Loren, it is worth chasing this problem down.
My guess (and it is only a guess) is that the software is not seeing the the "proper" drivers and library. Go through the list it provides (I think it is a "properties dialogue" within Holmimpulse) and make sure it is "seeing" the correct hardware, channel, and drivers. If there is a choice of the libraries (these are some sort of protocols), try using ASIO, if that does not work, then try MME. Stay away from the DS ones. Depending on what is selected, the software (Holmimpulse) may not have control over the volume settings, so check these in Windows to see if the volume is at max and there is no mute switched on.
If worse comes to worse, you can get an outboard device (sound blaster or M-Audio) that is not too expensive and will connect via an USB port. If the notebook is old make sure the outboard device supports USB 1.1 as well as 2.0.
When you are testing to see if it works and recognizes things, do not use the microphone. Instead use the loopback feature (DAC directly into ADC via a cable). Set it up to use only a single channel for input. By avoiding the mic at this testing stage you have eliminated some possible problems and can directly test whether the software is interfacing with the hardware and drivers.
Good Luck
I guess I'm unclear if Holm would not install or its just won't run. Those are different problems. Sometime laptops have sound cards that don't fully meet the requirements. I ended up using a USB sound card for my laptop.
Each time I go to install, it asks for MS .NET to be installed. When I first installed .NET it claimed a driver was missing. Finally got that driver installed (no thanks to the lack of directions, thank you, Microsoft), restarted the system and tried installing Holm. Asked for .NET to be installed (again), did that, but when I go back to reinstall Holm it tells me to install .NET again. Or am I installing .NEYT? That's a little joke... I do it all the time.
Back to reality... The process repeats ad infinitum.
So I can't get to first base. I also am using M-Audio Mobile Pre for the USB sound card.
I can run Holm on my desktop, but can't load on my laptop. Both running XP Pro.
I can just use a longer mic cord for my desk top, but the laptop makes it portable.
Last edited:
Well, maybe we are getting closer. I had assumed that the program had installed okay (and was giving advice on mistakes that I had personally made myself).
I am also using XP on a notebook. I did not bother testing whether it would work using the built in sound card since they are typically of limited utility. I use a sound blaster and once I set up the channels correctly and used the ASIO stuff, I was okay.
Back to the installation issue.
IIRC, I had to be meticulous about downloading & installing all the NET updates (there were many of them). But Holm was able to install after that. So unfortunately I am not getting the same error you are getting and I can't be of help.
One thought, before (re)installing Holmimpulse, did you clear out any directories it may have created? Was there enough of a program installed to where you would need to do an "uninstall"? Or did the install problem happen before anything was created, written, transferred?
I am also using XP on a notebook. I did not bother testing whether it would work using the built in sound card since they are typically of limited utility. I use a sound blaster and once I set up the channels correctly and used the ASIO stuff, I was okay.
Back to the installation issue.
IIRC, I had to be meticulous about downloading & installing all the NET updates (there were many of them). But Holm was able to install after that. So unfortunately I am not getting the same error you are getting and I can't be of help.
One thought, before (re)installing Holmimpulse, did you clear out any directories it may have created? Was there enough of a program installed to where you would need to do an "uninstall"? Or did the install problem happen before anything was created, written, transferred?
Last edited:
I've had nothing but problems with the .net framework. I uninstalled it and don't bother with programs that need it anymore. I started to install HOLM -- just out of curiosity, I already have an ARTA license -- but I bailed when I saw it needed .net.
Guys, if you are going to stay on Windows, you will need .Net. .Net is the future to all software on Windows. Win 7 is basically .Net. After I switched to .Net as a programming environment I could see how Ask was able to do such great software so quickly. Its simply the best code development environment available.
You need to bite the bullit and get .Net. Everything that I do from now on will be .Net. I won't go back to the older Win32 and MFC stuff as its now completely obsolete.
You need to bite the bullit and get .Net. Everything that I do from now on will be .Net. I won't go back to the older Win32 and MFC stuff as its now completely obsolete.
Earl,
This is true. This is also why we have an educational system. Most of the time, industry wants only standards for minimum acceptable quality, but never want the comparitive quality to become a standard method. From an engineering point of view, if we can find a means to have a reliable means of comparitive quality, then we can more tune the results to match a specific desired listening experience. I have an old classmate that has developed a water treatement system and claims the user can adjust it to taste like whatever brand of bottled you may desire. I think it's possible with audio as well. So what I'm saying, let's not stump out different ideas, but rather, find the appropriate.
