Moderators: Dave Moll, Andrew Davie, Steve Anderson
gary wrote:Very poor frequency response both lower and upper - can you post a block diagram of the signal path showing what each section is doing please?
gary wrote:Ok, so it seems it could only be the head amp causing the low and high freq roll off.
I am not happy with the phone/mp3 source but, and correct me if I am wrong, that is what you were using in the previous tests right? (i.e. direct from phone to laptop) and that seemed fine. I suppose the only test you didn't do then is with the sweep - perhaps you should back up and do that before looking at the head amp.
Harry Dalek wrote:Hi Gary
Just a quick one before i review in your last post .
It looks like i have lost the original sweep its not one my old laptop ,i was just doing another one ,but it might be best if i do get a good one via you before i go on testing.
gary wrote:Yes Harry, mp3s are not a good choice for video - the compression is done at the expense of audio we don't hear - but when we convert it to video we very definitely see it.
Oh i see that was a problem....It is not so much using .wav files as they can in fact be mp3 encoded as well - what you want is PCM (i.e. no compression) - thankfully nearly all .wav files are.
Will keep to wav files from now on,
Thanks for up loading the sweep that will make it easier a factor i know is correct .Harry I strongly recommend that you download and install Audacity - it is totally free and will give you everything you need to analyse the audio files.
I will down load that i was just using NCH tone generator but that one sounds good ...
In addition, when recording these tests always record in mono 48kHz if possible.
Harry could you also post the sweep file as well? so we have a benchmark.The secret to testing and debugging is to minimise your variables - ideally only change one thing at a time.
Steve Anderson wrote:Suggestion - if you're wary of the original mp3 sweep file, as Gary suggests install Audacity. Within it you can generate sine-wave tones across the audio band, though I don't think a sweep. Say at 1kHz, 1.5kHz, 2.2 etc, much as the spacing of resistor values, edit these step-tones together, 30secs of 1kHz as a line-up, then say 10secs of the others. I think the highest it will generate is 20kHz - enough for NBTV.
Make sure that they are displayed at equal levels in Audacity then check the output of the phone/laptop/whatever is the same - you could well be surprised, especially with a phone!
When you measure do not use an AC multi-meter, most give up at 1kHz or less, especially the digital ones, use a scope. They are mainly intended to just measure AC mains frequencies, not full-band audio. (Exceptions - expensive Flukes). Make sure your test set-up isn't feeding you duff info...
Steve A.
gary wrote:Harry both of your files are still 44.1kHz stereo - so you are doing something wrong when recording.
Here are the response graphs - as you can see the samsung appears to be far superior in the low frequencies but has a glitch in there - perhaps your cable or plug came loose?
gary wrote:That glitch at the start looks like a buffering problem in either the output player or the input recorder (see the gap in the audio below). Perhaps it is not worth worrying about as it only happens at the very beginning of the audio.
Yes the audio is still stereo see below for how to set it to mono.
Users browsing this forum: No registered users and 91 guests