been using REW with a UMIK-1 for measuring drivers to import into crossover design software to build speakers, but the timing info was not perfect and the phase between drivers was not always giving me the results in real life the crossover design software said I...

www.avnirvana.com

But then you answer the post by saying to use another mic so I am alittle confused if this method works.

So I essentially connect the front trs socket on my Focusrite Solo to the back. Can the cable be TRS at the front with RCA into ao TRS adapter on the back or even vice versa or does the cable have to be headphone jack into a TRS adapter on the back as well?

And if I use one of the inputs in the back how do I get stereo out to the amp? DO I need a splitter?

Laslty, whats next calibration of some sought?

And do i use the method similar to that of acoustic timing reference in which the sound beeps from one speaker and then measures through the other?

 

[jtalden]

• It works very well.
• No need for a new mic. The EMM-6 is fine.
• Place a TRS/TRS or TS/TS cable connecting channel 2 output to channel 2 input. Do not use a TRS to 2 RCA adaptor. It will not work as it cancels the balanced signal. There are other adaptor choices that work, but a low-cost cable is all that is needed.
• It is recommended to measure and EQ each channel separately using channel 1 output. If you like, you can then sum the 2 measurements within REW using vector arithmetic after aligning the 2 impulses.
• It is not recommended to measure 2 channels at once. For bass range the SPL will be fine, but the upper midrange and treble SPL will likely be very distorted and thus misleading. This is the result of comb filtering if the mic is not exactly equidistant to both speakers. Again, a TRS to 2 RCA adaptor will not work. A TS to RCA adaptor followed by an RCA to RCA splitter will work.
• In REW preferences/soundcard choose the solo device and assign the channel 1 for measuring and channel 2 for reference.
• When measuring, be sure loopback timing is selected in the measure panel.

 

[Trdat]

• It works very well.
• No need for a new mic. The EMM-6 is fine.
• Place a TRS/TRS or TS/TS cable connecting channel 2 output to channel 2 input. Do not use a TRS to 2 RCA adaptor. It will not work as it cancels the balanced signal. There are other adaptor choices that work, but a low-cost cable is all that is needed.
• It is recommended to measure and EQ each channel separately using channel 1 output. If you like, you can then sum the 2 measurements within REW using vector arithmetic after aligning the 2 impulses.
• It is not recommended to measure 2 channels at once. For bass range the SPL will be fine, but the upper midrange and treble SPL will likely be very distorted and thus misleading. This is the result of comb filtering if the mic is not exactly equidistant to both speakers. Again, a TRS to 2 RCA adaptor will not work. A TS to RCA adaptor followed by an RCA to RCA splitter will work.
• In REW preferences/soundcard choose the solo device and assign the channel 1 for measuring and channel 2 for reference.
• When measuring, be sure loopback timing is selected in the measure panel.

Well I got the Behringer with the Focusrite solo but technically same thing.

I had a feeling the RCA adapter wont work, so I have a headphone cable one side I can use a TRS 3.5mil adapter the other side a 2.5 TRS adapter that fits onto the jacks and then into the soundcard I think that should work.

So the front TRS input connects to the back left or right output? Or doesnt matter as long as I chose the correct option in preferences?

So I connect the remaining output from the back to the amp and that means technically 2 channels is not needed as I am measuring one driver at a time.

To get 2 channels at once, I can get a TRS adapter to XLR and split the XLR into the back of the amp but I really do not need this as I am measuring one driver at a time. But that means that I dont get the beep signal like I do in acoustic timing reference in which the signal beeps out of one side then measures the other?(you mentioned this in one post but I suppose that the loopback is another method)

Assinging is what I needed to know, but im confused. In preferences I chose the loopback channel as the channel that is linked with the loopback cable(left or right channel) from front to back and the other channel is for timing reference output, I think I got this right yeh?

Got the last one.

Actually, I found a post with John mentioning that you have a few responses on how to work out delays between drivers so I have them all open and working through them. Just trying to get this part right so I can start measuring then I will work out alignment tool in REW or potentially tune real-time. Slowly but surely. Appreciate your response.

 

[jtalden]

Well I got the Behringer with the Focusrite solo but technically same thing.

