|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
73.129.54.230
In Reply to: RE: First fruits of HRT Line Streamer posted by E-Stat on September 20, 2016 at 12:41:13
Looks like you are clipping off the waveform. The amount of distortion that something like this can generate has to be experienced to believe it.
Check out the maximum sample amplitude and location from the original and the recorded sample. I cannot help you in how to do this in Audicity because I do not use this software to record. I use Sony Sound Froge
Before you start recording you should fix whatever is causing this issue or you will end up with a lot of lousy recordings (unless you are a fan of the loudness wars).
Ed
We don't shush around here!
Life is analog...digital is just samples thereof
Follow Ups:
Yeah, I need to better acquaint myself with the software. There is an active meter that didn't seem to clip during the recordings, but it got really close at times.
Perhaps I should try adding some headroom. :)
Ed gave you good advice on the amount of headroom to leave.
It is important to leave at least 3dB margin from FS not least because of the nasty effects of overloading the ADC, but because the intersample amplitude (in the recovered analogue waveform), especially with upsampling reconstruction filters, may actually end up with a peak greater than the sample peaks. This is of greater importance if you intend to downconvert to RedBook.
I personally don't worry about trying to set the initial record level to have peaks at exactly -4dB(in my case) - that is done in the processing phase where I will then apply gain (at the very last stage) to bring the peak up to -4dB.
This is fine since recording digitally doesn't have the same restrictions as recording on analogue tape in terms of the limited dynamic range.
Take care on how the gain is applied - different packages refer to "Normalize" in different senses. In some package "Normalize" or "Maximize" will apply variable gain to a desired RMS value which will compress the signal (Goldwave for example). RX uses Normalize in the sense of applying fixed gain to achieve a peak of whatever you choose. There is a separate function to literally apply a gain of (say -4dB).
Regards Anthony
"Beauty is Truth, Truth Beauty.." Keats
I have at least a half dozen of these here. Are they yours?
I also have RX5 and would be happy to test files for you.
I should have one in the listening room. :)
Thanks for the offer, but I'm going to experiment with recording levels and see what difference that makes. Each file is about 150 MB anyway.
It looks like you should turn down the input by 4-5db. I don't know your setup so take that comment with a grain of salt if you already have a plan.
Its best to have it turned down so your high samples are somewhere around -3db. This usually gives you headroom for transient peaks but that could be insufficient with very dynamic LPs.
Ed
We don't shush around here!
Life is analog...digital is just samples thereof
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: