Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Nyal Mellor]

I'm sure someone else has gone through this. What I am trying to do is play REW through JRiver so I can calibrate my system (adjust distances, levels, eq, etc).

I have REW set up using Java engine, sending signal to the default device so it goes through JRiver loopback. Recording is via a USB soundcard.

JRiver is setup to play through a Lynx AES16e.

I can get sweeps to playback through my system but there are ticks, pops and even when the sweep sounds clean the results in REW are all over the place i.e. not consistent from one sweep to the next which tells me something is not right.

Any ideas?

[Matt]

I don't have much experience with REW, although I know a lot of users like it.

However, for basic correction you could use DSP Studio > Room Correction.

Set distances first (with a measuring stick), then set levels using a decibel meter (or an app on your phone) and MC > Tools > Advanced Tools > Audio Calibration.

For more advanced frequency based correction, you might consider the third-party packages for doing convolution based correction (they build the filter and you use it in DSP Studio > Convolution).  mitcho wrote a nice series of articles on this approach at computeraudiophile.com.

[mojave]

It looks like you have everything setup properly so I'm not sure where the issue lies. I have ran REW sweeps through JRiver in the past without any issues. I just got a Lynx AES16e and a Lynx Aurora 8 is arriving tomorrow. I plan to use REW on Saturday or Sunday afternoon to perform some measurements as I get ready for the Iowa Spring Audio GTG. I'll let you know how it works out. I will be using either my Steinberg UR824 or Tascam US144MKII as a USB device for input.

Have you made sure you are using the lowest latency buffers that still work?

[Nyal Mellor]

Are you referring to the ASIO buffers in jriver? How should these be set? I have not changed them from defaults (.5s, large hardware buffers)

[mojave]

I use .07 or .1 and uncheck "Use large hardware buffers." However, on most devices the "Use large hardware buffers" doesn't do anything since the device's own ASIO drivers control the buffer size.

In the AES16e control panel try selecting different buffer sizes. I think if these are too large the latency may be too great for REW.

The input device is using WASAPI so there is nothing to adjust on that end.

I would suggest getting the loopback to work smoothly with something like a YouTube video played in a web browser. Start the loopback in JRiver and then play the video. If the audio plays fine then you know the loopback is working as intended. After you get this working well, then you can use REW.

I've requested at the REW forum that John make external test clips that can be played through JRiver or other software like the OmniMic clips, but that hasn't happened yet. It would actually be nice if they were integrated into JRiver's Audio Calibration section.

If you used the AES16e for input and output, I think you can then use the ASIO Line In in JRiver and eliminate the WASAPI loopback altogether. I'll use the Lynx AES16e and Aurora for examples. I could set REW to output via ASIO on Output 1 in the AES16e. In the AES16e mixer I could route this output back into Input 1. In JRiver I activate the ASIO line in and open Input 1. The sound now goes through JRiver and its DSP. In JRiver I would use the ASIO offset and set it to 9 so that the output goes out Output 9. Lynx told me I can have one software using outputs 1-8 (REW) and other software on outputs 9-16 (JRiver) simultaneously with the their ASIO driver. The mic would be connected via a preamp to the Aurora 8 and it would go through Input 2 back into REW. Now the sound has completed a loop all handled by the Lynx ASIO driver.

If you add a mic input section to the Metric Halo you should be able to do the same thing.

[Nyal Mellor]

I played around with it some more today. I think my main problem was that the default audio device being used for loopback was set at 44.1kHz whereas Room EQ Wizard was working at 48kHz. Fixed that and results are better and more repeatable.

My Lynx latency was set at 1028, I reduced that to 64, also reduced the JRiver buffer to 0.1s. Honestly the latency changes didn't seem to do much, it was more the sample rates were screwy.

However, there is still an issue with the REW results through loopback differing to the results when just measuring through REW. The measurements are different above about 2k. I did a bit more investigative work by playing pink noise in REW and using the analyzer in JRiver to visualize. It seems like the signal is getting through JRiver ok but something strange is going on with the recording side. The main thing is at least it's a consistent deviation and it's well out of the range I would need to look at anyway. Any ideas? The out soundcard is an internal Lynx AES16e and the in soundcard is a USB M-Audio MobilePre.

See attached image.

[natehansen66]

I use the loopback with REW all the time. Every once in a while I'll get an audio glitch but that's mostly when I'm doing linear phase convolution. I also use a MobilePre. Like you said having the sample rates all matched up is huge. The device you set to default will always resample, which I can't say I'm real crazy about but it is what it is. Try a .05 sec buffer. That's what I use. Are you resampling in MC at all? I know that can cause issues with loopback and convolution so it wouldn't hurt to leave the sample rate alone there.

[mojave]

I used REW yesterday and the loopback in JRiver. My measurements matched without any issues. I also used the same device for recording and playback (Steinberg UR824) and a different device for playback (Lynx Aurora 8 via AES from AES16e) and recording (Steinberg UR824). These differed slightly above 10,000 Hz. I'm not sure what is causing your anomaly in the measurements.

I am once again absolutely amazed how easy it is to get great sound using Audiolense vs manually entering filters and using REW.

[Nyal Mellor]

I used REW yesterday and the loopback in JRiver. My measurements matched without any issues. I also used the same device for recording and playback (Steinberg UR824) and a different device for playback (Lynx Aurora 8 via AES from AES16e) and recording (Steinberg UR824). These differed slightly above 10,000 Hz. I'm not sure what is causing your anomaly in the measurements.

I am once again absolutely amazed how easy it is to get great sound using Audiolense vs manually entering filters and using REW.

Yes I need to do some electrical measurements post dac to rule out any issue with the lynx card itself...

[Nyal Mellor]

I did some more testing. It seems like it is a MobilePre USB / Lynx interaction issue.

I ruled out the DAC being defective by testing with the Metric Halo and DEQX...same issue using REW-JRiver-Lynx-DAC-MobilePre USB-REW

I ruled out the Lynx being defective by doing a test purely through REW (lynx digital out-digital in).

I also ruled out JRiver by doing a test from REW-JRiver-Lynx digital out-Lynx digital in-REW.

See attached..note ignore below 200Hz as I know what is causing that - it's filters I have set in JRiver...

Is what I am seeing some kind of clocking issue? I verified everything was set at 48kHz.

[mwillems]

I did some more testing. It seems like it is a MobilePre USB / Lynx interaction issue.

I ruled out the DAC being defective by testing with the Metric Halo and DEQX...same issue using REW-JRiver-Lynx-DAC-MobilePre USB-REW

I ruled out the Lynx being defective by doing a test purely through REW (lynx digital out-digital in).

I also ruled out JRiver by doing a test from REW-JRiver-Lynx digital out-Lynx digital in-REW.

See attached..note ignore below 200Hz as I know what is causing that - it's filters I have set in JRiver...

Is what I am seeing some kind of clocking issue? I verified everything was set at 48kHz.

Have you tried measuring in a different measurement suite, like Holm Impulse?  Holm displays phase and impulse response info as well, and looking at the phase in that area may be a clue as to what's going wrong and where. For example, I was getting strange measurement results with loopback for a while until I realized that my loopback soundcard was inverting polarity.  It produced some high frequency anomalies, but I couldn't figure it out until I looked at phase in Holm, and saw it was all goofed up (that's what tipped me off).  

To be fair, my anomalies were not as severe as what you're seeing.  That null is pretty intense; the only place I've ever seen anything like that in measurement is a crossover with one element 180 degrees out of phase (but that's obviously not your situation unless you've got a supertweeter crossed at 10 KHz that's getting inverted somehow  ).

[Nyal Mellor]

I looked at the phase response and you are right it is not flat, there is around 520 of phase rotation at the exact null frequency. Honestly I have no idea. I am pretty sure my test approach has conclusively ruled out any issue with any of the components in isolation, it seems to be the interaction of the MobilePre USB when used to record and the Lynx AES16e when used to playback.

[mwillems]

I looked at the phase response and you are right it is not flat, there is around 520 of phase rotation at the exact null frequency. Honestly I have no idea. I am pretty sure my test approach has conclusively ruled out any issue with any of the components in isolation, it seems to be the interaction of the MobilePre USB when used to record and the Lynx AES16e when used to playback.

What does the phase look like in that region measuring with just REW with JRiver out of the equation (flat? ripples?).  If it's flat than I'm stumped, except to repeat what someone else mentioned above:  have you looked at your JRiver output format tab to ensure that it's not doing any resampling?  I've had consistently weird results trying to use convolution, loopback, and resampling together.

Also (and I know this sounds like a longshot), but have you tried doing all this at a completely different sample rate (like 44.1)?  I'm not sure why that would make any difference, but might provide diagnostic info (like if the null moves to a different position or something).  I just have to imagine that one of the pieces in your chain is resampling something with weird results.

[harlington]

Guys, I need some simple guidance to measure jriver parametricEQ via REW.  What I have is REW software, Umik-1 usb mic that I have been using to measure my FR.  My HTPC is connected to my Denon receiver via HDMI.  I have been playing with jriver pEQ to flatten a 4-5 db dip at 45Hz and would like to see a REW graph for it. Can someone post a simple step by step guidance please?  Thanks.