Okay

I had a feeling the RCA adapter wont work, so I have a headphone cable one side I can use a TRS 3.5mil adapter the other side a 2.5 TRS adapter that fits onto the jacks and then into the soundcard I think that should work.

?? Okay so long as it gets the signal from output to input.

So the front TRS input connects to the back left or right output? Or doesnt matter as long as I chose the correct option in preferences?

Yes, but why would we cross channels? Hmm, maybe I am the only one that gets easily confused.

So I connect the remaining output from the back to the amp and that means technically 2 channels is not needed as I am measuring one driver at a time.

Yes

To get 2 channels at once, I can get a TRS adapter to XLR and split the XLR into the back of the amp but I really do not need this as I am measuring one driver at a time.

Yes

But that means that I dont get the beep signal like I do in acoustic timing reference in which the signal beeps out of one side then measures the other?(you mentioned this in one post but I suppose that the loopback is another method)

You are selecting the loopback timing mode (cable) not the acoustic timing mode thus no audible beep.

Assinging is what I needed to know, but im confused. In preferences I chose the loopback channel as the channel that is linked with the loopback cable(left or right channel) from front to back and the other channel is for timing reference output, I think I got this right yeh?

No? I think you misstated this.
Follow post-2:
Channel 1 = left I/O >> for mic and signal to amp/speaker. (measurement channel)
Channel 2 = right I/O >> for the timing reference signal - has the loopback cable connected. (reference channel)
However, you can optionally swap channels since you have mic inputs on both channels (the Solo doesn't)

There are other ways to connect as in crossing the channels, but why do that? Just make sure you set REW preferences accordingly.

 

[Trdat]

Hello JTalden,

To be honest you were very clear which channel goes to which channel I was just confused because on the rear there is a left and right not channel 1 and 2 but I think I got it.

So, Channel 1 upfront is to the mic and and left channel (left on the back) is out to the amp.
While Channel 2 on the front(TRS) is connected as loopback to the right channel on the back.

Then, in preferences when you say channel 2 for reference that means the "right channel" is chosen for "timing reference output" is that correct?

Then I am confused because what channel is chosen for loopback option in preferences "right channel" again yes? Because it mentions loopback input which is the front 2nd TRS channel considered as the right channel.

 

[jtalden]

Yes.
Typically:
Drivers = ASIO
I/O Devices = Behringer USB AISO (or similar)
Output = 1 or left
Input - 1 or left
Timing Reference Output = 2 or right
Loopback Input = 2 or right

You can post a screen capture of the preferences/soundcard panel if you are unsure, or you can just make a test measurement to confirm it is working correctly.

 

[Trdat]

Yes.
Typically:
Drivers = ASIO
I/O Devices = Behringer USB AISO (or similar)
Output = 1 or left
Input - 1 or left
Timing Reference Output = 2 or right
Loopback Input = 2 or right

You can post a screen capture of the preferences/soundcard panel if you are unsure, or you can just make a test measurement to confirm it is working correctly.

I got it. Thanks so much, really apprecate it. I cound't find a 2.5ml TRS adapter so I have to wait for some cables and adapters to arrive then I can give it a try.

But just so I can get my head around what comes next in my goal of time alinging/setting delays in my Hypex Design Filter I just want to confirm the overall process.

1.Take measurements of the tweeter and mid woofer seperately and also from the acoustic centre from 1 metre away with the loopback method.

then I have the option to use the alignment tool in REW.
2. So I upload the seperate files of the tweeter and midwoofer and align them looking for the biggest reverse null and that would be the best alignement. And use the delay from there? How does the alignment tool give me the delay?

Or another option would be to use some type of real-time tuning using something like opensoundmeter or even REW which produces a pink noise and I could see in rea-time where the reverse null is.

I admit I don't know too much details of the steps of the 2 options I have given, I am still learning but am I think I am on the right track?

 

[jtalden]

Yes, that is basically the process.
Activate your chosen XO in the Hypex (I have no information of that process), then take the measurements of the MW and TW. You can use full range measurements for both drivers. Use a minimum of 50-10k Hz wide to assure the entire XO range is well covered in both measurements.
For timing analysis there are several methods that work well so you can use the one you like. There are YouTube videos on several different methods. The REW alignment tool is great for this purpose, but it may not be the easiest to understand. It's the method I use, and I can help with questions. I can also confirm your timing and/or suggest a favorable setting.
I posted a couple of files here. You can use them to practice how to use REW controls in the alignment tool if you don't have you own measurement yet. That thread is about SW XO timing, but the process is the same. Your higher XO frequency and 1 m mic placement will make it very easy to clearly see what is happening to SPL and phase as timing is adjusted. A good target is to maximize SPL through the XO range and have the phase of both drivers closely follow each other through the XO range.