For example, you find that constant directivity is very important, as do many others in the audio field, but most of this research do not go much lower than 500Hz. Then there is a question, if we want to accomplish CD to lower frequency, what kind of tradeoffs do we have to give up? What do we gain? I'm sure that most people cannot tell what we gain. Is it worth pursuing? From a commercial concern, probably not unless there might be some way to reasonably implement it into an existing home. But we never know until someone has actually done it.
Now, if we look at measurements, each type of measurement was created because there was a problem to be solved. Recently, I was looking at noise spectrum in a little plate amplifier. I had two, and the only difference is that the plate they were installed on were of different material, one aluminum, and one iron, both of the same material thickness. This difference in configuration was deliberate because cost and performance were both taken into consideration. Just from this knowledge, can anyone say which would give better noise figures, and why?🙂I already know that the data was consistent with our judgement, but I'd like to hear what others would think, and see if my little detour was worth it or not.😉
I can speak to this, I think. The thought seems a little disjointed, so maybe I am interpreting this wrong, but...
I think many people and even more engineers loose sight of he big picture. People get hung up on quality, but the bottom line is any company only wants to expend the minimum amount of energy and resources to captured the intended market.
The end game is profit and any marketing department knows you can't be all things to all people (and make a profit).
I don't think most engineers see this and are happy to try to build the best widget that humans can conjure up with no regard to cost or return on investment.
Marketing rolls its eyes and restricts engineering's requirements enough so they can get the job done within budget and on time.
Engineers don't seem to get it, point fingers at marketing claiming they are a bunch of meatheads and if engineers ruled the world it would be a better place. Meanwhile, marketing is having the same thoughts about engineers and the old adage that you need to shoot the engineer to get the project done is only foiled by the no-fiream policy from the company management (which have their own view of things).
All this boils down to what I said in that there is no profit in being all things to all people. You can't make one loudspeaker box with a large knob on it that allows the customer to dial in whatever sound they want because most would not be able to afford it. Even if the knob was only controlled by manufacturing, it simply is not cost effective.
Yeah, I'm a Luddite still running XP / 32. I'm planning on buying a notebook with Win 7 / 64 so maybe it will work better.
This is why the Apple started in the garage, isn't it? Heh, this makes me think of me telling my kids that I've been paying for all the electronics stuff around the house, and my daughter tell me "Yeah, but we get what you phased out...". Whenever my daughter wants to get something higher priced, she know shes not going to get her mom to pay for it. when she comes to me, I point is "dear, can you explain how you decide what you want to buy in a logical way?"😀
Personally, I think it's a communication problem, and it is hard to manage the right balance between creativity and profit. The book "From Barbarians to Bureaucrats" had inspire me to think about managment issues like these. Sometimes you have to make near term money to support visionary development efforts. Sometimes you just need to get the sales, engineering, etc. working as a joint group instead of each trying to fight for control over the product and not willing to be responsible for the consequences.
Putting on my consumer hat, I'd like to think whenever I buy a product, I'd like to trust the company is providing the absolute best quality they can provide at the price level I feel I'm getting good value. There is so much variety out there that it really hurts my mind to try select from a whole bunch of brands. Up to now, Toyota, Erricson, Asus, have not let me down, but I do notice that the youger generation know less and less about their products because too much has been standardised. I once just outright told a support engineer, "hey, I had 3 people answer the same question, and no two gave me the same answer, the manual does not explain it in detail, and if you cannot give me a better answer than what is in the manual, why do we need your support?"
I must be getting old because these interesting things are starting to flash back. I used to have an engineer in my group, it was really hard to get him working on projects deliverable to customers. But when it came to technology demonstrations to impress, boy, he worked night and day to get things working, and really does a good job giving the presentation as well. But if we really wanted it to be deliverable to customer, we always had to get someone else to work the bugs out of it.
Back to speakers, as in any design, we always want a vision, but also a plan to do things incrementally, and make necessary adjustments along the way.
Last edited:
Not from my experience. Most engineers are smart guys and are well aware of what it takes to bring a product to market within it's budget. I know the guys I work with are.
Rob🙂
Yes I would say it is often more a problem with senior management or finance. I've seen the following:
-------------------------------------------------------------------------
Sales promises some level of specification, the designers/engineers design something to meet that specification, the estimators estimate how much it will cost to build and then management/finance says to the estimators that is too expensive redo the estimate (but keep the specifications the same).
The whole thing goes ahead, and of course it is impossible to build the designed product for the "adjusted" estimate, so either something is compromised and the product doesn't meet the spec any more (or is severly compromised in quality) or if there is a contractual obligation the product is developed and sold at a loss.
-------------------------------------------------------------------------
anyway this has very little to do with what measurements are important, at least from a DIY sense, in that every DIY'er can set their own budjets and expectations, and adjust them as things develop.
I think one thing has become clear however, and that is that there still is NO concensus (among experts) about what are the most important measurements.
As I alluded to earlier, regardless of what measurements we deem as important or not important, just having the measurements isn't enough. One needs experience in interpreting them (unless you have say for example a freq response curve that is ruler flat say plus or minus 0.1db from 30 hz to 20Khz) all the way out to 30degrees (or greater) which is almost certainly not going to happen, then comparing the measurements of one speaker vs another and coming to a conclusion of which one the majority of people will prefer, without actually hearing them, I think is going to be a difficult task.
I think that in reality (whether people believe so or not) they MUST have some level of subjective interpretation of the measurements, otherwise how could one determine that a 3db peak at freq X is more of a problem than say a 3db dip at Freq Y (maybe they are of equal importance I don't know)? I think only with experience, either making measurements and trying to make ones own correlations to those measurements, which is obviously going to be subjective, OR looking at measurements and trying to spot what it is that is common in them that caused a panel of expert listeners to judge them as being of superior quality. I think that this second one is would probably be more objective, but probably harder, and most certainly out of the reach of the average DIYer.
In order for this second "training" method to be as objective as possible then the trainee should only be shown the results of the measurements and the relative subjective ranking of the speakers, and no information about what it was that made the people rate the speakers as they did. Sound like too hard a task??
So in the end, for us who are not working for audio companies and instead, designing our own stuff, I think that rightly or wrongly, we have to make measurements take an educated guess on what is problematic from those measurements, try and do something to rectify it, make more measurements, but also listen to the result, and add what it is we discovered to our personal (subjective) knowledge base.
I know that one of the things I did whilst working on my crossover resulted in better freq response measurements, but definitely resulted in inferior (subjectively) sound. Another measurement that I hadn't done (I've only done on and off axis freq response and impedance measurements) might have told me why that was, but for me fix was worse than the problem (a freq response hump between 1K and 2K which I flattened out with a low q notch). I removed the notch and I'm now much happier with the sound (despite the fact the measurements, by my interpretation, tell me it should be otherwise).
Tony.
Last edited:
not right at the moment 🙂 they are on my computer at home. I'm also unsure whether I have X Y and Z measurements or only X and Y measurements, I'd have to check (Z being the current)... I have definitely have X and Y (no notches and two notches 1 at 1-2k and a higher q one at 4k) but I don't think I took measurements with only the 4K notch, instead deciding to go with the simulated result as the sim without the two notches and with them matched the actual measured result quite well.
Unfortunately doing more measurements isn't going to be practical for a while as my 2 YO daughter got back yesterday, and my other half is still away.
Tony.
Unfortunately doing more measurements isn't going to be practical for a while as my 2 YO daughter got back yesterday, and my other half is still away.
Tony.
Here's what I got when I flattened for one axis leaving in place some measures for othe off axis response:
Strangely, this reminded me of a mini monitor sound. Man I did too much experimentation.
Dan
An externally hosted image should be here but it was not working when we last tested it.
Strangely, this reminded me of a mini monitor sound. Man I did too much experimentation.
Dan
OK I pulled up the on axis measurement of the speaker black trace, and then put in the simulated crossover (that was in use at the time of that measurement) red trace, and finally the simulated crossover with only the 4k notch filter on the woofer and no 1k notch (blue trace).
revisiting the measurements and comparing actually made me think about it some more (this is where I was going with the needing experience in order to interpret what is better). The 1-2K notch is having an effect on much more than just the 1-2k freq it is supposed to, dropping the sub 1K freq quite a bit.
If we take the 0db line as the mid point then the freqs below between 500 Hz and 1K without the notch are closer to what they are in the past the 1k - 2K hump. Looking at the freq response alone I'd say that the effect that the notch has below 1K is having a negative effect on the perceived subjective response (in my case) which is worse (to me) than the hump between 1 and 2K.
Note that it wasn't a tonal imbalance that I was perceiving as being wrong though, if anything the tonal balance seemed better, it was a case of the speaker sounding noticeably distorted. The funny thing was the distortion measurements of the raw driver compared to with the notch filters were virtually identical. I also found that listening at normal levels after a short while my ears began to hurt (and it wasn't due to excessive spl).
edit: I just noticed another thing, the simulated 4K notch seems more effective without the 1K notch in place, so it may in fact be the drop at 4K which is making the difference...
The second graph shows three different simulated crossovers. The first blue is the one that is close to what is currently running. Which consists of nothing more than a notch filter at 4k on the woofer, and a 4.3uF capacitor and 1.0 ohm series resistor on the tweeter. This gives a Bessel 2nd order acoustic slope at 3Khz cross point. (note that I found the treble too prominent and put in a 2.2 ohm resistor in which sounds much better to me (again don't have a measurement of that). The red graph is a sim of a 3rd order butterworth acoustic rollof at 2Khz, and the green is (I think) a 4th order LR acoustic rolloff at 1.8Khz (with significant padding on the tweeter). Note I think that there are two problems with this one. The first being throwing away 2db of sensitivity to get a marginally flatter response, and the second is that it may be too flat and once room gain comes into effect will sound too bass heavy.
I'm leaning towards the third order butterworth at this point, the idea was to cross at 2k to take advantage of the 3db hump to fill in some of the trench in the woofers response there. But I can't tell from just looking at the graphs whether it will sound better (probably with some active BSC in my active crossover from the stereo subs) than the currently simpler setup. I'm a bit hesitant to shell out for the substantially bigger inductors I'd need for this (the biggest one being on the tweeter ironically) in case it sounds worse! I guess what worries me about the 3rd order butterworth sim is the sharp jump up at around 1Khz, but the thing that appeals to me is the overall (I think) smoother response and the higher overall sensitivity than either the current implementation or the 4th order LR.
Unfortunately I think I really need to take off axis measurements of the individual drivers and sim at the various angles to get any possible clue, as the on-axis alone is really not going to tell me (I don't think).
Tony.
revisiting the measurements and comparing actually made me think about it some more (this is where I was going with the needing experience in order to interpret what is better). The 1-2K notch is having an effect on much more than just the 1-2k freq it is supposed to, dropping the sub 1K freq quite a bit.
If we take the 0db line as the mid point then the freqs below between 500 Hz and 1K without the notch are closer to what they are in the past the 1k - 2K hump. Looking at the freq response alone I'd say that the effect that the notch has below 1K is having a negative effect on the perceived subjective response (in my case) which is worse (to me) than the hump between 1 and 2K.
Note that it wasn't a tonal imbalance that I was perceiving as being wrong though, if anything the tonal balance seemed better, it was a case of the speaker sounding noticeably distorted. The funny thing was the distortion measurements of the raw driver compared to with the notch filters were virtually identical. I also found that listening at normal levels after a short while my ears began to hurt (and it wasn't due to excessive spl).
edit: I just noticed another thing, the simulated 4K notch seems more effective without the 1K notch in place, so it may in fact be the drop at 4K which is making the difference...
The second graph shows three different simulated crossovers. The first blue is the one that is close to what is currently running. Which consists of nothing more than a notch filter at 4k on the woofer, and a 4.3uF capacitor and 1.0 ohm series resistor on the tweeter. This gives a Bessel 2nd order acoustic slope at 3Khz cross point. (note that I found the treble too prominent and put in a 2.2 ohm resistor in which sounds much better to me (again don't have a measurement of that). The red graph is a sim of a 3rd order butterworth acoustic rollof at 2Khz, and the green is (I think) a 4th order LR acoustic rolloff at 1.8Khz (with significant padding on the tweeter). Note I think that there are two problems with this one. The first being throwing away 2db of sensitivity to get a marginally flatter response, and the second is that it may be too flat and once room gain comes into effect will sound too bass heavy.
I'm leaning towards the third order butterworth at this point, the idea was to cross at 2k to take advantage of the 3db hump to fill in some of the trench in the woofers response there. But I can't tell from just looking at the graphs whether it will sound better (probably with some active BSC in my active crossover from the stereo subs) than the currently simpler setup. I'm a bit hesitant to shell out for the substantially bigger inductors I'd need for this (the biggest one being on the tweeter ironically) in case it sounds worse! I guess what worries me about the 3rd order butterworth sim is the sharp jump up at around 1Khz, but the thing that appeals to me is the overall (I think) smoother response and the higher overall sensitivity than either the current implementation or the 4th order LR.
Unfortunately I think I really need to take off axis measurements of the individual drivers and sim at the various angles to get any possible clue, as the on-axis alone is really not going to tell me (I don't think).
Tony.
Attachments
Last edited:
What does the impedance look like for all these? Have you tried measuring with the driving amplifier as part of the DUT?
Last edited:
ok here is the impedance curve of the current crossover. 4K notch and 4.3uf cap on tweeter with 2.2 ohm resistor in in series. I don't have measurements for the other config (that I can find). I'm pretty sure I did one with the 1K and 4K notch but I think speaker workshop crashed on me and I lost it.
I'll see if I can get SW to sim the impedance for the simmed crossovers. Note the measurement shown here was taken with the speaker sitting next to the TV and with other stuff nearby not on the proper measurement platform more than a meter away from anything else. Also it has been 1/16th octave smoothed to get rid of some furryness 🙂
I'm not sure what you mean by measuring with the driving amp as part of the DUT. I am using a walin II jig, and the signal that comes from the amp gets fed back (via a divider network) into the other channel of the sound card. Mic on one channel, signal as passed through the amp into the other channel. F/R Measurements were taken using holm impulse and imported into speaker workshop. Impedance measurements (of the raw drivers in cabinet) were taken with the speaker in position on the measuring platform with speaker workshop driven directly off the soundcard, no amp. Picture of the measurement environment also attached (note that the measurements shown had slightly less furniture in the room the closest object was the ceiling at around 1.2M next closest were items to the side (sofa to the right and TV to the left) at around 1.3M each and floor at around 1.3M.
I've also attached a graph with on axis (black) and then 15, 30 and 45 degrees off axis measurements. This is with the 1K and 4K notch filters on the woofer.
edit: the simulated impedance will have to wait because it is past my bed time, and I need my beauty sleep 😉
Tony.
I'll see if I can get SW to sim the impedance for the simmed crossovers. Note the measurement shown here was taken with the speaker sitting next to the TV and with other stuff nearby not on the proper measurement platform more than a meter away from anything else. Also it has been 1/16th octave smoothed to get rid of some furryness 🙂
I'm not sure what you mean by measuring with the driving amp as part of the DUT. I am using a walin II jig, and the signal that comes from the amp gets fed back (via a divider network) into the other channel of the sound card. Mic on one channel, signal as passed through the amp into the other channel. F/R Measurements were taken using holm impulse and imported into speaker workshop. Impedance measurements (of the raw drivers in cabinet) were taken with the speaker in position on the measuring platform with speaker workshop driven directly off the soundcard, no amp. Picture of the measurement environment also attached (note that the measurements shown had slightly less furniture in the room the closest object was the ceiling at around 1.2M next closest were items to the side (sofa to the right and TV to the left) at around 1.3M each and floor at around 1.3M.
I've also attached a graph with on axis (black) and then 15, 30 and 45 degrees off axis measurements. This is with the 1K and 4K notch filters on the woofer.
edit: the simulated impedance will have to wait because it is past my bed time, and I need my beauty sleep 😉
Tony.
Attachments
Last edited:
First of all, I would seriously try to get the minimum impedance above 6Ohm unless the amp really can double the output power when you half the load impedance. Second, I would try to get the speaker impedance as constant as possible. Normally I try to get as close to 2Ohm diff between min and max with the least components possible.
During SPL measurement, if you have the ref input connected to the input of the amp instead of output of the amp, then the amp is part of the DUT.
During SPL measurement, if you have the ref input connected to the input of the amp instead of output of the amp, then the amp is part of the DUT.
HI Tony Several disagreements
1) most experts that I know are fairly well in agreement on what needs to be measured and how to interprete the results. Most DIY are not.
2) Subjective plays a very important role when you goal is "preference" as you clearly state that yours is. However, "accuracy" does not require this same dependence on the subjective. You have to decide if what you want is "preference" or "accuracy" - they will almost never be the same in the short term - longer term they probably will - because "preference" is not a stable assesment. It can and does change with the environment, time of day, person, mode and on and on. Its a moving target.
3) most of the "peaks and dips" things that you discuss have been quantified. Knowing the liuterature goes a long ways to helping with interpretation.
Regarding engineering - its no different than anything else, all aspects of a company have their point of view. After doing engineering for 40 years I have reached one conclusion - only when all of these aspects agree on one thing, and that is "customer value", does it all work. If the marketing approach offers "value" then it will work, if engineering is "value oriented" then it will work as well, and only if managment is committed to giving the customer "good value" will the company survive. Example - Honda - not the best cars in the world in many senses, but clearly the best value, and probably the strongest of all the car companies at the moment - and always in the top three.
- Status
- Not open for further replies.