Please explain to me the difference 10025 Hz, 48 Kbps, mono vs 8000 Hz, 16 Kbps, mono. These are just talks or voice recordings of a low quality. Of course there is a big difference in file size. 42,000 kb to 14,000 kb. I don't have much disk space on the HD & have about 200 of these talks that I am moving to the HD from old casset tapes, which are not of very good quality after 20 yrs or so. I just want to understand the speaker.
Do you think it would be ok to save them as 8 or 16 kbps? What will I be losing?
MP3
-
Blandine Catastrophe
- Posts: 253
- Joined: Sun Jul 24, 2005 3:25 pm
1 10025 Hz & 8000 Hz are the sampling rate = the number of signal measures in each second to "draw" the wave. The maximal frequency that can be heard is the half of this sample rate.
The frequency of a wave is measured in Hz, and corresponds to the number of cycles per second. Here the width shown by the arrows illustrates one cycle.
On an Audio CD, the sampling rate is of 44100 Hz, what allows maximum a 22050 Hz high frequencies limit. More you lower the sample rate, more you restrain the treeble range.
2 48 kbps & 16 kbps is the resulting speed of the "stream" obtained after the compression, called the bitrate. Due to the compression system applied to save space, the quality of the sound will be better with high values and will decrease with low ones, because the compression makes to loose precision.
3 Combinations of sampling rate & streaming debit : If you use a lower sample rate on the same mp3 debit, you will have less compression, then less compression artifacts, but you will have a lower bandwidth in the treeble range.
Old tapes with low quality recording talk: a 22050 kHz sampling rate would be enough, or even 16000 Hz, just see the spectrum analyser on hissing vowels (SSS's) before to choose, take the double of the highest frequencies who are not tape noise.
Talks: I suppose there is no stereo need, that allows to gain space also to get a monaural file. You will perhaps get a better signal if you digitize only the best of the stereo channel to use it as monaural.
Final choice of the bitrate: it's to you to decide if the quality is good enough at your taste, if not convinced, then choose a higher bitrate until you can listen to it comfortably. As long as you are not sure of what sample rate and bitrate you prefer for that, keep the recording on a *.WAV file to redo it, because you cannot remove the mp3 artifacts. Once you have decided, you can encode directly in mp3 after editing and remove the then useless *.WAV files.
The frequency of a wave is measured in Hz, and corresponds to the number of cycles per second. Here the width shown by the arrows illustrates one cycle.
On an Audio CD, the sampling rate is of 44100 Hz, what allows maximum a 22050 Hz high frequencies limit. More you lower the sample rate, more you restrain the treeble range.
2 48 kbps & 16 kbps is the resulting speed of the "stream" obtained after the compression, called the bitrate. Due to the compression system applied to save space, the quality of the sound will be better with high values and will decrease with low ones, because the compression makes to loose precision.
3 Combinations of sampling rate & streaming debit : If you use a lower sample rate on the same mp3 debit, you will have less compression, then less compression artifacts, but you will have a lower bandwidth in the treeble range.
Old tapes with low quality recording talk: a 22050 kHz sampling rate would be enough, or even 16000 Hz, just see the spectrum analyser on hissing vowels (SSS's) before to choose, take the double of the highest frequencies who are not tape noise.
Talks: I suppose there is no stereo need, that allows to gain space also to get a monaural file. You will perhaps get a better signal if you digitize only the best of the stereo channel to use it as monaural.
Final choice of the bitrate: it's to you to decide if the quality is good enough at your taste, if not convinced, then choose a higher bitrate until you can listen to it comfortably. As long as you are not sure of what sample rate and bitrate you prefer for that, keep the recording on a *.WAV file to redo it, because you cannot remove the mp3 artifacts. Once you have decided, you can encode directly in mp3 after editing and remove the then useless *.WAV files.
Last edited by Blandine Catastrophe on Thu Aug 04, 2005 9:10 am, edited 1 time in total.
Gloup? :-°
Thanks for your very detailed response. As far as stereo is concerned, there is no stereo. These talks were made with a hand held tape recorder in the audience, with background noise. Since they average about 20 yrs in age, they have deteriorated quite a bit. Some parts can not be understood. I hope to learn about filtering, which may help make them more understandable.
I don’t know what the treeble range is with such a recording? I don’t really care about that, but do care about just understanding what was said & file size.