 

[Trdat]

Activate your chosen XO in the Hypex (I have no information of that process), then take the measurements of the MW and TW. You can use full range measurements for both drivers. Use a minimum of 50-10k Hz wide to assure the entire XO range is well covered in both measurements.[/QUOTE]

Crossover slopes are set in Hypex just awaiting to take measurements with loopback.

For timing analysis there are several methods that work well so you can use the one you like. There are YouTube videos on several different methods. The REW alignment tool is great for this purpose, but it may not be the easiest to understand. It's the method I use, and I can help with questions. I can also confirm your timing and/or suggest a favorable setting.

I will check out other methods and will work on calibrating in realtime in future but for now it would be great to learn the alignment tool in REW, I suppose I can learn a lot just from the tool itself. I will get back to the thread or a new thread once I get there to get some help for you to confirm a favourable setting thanks.

I posted a couple of files here. You can use them to practice how to use REW controls in the alignment tool if you don't have you own measurement yet. That thread is about SW XO timing, but the process is the same.

That is great, I will take a look at them and try teach myself. I get that SW alignment process will be just about the same as aligning tweeter and mid woofer.

Your higher XO frequency and 1 m mic placement will make it very easy to clearly see what is happening to SPL and phase as timing is adjusted. A good target is to maximize SPL through the XO range and have the phase of both drivers closely follow each other through the XO range.

I thought we were looking for the biggest reverse null not the highest SPL through the XO range?

And lastly any threads on basics on phase interpretation?

 

[jtalden]

I thought we were looking for the biggest reverse null not the highest SPL through the XO range?

That is one method that can work fine. It is just reverse of saying 'find the most SPL support' (without the polarity reverse). They should agree.

And lastly any threads on basics on phase interpretation?

With the mic on the listening axis for the measurements, check to find the closest phase tracking through the XO range. It will assure that you are not 1/2, 1, 1-1/2, or more away from the preferred timing. The SPL and null can look very good at those points, but the 2 drivers are offset on different cycles and thus the horizontal dispersion is not as uniform. The side lobes are impacted, i.e., there is more variability in the horizontal SPL. This may not be critical for sound quality, but there is no reason not to do it right. With a 2-3 kHz XO the job is very easy to get right.

 

[Trdat]

I can't believe I have to ask a stupid question like this but how come the alignment tool is not showing up in my control? Every video and example has it there.

 

[John Mulcahy]

Are you looking in the All SPL graph controls?

 

[Trdat]

Are you looking in the All SPL graph controls?

Got it! Thanks.

 

[Trdat]

So I imported the files and gave it a go in the alginment tool.

I put the cursor at the crossover point exactly where I want the frequency to flatten and then chose "align phase slopes at cursor" which must be the easy way of doing it but I am presuming that I should do it manually to see what I am actually trying to attempt.

Few questions,

1. You have two roll off slopes, are they meant to be simulating a frequency repsonse? As in I will be chosing the frequency response of the tweeter and the woofer within the alignment tool when measured with loopback...?
2. You also have the aligned sum on one of the examples is that the what I am trying to emulate when trying to align in the alignment tool?
3. If so, what am I trying to align? A flat SPL? What am I trying to match ? Is it the bottom graph that has those funny lines are they trying to line up...?

 

[jtalden]

1. You have two roll off slopes, are they meant to be simulating a frequency repsonse? As in I will be choosing the frequency response of the tweeter and the woofer within the alignment tool when measured with loopback...?

Yes, the 2 responses represent an idealized case of a SW and front main XO. The idea was just to practice a little to understand the alignment tool controls without distraction of using real data.

2. You also have the aligned sum on one of the examples is that the what I am trying to emulate when trying to align in the alignment tool?
3. If so, what am I trying to align? A flat SPL? What am I trying to match ? Is it the bottom graph that has those funny lines are they trying to line up...?

Yes, it is intended to clarify what an idealized alignment looks like, i.e., maximum SPL support and phase overlap throughout the XO range. Real data will be a compromise situation.

 

[Trdat]

I am a bit slow with this because I have been using Audiolense DSP automatically aligning everything for me now I am forced to learn as my centre speaker needs manual aligning with Hypex,

So after some playing around with both your examples and with my measurement of my tweeter and midwoofer without loopback,

It seems the idea is to match the tweeter and woofer step response or phase(if it is a step response) on the bottom graph with them falling exactly on top of each other. This is what I got with the examples you gave. I am presuming this is correct?

In the bottom graph brown is the mid woofer blue is the tweeter and the black? Is the black what we are trying to match?

But, trying to match the 3 lines in the bottom graph doesn't necesseraly give you the flattest spl right?

So are we trying to find the perfect balance between these two? Is this what you mean by comprimise situation?

Lastly, the delay I get will be delaying the tweeter right even though the delay amount is given for the woofer in the alignment tool?

 

[jtalden]

It seems the idea is to match the tweeter and woofer step response or phase(if it is a step response) on the bottom graph with them falling exactly on top of each other. This is what I got with the examples you gave. I am presuming this is correct?

Yes, and it is a phase response (a step response is different).

In the bottom graph brown is the mid woofer blue is the tweeter and the black? Is the black what we are trying to match?

Yes? But, more correctly, the black does not show a 'target'. It shows the SPL and phase that will result if a measurement is taken using the timing chosen in the alignment tool.

But, trying to match the 3 lines in the bottom graph doesn't necesseraly give you the flattest spl right?

Yes. It does in this case as it was constructed to that as an example. There are many factors that impact the result of a real measurement situation; XO filters chosen, driver characteristics, box design and room effects. In this case and in real measurement situations finding the closest tracking of the of the phase results is the most SPL support in the XO range. That is normally the best situation.

So are we trying to find the perfect balance between these two? Is this what you mean by compromise situation?

I mean by 'compromise situation' that we want to approach this phase tracking situation as closely as possible. The other factors impacting our measurements will prevent a nice clean situation like this, but we want to approach it as closely as possible. That means the phase should cross at, or very near, the XO frequency and the phase follow each other as closely as possible to either side of the XO frequency.

Lastly, the delay I get will be delaying the tweeter right even though the delay amount is given for the woofer in the alignment tool?

It doesn't matter which driver timing is changed to create the proper timing relationship for the 2 drivers. The Hypex probably has no delay on either driver by default, so it is normal to increase the delay on one that creates the proper relationship. For my typical 27mm tweeter and 180mm mid-woofer on a flat baffle the mid-woofer needed a very small delay.

You seem to be struggling a little on the basics. If this is a one-time job, i.e., you are not needing to learn how to do this for future hobby work and just want to get the job done, I can provide the timing for you. You don't even need to use loopback timing in that case. If I am given a mdat file with full range measurements of the TW, MW, and TW+MW with mic placed anywhere between 0.5-1m away, I can determine the current timing relationship and also what timing change is needed to achieve the most favorable condition.

 

[Trdat]

It doesn't matter which driver timing is changed to create the proper timing relationship for the 2 drivers. The Hypex probably has no delay on either driver by default, so it is normal to increase the delay on one that creates the proper relationship. For my typical 27mm tweeter and 180mm mid-woofer on a flat baffle the mid-woofer needed a very small delay.

You seem to be struggling a little on the basics. If this is a one-time job, i.e., you are not needing to learn how to do this for future hobby work and just want to get the job done, I can provide the timing for you. You don't even need to use loopback timing in that case. If I am given a mdat file with full range measurements of the TW, MW, and TW+MW with mic placed anywhere between 0.5-1m away, I can determine the current timing relationship and also what timing change is needed to achieve the most favorable condition.

Yes, I am struggling I have always wondered how many years or what does it take for a DIY guy to learn these concepts I have been at it for about 5 years but 3 years pretty hard. My efforts are divided between speaker design and acoustics so that is my excuse. But I think it is because I have never tried learning about phase in the SPL graph which should be a prerequisite before attempting what I am now. So I will read up on how to interpret phase first.

In this case and in real measurement situations finding the closest tracking of the of the phase results is the most SPL support in the XO range. That means the phase should cross at, or very near, the XO frequency and the phase follow each other as closely as possible to either side of the XO frequency.

I think these two sentences are key and this is what I am struggling to understand. A video might help so I will take a look in youtube. I think once I read up on phase I should be able to pick up on what these actually mean in practise.

It is for future hobby work and I am trying to learn vituixcad as well so I am going to persevere. I still have a lot to learn, off axis measurements, vituixcad, and apart from this acoustic modeling software. But in saying that I have analysed my tweeter's polar plot with the waveguide and have the ideal crossover point all that is left is the delay and my audyssey will give me some frequency correction for a half decent centre speaker till I manage to use vituixcad to enter biqauds in the hypex DSP which is my ultimate goal.

I know it will be a mild delay.

For next time I will choose drivers that have the FRD files with off axis measurements or polar plots to make things easier, I suppose not everyone has to go through taking measurements but while I am on this process I will keep going that way I can officially design my own speaker one day. My tri-amped system uses audiolense, superb sound, good directivity, perfect drivers but kind of cheated the process. Ha ha... But that was rocket science in itself. Lol!

So I need to take measurements of the tweeter and mid woofer seperately and then the tweeter and mid woofer together and while your at it would it be possible for a delay between the woofer and mid woofer? So that means a seperate measureemnt for the woofer and with all 3 drivers together or the woofer doesn't need phase alingment? I do have subs so the woofer in the 3 way will play between 80hz to 350hz.

I'll take these measurements and get in touch. I would really appreciate it.

 

[jtalden]

Thanks for the background information. It helps me understand better your objectives and situation. You are much further along in the process than I was guessing. I like that you choose the XO frequency with dispersion matching as a consideration. The dispersion will be retained horizontally off-axis more uniformly using the correct timing, so it's good you considered this.

Okay, I missed, or forgot, that this was XO timing needed for 3-way mains and also for SWs. I suggest you continue to get loopback timing working for your XO timing efforts. It will be helpful here and in your future work. I suggest we address these XOs in separate steps in consideration of file size and ease of understanding, [For a given finalized room setup all measurements could be taken in one session with mic at the LP. It can be done, but it is not the way to learn the process.] Let's continue to use this TW/MW XO as a good first example. Once it's sorted, I can continue help with the other XOs as needed.

If we know our room setup geometry and want to be pedantic, we would:

Setup

• Mic located on the intended listening axis to point between the two drivers and placed a distance away from the baffle maybe 4-5 times the diameter of the larger driver (MW driver in this case)
• REW loopback timing active
• Full range sweeps for all measurements

Measure

1. TW
2. MW

That is all that is needed to determine timing with highest accurately. [If we are more practical or don't know the room setup geometry, we can place the mic on the TW axis as is typically done commercially.]

I would still like to determine the timing needed for the TW/MR XO prior to your having loopback timing available if you still want to do that. The above process is the same except a 3rd TW+MW measurement is needed, i.e., TW and MR together. I haven't used my technique for doing that on a higher frequency XO yet but expect it to work just as well.

 

[Trdat]

Thanks for the background information. It helps me understand better your objectives and situation. You are much further along in the process than I was guessing. I like that you choose the XO frequency with dispersion matching as a consideration. The dispersion will be retained horizontally off-axis more uniformly using the correct timing, so it's good you considered this.

Okay, I missed, or forgot, that this was XO timing needed for 3-way mains and also for SWs. I suggest you continue to get loopback timing working for your XO timing efforts. It will be helpful here and in your future work. I suggest we address these XOs in separate steps in consideration of file size and ease of understanding, [For a given finalized room setup all measurements could be taken in one session with mic at the LP. It can be done, but it is not the way to learn the process.] Let's continue to use this TW/MW XO as a good first example. Once it's sorted, I can continue help with the other XOs as needed.

If we know our room setup geometry and want to be pedantic, we would:

Setup

• Mic located on the intended listening axis to point between the two drivers and placed a distance away from the baffle maybe 4-5 times the diameter of the larger driver (MW driver in this case)
• REW loopback timing active
• Full range sweeps for all measurements

Measure

1. TW
2. MW

That is all that is needed to determine timing with highest accurately. [If we are more practical or don't know the room setup geometry, we can place the mic on the TW axis as is typically done commercially.]

I would still like to determine the timing needed for the TW/MR XO prior to your having loopback timing available if you still want to do that. The above process is the same except a 3rd TW+MW measurement is needed, i.e., TW and MR together. I haven't used my technique for doing that on a higher frequency XO yet but expect it to work just as well.

Apologize for the late reply, you always reply in a timely manner. Yes, my situation is all over the place, a bit advanced somewhere and totally weak somewhere else.

Thanks so much for your help, we can go one step at a time and stick TW and MW then woofer and then hopefully I can manage myself.

I will wait to take loopback measurements, the above doesn't seem to difficult I can manage that.

For now you mentioned you still would like to see the measurements without loopback so have added a mdat file of the measurements that I have already taken without loopback, each measuerement has a description. Let me know if there okay for your technique I did read in another thread you have your own way of figuring out the delay, I think that is what your getting at.

 

[jtalden]

I cannot use this data for XO timing. The measurements provided are not the ones required. All 3 measurements must be taken with; the mic located at the same position, the Hypex XO active and with all EQ disabled. It's no problem if you prefer to wait until you are setup for REW loopback measurements to do this.

I have no experience using Audiolense or other commercial DSP software. I only have limited experimentation with DIY FIR EQ filter usage, but my approach in your case would be to first set that center speaker TW/MW XO timing in the Hypex using REW alignment tool with no EQ active. Audiolense can then be run to EQ and time align this 2-way portion of your 3-way center speaker along with all the other independent portions of your setup. When Audiolense is run the proper driver TW/MW timing in the Hypex will be retained. The result will be a more uniform horizontal off-axis response across that XO frequency range. This has been my understanding of your objective.

 

[Trdat]

I cannot use this data for XO timing. The measurements provided are not the ones required. All 3 measurements must be taken with; the mic located at the same position, the Hypex XO active and with all EQ disabled. It's no problem if you prefer to wait until you are setup for REW loopback measurements to do this.

I have no experience using Audiolense or other commercial DSP software. I only have limited experimentation with DIY FIR EQ filter usage, but my approach in your case would be to first set that center speaker TW/MW XO timing in the Hypex using REW alignment tool with no EQ active. Audiolense can then be run to EQ and time align this 2-way portion of your 3-way center speaker along with all the other independent portions of your setup. When Audiolense is run the proper driver TW/MW timing in the Hypex will be retained. The result will be a more uniform horizontal off-axis response across that XO frequency range. This has been my understanding of your objective.

Yes, my apologize the mic was moved to each driver I totally forgot about that sorry. But the hypex crossovers were in place with no EQ.

We can wait for the loopback but I always take measurements anyway if I get around to measuring my main system I will also measure my centre speaker, if my trs cable arrives soon i'll just send the mdat file with loopback So, all three measurements with the mic exactly the same place, got it.

I will not be using Audiolense for the centre speaker, unfortunately without a seperate DAC the software can't detect the centre speaker as it is set up through the home theatre. Essentially, all the FIR filters will done through the hypex filter design software so its up to me to manually input biquads. Audiolense would have been the easy and effective way but it would have only worked for music and not through the home theatre.

Actually, I had set it up through a seperate DAC for a 5.1 for music it was a long and laborious process of a year or so but I didn't like the surround for music, I prefer 2 channel for music so now I am setting up the biquads so the centre speaker can be run through the home theatre and use it for a centre voice channel.

 

[jtalden]

Okay, I'll be here whenever the measurements are ready for the TW/MW XO timing analysis. When that is done, we can then address the MW/W XO timing if you like.

 

[Trdat]

Hey jtalden,

So, I got the hang of it.

I measured from the exact position for TW and MW and again for both together from about 50cm away. I was a bit confused on exactly where to measure from as listening position was mentioned as well as 4 to 5 times the diamter of the large driver which is about 50cm.

Please let me know if there is any anomalies, I am happy to retake measurements for it to be accurate. I double checked that all I had in the hypex design filter was the crossover slopes, no delay, gain just the crossover points.

My only confusion was in the measurement with both the TW and MW together I didn't see crossover dip at the 2200 crossover point?

Anyway, let me know if the measurement was okay.

 

[jtalden]

Is the correct file? I was expecting 3 measurements, not 9.