Member
Joined 2003
REW is indeed "clunky" but rather featureful. I dislike the UI myself and lack of proper dual channel capability, so my recommendation is for ARTA even though ARTA development is likely to move at a slow pace, the developer has retired a few years back so I can only assume his energy for future software development will be declining.For me there are several things that HOLM does better than REW.
That said, REW has made very important improvements over the years, and gets better all the time. But its funky need of a specific timing reference is still a big handicap IME. That's the main part of HOLM that REW could do well to adopt. I used to use ARTA quite a bit, but haven't in years so am not up to date on it. I liked it a lot, but would normally turn to HOLM for routine work.
- Sound cards. HOLM is better than any measurement software I've ever used at dealing with sound cards. REW seems very clunky in that regard.
- Timing. HOLM needs no second channel or acoustic markers to measure time. It is very accurate and repeatable at measuring time offsets. The time lock feature is extremely useful and easy. This might be related to point #1
- Phase. Because of the difficulties in timing, I've never gotten a useful phase measurement from REW. In HOLM, it's just there and makes sense. This is probably related to point #2
- Measurement import. HOLM can compare any two files and show the FR difference. Unlike REW that needs a very specific format of sweep and even then often refuses to import even it's own sweeps. HOLM does care as long as the files are at the same sample rate. You can compare two sweeps, a sweep and an impulse, an impulse and a sound or even two songs. It's a handy feature.
- Impulses. It is called HOLM Impulse for a reason.
The way HOLM handles impulses feels more solid and useful than other softwares.
1. I don't understand what you mean "dealing with soundcards".. measuring sound card performance or just overall support? Both REW and ARTA support WDM and ASIO drivers which covers pretty much all possibilities.
2. The timing lock function is unreliable in most cases (especially for USB mics that seem to be popular these days), and the sole reason why dual channel measurement systems exist. Maybe it provides repeatable measurements in your specific configuration, I would suggest that every user should do an evaluation of their configuration to guarantee repeatability of that function. I would say it's good enough to set up delays for HT speakers, but for individual drivers where 1 sample of error presents significant phase error, just avoid. 2 channel measurement systems can be dead simple these days, just get a 2 channel USB interface and XLR mic, 1 patch cable loops back from output to input on the interface, job done.
3. This sound like your lack of knowledge to set FFT start and window length properly, not a software deficiency.
4. There are many ways to compare files with various software. If Holm is your preference then have at 'er. I prefer VituixCAD for this purpose.
5. Try IR to FR tool in VituixCAD, it may surprise you. REW is also very featureful in this regard, and ARTA is perfectly fine, for speaker design I still recommend simply measure IR with ARTA, save the PIR file and process through VituixCAD for perfect results every time, especially processing bulk measurements for full 360 degrees of time accurate data.
In this list, I'm not sure you've captured any functionality that isn't available elsewhere. 1,3,4,5 fall under lack of knowledge/experience and personal preference, rather than functions that can't be achieved in other software. #2 is absent from other software for the reliability issue I mentioned, other developers have simply avoided this option as a "bad idea", and since 2 channel configuration for a timing reference is simple and reliable, that is the preferred method. To verify for myself, I ran Holm and followed the locked timing reference process. Using a USB audio interface as the output, and USB microphone as the input, the locked timing reference provided results +/- 2.5 meters in distance between measurements, not even close to repeatable locked timing.
In any case, Holm is still an option if it's your preference, just no longer developed and supported other than from community support like this forum.
Member
Joined 2003
Yep. I explained what I can easily achieve in HOLM that isn’t as good or easy in other software. In other words things that could be improved in REW. That was your question, to which I gave my best reply. You claim user error or ignorance on my part, then throw in other software that wasn’t even part of the question!
There are much better ways to respond when someone gives you a good faith answer. No worries, I won’t do that again.
There are much better ways to respond when someone gives you a good faith answer. No worries, I won’t do that again.
That's OK... A certain software company once introduced the idea that software can't be used if not currently maintained, disingenuously I might add.
Dual channel techniques are not new. I take it you meant the acoustic reference.
The nice thing about HOLM is that it simply doesn't need any external timing reference, it seems to find timing from the audio stream itself. I can not confirm that it can do so with a USB mic, as I have almost always used a USB soundcard that includes both input and output channels on the same clock. I have tried using two different sound cards (in imitation of a USB mic) with HOLM and REW and not found the offset to be reliable.
I keep thinking that one day I need to sit down and sort out the timing issues with REW (tried a few times), but with HOLM just a mouse click away, I have not been too motivated.
I keep thinking that one day I need to sit down and sort out the timing issues with REW (tried a few times), but with HOLM just a mouse click away, I have not been too motivated.
Member
Joined 2003
With comments like "I've never got useful phase from REW", and troubles setting up a simple loopback timing reference, it sounds like you haven't really tried, so forgive me for making such claims. I did not mean to be rude, just wanted to convey that everything that Holm does can be done with other software, free options even. It sounds like you are happy with Holm anyway so I don't blame you for lack of motivation., but if/when you do feel motivated, I would spend time with ARTA personally over REW, it is more purpose built with loudspeaker design in mind, rather than REW which was built with room EQ in mind that happens to have enough features to be suitable for speaker design. REW also is not truly dual channel, just timing reference, but most people seem fine with that limitation it seems. ARTA does have a price tag though to save impulse response files directly, but otherwise is free to use/play/learn. I am happy to help as well.
I tried with output from USB audio interface, input from USB mic, each subsequent measurement had increased delay, in cm then to meters. Not reliable at all. Tried again using just USB interface only with XLR mic, better results, it initially appeared to have good repeatability until I took a couple min break between measurements then it was off by more than 5 meters, another few min then 10 meters of error. I'll pass on this feature, just run a patch cable from output to input for a reliable timing reference instead of using software guesswork. Since you're using a USB soundcard, I see no reason that you cant do the same, but it will require moving on from Holm.
You've just described all "web apps". Too many these days, but I guess people like convenience more than long term control over their own software and files.
I meant dual channel, and I know it's nothing new, just "the right way to do it". I don't know what you think I mean about acoustic reference, but I am not talking about acoustic references.
Oh yes, I tried and tired. Followed all the help and instructions. Never found it it work right. But I've had some other serious problems with REW, like its refusal to open files sent to me by other users. Luckily the latest update has fixed that problem, so maybe some other problems got fixed.
Yes, that similar to what I found, it would be OK for 3 or 4 runs, then go completely off.
I have not seen that happen, and I have specifically test for it. Even after 10 mins it was not 1 sample off, which I found amazing. Perhaps it's soundcard or USB bus specific? But I tend to restart the audio stream for each new batch of measurements anyway, so drift has not been a problem. As the years pass by, and Windows and my hardware get better, I find all of the measurement software to get more and more stable, with fewer glitches drop-outs and crashes. When I was running measurements under WinXP on an old Dell laptop, too much time was spent fighting audio problems. Those problems are now rare.
ARTA is good, and when I get back into serious projects I should buy a copy. Or maybe just learn REW better.
REW has a lot of features that are very good, many things in the UI are much better than HOLM, and of course the analysis windows are far beyond anything HOLM does. My main complaints with REW have been timing and its need for its own specific sweep signal. Those need improvement.Is that because you have some proprietary stimulus/response pairs you want to deconvolve? You can do that by loading them using File -> Import audio data and using trace arithmetic A/B, as long as the files aren't longer than a minute or so. If you just want to see the frequency content of some WAV file then Import audio data or dropping the file on the RTA will do that.
Sometimes, yes. I've gotten very used to using the import comparison function in HOLMIpulse, so I decided to spend some time with importing "responses" into HOLM versus REW. The differences are interesting. Both have good points. Below I've tried to outline what I see as the differences.
HOLM compares two files, it has to have a reference, it can't just import a recording.
REW does not use an external stimulus reference and simply imports the response file. (but further work can be needed)
HOLM will show a log sweep as a flat response if it's compared to a log sweep.
REW shows a log sweep as a falling response, and linear sweep as flat with a low shelf.
HOLM seems to display minimum phase as default.
REW needs to have minimum phase generated to see it. But then lets you see different phase types.
REW and HOLM show different overall phase shifts on the same recording, but relative phase shifts within that signal seem to be the same.
REW shows basically the same thing as HOLM if the stimulus file is imported then A/B math is used on the stimulus/response pair. This is logical as HOLM is just doing A/B during the import.
Bottom line. The same import of a recording results can be achieved in both softwares. HOLM is a little faster for this use as it does the offset, minimum phase and A/B as part of the import. For me it's nice to have everything done at import, as it's quicker than doing further processing . Maybe I need a macro in REW.
Next will be to test measurement timing on both. Timing in HOLM is super easy, I need to learn how to make it as easy in REW.
Esteemed members, as a noob in speaker modelling and adjusting / validating a crossover: Is my impression correct that if I want to measure phase and response curve from an existing speaker in order to compare it with the same of a speaker under development:
That HolmImpulse is a one stop application with a intuitive interface whereas with other software I need either to use multiple pieces of software or get used to clunky interfaces?
Your input will be greatly appreciated, thank you in advance for spending the time to reply.
That HolmImpulse is a one stop application with a intuitive interface whereas with other software I need either to use multiple pieces of software or get used to clunky interfaces?
Your input will be greatly appreciated, thank you in advance for spending the time to reply.
Thank you, that saved me a lot of time and prevented from going on a wild goose chase.
I'd it going in a very short time and does what I want to do.
I suppose calibrating the soundcard(s) is not needed if I only want is to compare one speaker to another.
FWIW I'm using a FocusRite solo for output and an Umik-1 calibrated microphone and repeated measurements of frequency and phase came through the same. Exported a frequency response file but for some reason or another X-Sim did not import it correctly - I'm only seeing the phase. Will later spend some more time on this to check what I'm doing wrong.
Did that, suspect some invisible markers somewhere. Don't worry, it's not a biggy - I've ordered extra speaker connectors (presently only a single set of connectors - no bi-wire, bought a matching pair) and then will have the tweeter and mid/woofer each going outside so I can tinker with the crossover outside the box without having to open the cabinet.
Last edited:
