Hi Ask
Had a chance today to use the software a lot. Comments and bug list:
1) it is very hard to figure out the gain settings to get the maximum SNR and to drive the speaker at a given level. The system kept telling me that there was digital clipping but yet, another piece of software with the same levels settings in Windows, reported a very low level. Its baffling to me how I get exactly 2.83 volts into the system and then get all the gains set so that everbody (your software and others) are happy. In my other software SpectraLab, I can always see the waveform directly from the sound card. I can see cliping etc. and there is a meter at the bottom for setting the levels. In your software its a pure guess.
2) it gets the cleanest impulse responses that I have ever obtained - thats great
3) on my system there is an impulse at about -280 samples. I'm guessing that this is crosstalk or something bleading through the electronics. Normally this would not be a problem - just set it outside of the window - but you have to do this in "impulse" every time. The system always resets back to "auto window" setting that starts well ahead of 280 samples. You have to go in and turn off "Locked" each and every time you take a measurement. It doesn't stay "unlocked" and the window doesn't stay were you set it.
4) when saving the impulse responses you have to set the number of samples to export each time. It should remeber these settings.
Works pretty good otherwise. I like to do more myself than your "auto everything" approach however. I would prefer that the system let me set the window and not have the software find what it "thinks" that it should be.
Had a chance today to use the software a lot. Comments and bug list:
1) it is very hard to figure out the gain settings to get the maximum SNR and to drive the speaker at a given level. The system kept telling me that there was digital clipping but yet, another piece of software with the same levels settings in Windows, reported a very low level. Its baffling to me how I get exactly 2.83 volts into the system and then get all the gains set so that everbody (your software and others) are happy. In my other software SpectraLab, I can always see the waveform directly from the sound card. I can see cliping etc. and there is a meter at the bottom for setting the levels. In your software its a pure guess.
2) it gets the cleanest impulse responses that I have ever obtained - thats great
3) on my system there is an impulse at about -280 samples. I'm guessing that this is crosstalk or something bleading through the electronics. Normally this would not be a problem - just set it outside of the window - but you have to do this in "impulse" every time. The system always resets back to "auto window" setting that starts well ahead of 280 samples. You have to go in and turn off "Locked" each and every time you take a measurement. It doesn't stay "unlocked" and the window doesn't stay were you set it.
4) when saving the impulse responses you have to set the number of samples to export each time. It should remeber these settings.
Works pretty good otherwise. I like to do more myself than your "auto everything" approach however. I would prefer that the system let me set the window and not have the software find what it "thinks" that it should be.
I had a little trouble setting the levels, too. But as my sound card mic input has actual knobs on it, it didn't take long. I just ran a few sweeps and looked at the resulting levels.
The line input doesn't have a knob, just a -10dB or +4dB switch. In the +4dB position, I see a -10 level input as -10, so that seems to work well.
As for setting the window, you might try unchecking the "Auto apply time window" button on the lower right of the Setting tab. I never have, but maybe it will do what you want.
Great software - features getting better and better.
Next up - the distortion graphs!
The line input doesn't have a knob, just a -10dB or +4dB switch. In the +4dB position, I see a -10 level input as -10, so that seems to work well.
As for setting the window, you might try unchecking the "Auto apply time window" button on the lower right of the Setting tab. I never have, but maybe it will do what you want.
Great software - features getting better and better.
Next up - the distortion graphs!
panomaniac said:
Exactly, and every time you make a measurement you have to go back and uncheck it. It doesn't stay unchecked.
I think it is an excellent way that the software development is going together with users' feedback.
In a while, it will be a great piece of software.
Thanks again!
In a while, it will be a great piece of software.
Thanks again!
gedlee said:
I am using my SoundEasy hardware setup with HOLMimpulse. It includes attenuators to reduce the input signal to levels manageable by the sound card. In my sound card's mixer panel I am able to save particular mixer setting profiles. I believe that this is the common way to do it.
I think that having a level meter with clipping indicator in the software will be very helpful.
Levels and more
When I measure logsweep I use a full scaled PCM-signal - if you want a lower output you will need to adjust this using your windows mixer.
Thanks again for the feedback
You don't have to be guessing... See attachmentgedlee said:
When I measure logsweep I use a full scaled PCM-signal - if you want a lower output you will need to adjust this using your windows mixer.
Point taken - will not auto window anything if you choose not to
I'll keep it this way, I set the impulses to the maximum possible length, that is available
I'll try to make my "auto everything" optional for you experts 😀
Thanks again for the feedback
Attachments
Re: Levels and more
I simply don't understand why this can't be an option? If the user doesn't change these settings then they always get the maximum, if they do then they always get what they set it to. How does this compromise anything?
I am aware of the parts of the program that you show, but they don't help really. I think that the software may not be working with the windows mixer in the way that you think or not in my system. I do not seem to be able to turn down the output signal with the mixer. All I can say is that setting levels is really confusiong. It may be an odd part of my system - Vista 64 or sound card, but the input levels cannot be changed in the mixer because the software overrides it every time. You have to exit the software, change the input level and then strta the software again because it will only use the level set when the software starts. It resets it back to this level at every run. This may not be your program as I run two measurement systems at the same time.
askbojesen said:
I simply don't understand why this can't be an option? If the user doesn't change these settings then they always get the maximum, if they do then they always get what they set it to. How does this compromise anything?
I am aware of the parts of the program that you show, but they don't help really. I think that the software may not be working with the windows mixer in the way that you think or not in my system. I do not seem to be able to turn down the output signal with the mixer. All I can say is that setting levels is really confusiong. It may be an odd part of my system - Vista 64 or sound card, but the input levels cannot be changed in the mixer because the software overrides it every time. You have to exit the software, change the input level and then strta the software again because it will only use the level set when the software starts. It resets it back to this level at every run. This may not be your program as I run two measurement systems at the same time.
Mixer Levels
Output level
HOLMImpulse does not alter the mixer. In Windows Vista you can adjust output levels for individual applications and master. In Windows XP you can adjust master and PCM.
If your other applications adjust the mixer settings (Like audiotester does) while you are measuring, then I also get confused - why I have not included windows mixers.
Input level
The PCM wave I record is also given by the windows mixer level
I have tested this in windows XP and Vista - In Vista you might want to check the keep-stream alive box + restart - this way you will have the HOLMImpulse application in your Vista mixer before measuring.
No mixing in HOLMImpulse
Because of OS dependencies and other trouble I'll not implement mixer control in HOLMImpulse - I will leave all the control to you and your OS. Like audacity (http://audacity.sourceforge.net/) does not alter the mixer settings.
gedlee said:
Output level
HOLMImpulse does not alter the mixer. In Windows Vista you can adjust output levels for individual applications and master. In Windows XP you can adjust master and PCM.
If your other applications adjust the mixer settings (Like audiotester does) while you are measuring, then I also get confused - why I have not included windows mixers.
Input level
The PCM wave I record is also given by the windows mixer level
I have tested this in windows XP and Vista - In Vista you might want to check the keep-stream alive box + restart - this way you will have the HOLMImpulse application in your Vista mixer before measuring.
No mixing in HOLMImpulse
Because of OS dependencies and other trouble I'll not implement mixer control in HOLMImpulse - I will leave all the control to you and your OS. Like audacity (http://audacity.sourceforge.net/) does not alter the mixer settings.
Attachments
Re: Mixer Levels
You don't need to get involved with the Windows mixer - just don't generate the sweep at 0dB FS, why not allow the sweep to be generated at lower levels within HolmImpulse?askbojesen said:
Output & Input mixerlevels
HOLMImpulse Output software level 0-1 amplitude makes sense - My software mixer cannot be any worse than the Windows software mixer... I'll implement easily.
HOLMImpulse Input software level does not make sense
HOLMImpulse Output software level 0-1 amplitude makes sense - My software mixer cannot be any worse than the Windows software mixer... I'll implement easily.
HOLMImpulse Input software level does not make sense
I agree - keep away from the windoze mixer. But having a lower level signal might be a good idea, as John says. Some cards don't do well with 0dBFS.* Being able to set a precise level in HOLM would be nice. E.G. -3dB
It's all working fine for me - but every setup is different.
* Have a look at this clipping test.
http://www.hydrogenaudio.org/forums/index.php?showtopic=9772
It's in APE format. If you can't use it, let me know - I have it in FLAC and wave.
It's all working fine for me - but every setup is different.
* Have a look at this clipping test.
http://www.hydrogenaudio.org/forums/index.php?showtopic=9772
It's in APE format. If you can't use it, let me know - I have it in FLAC and wave.
Ask
It could well be that my confusion is with the new way that Vista works with the mixer. I have never liked the Windows mixer, but it seems like on my system I can't avoid it. I have to use it and its hard to configure with different software when one of them doesn't have any settings. Then everything has to be set to HOLMimpulse. I thing the variable output level is a necessity - thanks.
Again - why can't the output samples for the impulse response be set as an option?
It could well be that my confusion is with the new way that Vista works with the mixer. I have never liked the Windows mixer, but it seems like on my system I can't avoid it. I have to use it and its hard to configure with different software when one of them doesn't have any settings. Then everything has to be set to HOLMimpulse. I thing the variable output level is a necessity - thanks.
Again - why can't the output samples for the impulse response be set as an option?
What defines the resolution of the impulse response?
Would it be possible to have the IR be a very high number of points/sec combined with a long sweep to get greater detail? Mainly when doing CSD.
Would it be possible to have the IR be a very high number of points/sec combined with a long sweep to get greater detail? Mainly when doing CSD.
Using wav files
Many thanks for the excellent free program! 🙂
This may sound like a stupid question, but is there anyway that it could except plain old wave files and analyse them? I run a linux based crossover/playback system and it would be nice to test the whole system running, which would require the playback and recording to be separate.... Ie playing a generated sweep wav on the system and recording with a windows machine (either directly into the program, or into a wav file).... I know this is possible using aurora plugins with cooledit/audition, but its never worked that well for me.
Many thanks for the excellent free program! 🙂
This may sound like a stupid question, but is there anyway that it could except plain old wave files and analyse them? I run a linux based crossover/playback system and it would be nice to test the whole system running, which would require the playback and recording to be separate.... Ie playing a generated sweep wav on the system and recording with a windows machine (either directly into the program, or into a wav file).... I know this is possible using aurora plugins with cooledit/audition, but its never worked that well for me.
gedlee said:
I second the variable output/input testing if it is feasable - some recording engineers hit the meters no more that -6.02 dB as they think the ADCs perform better at this volume.
But I would recommend if your card doesn't have ASIO drivers to use ASIO4ALL instead of microsofts drivers. http://www.asio4all.com/
I for sure would not use XP's directsound and I am very skeptical of Vista's WASAPI platform.