I am not sure what to think about this?
> 48 kbps & 16 kbps is the resulting speed of the "stream" obtained after the compression, called the bitrate. Due to the compression system applied to save space, the quality of the sound will be better with high values and will decrease with low ones, because the compression makes to loose precision
I really don’t have much quality? I can hardly understand the words sometimes on some tapes.
I wish that I could understand this, since I think you are telling me how I can improve what I am doing?
>Old tapes with low quality recording talk: a 22050 kHz sampling rate would be enough, or even 16000 Hz, just see the spectrum analyser on hissing vowels (SSS's) before to choose, take the double of the highest frequencies who are not tape noise
How do I do this?
I guess what you are saying here is, that I should make them into wave files & then change them into whatever MP3 files that I decide?
>Final choice of the bitrate: it's to you to decide if the quality is good enough at your taste, if not convinced, then choose a higher bitrate until you can listen to it comfortably. As long as you are not sure of what sample rate and bitrate you prefer for that, keep the recording on a *.WAV file to redo it, because you cannot remove the mp3 artifacts. Once you have decided, you can encode directly in mp3 after editing and remove the then useless *.WAV files.
I thank you for your help. I still don’t understand this, but you have helped quite a bit.
What should I do?
I don’t know what the treeble range is with such a recording? I don’t really care about that, but do care about just understanding what was said & file size.
I am not sure what to think about this?
> 48 kbps & 16 kbps is the resulting speed of the "stream" obtained after the compression, called the bitrate. Due to the compression system applied to save space, the quality of the sound will be better with high values and will decrease with low ones, because the compression makes to loose precision
I really don’t have much quality? I can hardly understand the words sometimes on some tapes.
I wish that I could understand this, since I think you are telling me how I can improve what I am doing?
>Old tapes with low quality recording talk: a 22050 kHz sampling rate would be enough, or even 16000 Hz, just see the spectrum analyser on hissing vowels (SSS's) before to choose, take the double of the highest frequencies who are not tape noise
How do I do this?
I guess what you are saying here is, that I should make them into wave files & then change them into whatever MP3 files that I decide?
>Final choice of the bitrate: it's to you to decide if the quality is good enough at your taste, if not convinced, then choose a higher bitrate until you can listen to it comfortably. As long as you are not sure of what sample rate and bitrate you prefer for that, keep the recording on a *.WAV file to redo it, because you cannot remove the mp3 artifacts. Once you have decided, you can encode directly in mp3 after editing and remove the then useless *.WAV files.
I thank you for your help. I still don’t understand this, but you have helped quite a bit.
What should I do?
Later,
-- Richard --
-- Richard --
-
Blandine Catastrophe
- Posts: 253
- Joined: Sun Jul 24, 2005 3:25 pm
As long as you haven't finished to process the sound, it is better to keep it in WAV format, and to convert to mp3 only when finished.
You can use the noise reducer by picking a sample of only tape noise without any voice or room's/street's noise to the clipboard and using the clipboard option in the noise reducer. If there is much noise, you may reduce the scale to avoid artifacts (if you hear starwar noises when you verify the result, these are the artifacts)
You can filter the useless frequencies. If you are a complete beginner, a solution is to seek it in the highpass/lowpass filter. You set it on "highpass" and to the maximal steepness. Then you increase the Hz value until you don't hear anything from the voice, you change the setting to "lowpass" and reduce the steepness to about 5 and click "OK". Then you open the same effect, but you start to monitor with the "lowpass" enabled and reduce the frequency until you hear nothing from the voice. Once you have found, you enable the "highpass" and apply it with the steepness at five.
The recording isn't stereo, but what I thought was that the player is. If it is the case, you can verify with the "channel mix" by monitoring "LEFT & LEFT" and "RIGHT & RIGHT" the one who gives the best sound. By selecting only one channel when the tape is played on a stereo device, you avoid also some phasing effect : variable gaps in the frequency spectrum, giving a kind of "psychedelic" sound and due to micro-desynchronizations of less than a milliseceond, the tape has a trend to warp with age.
After you have restored what you can do, you will convert the WAV to mp3, and for that you must decide by comparing the quality by listening. For what you described, I think that a 22050 or 16000 Hz sample rate will be enough. When still in *.WAV format, use the resample button.
A monaural format will be the most appropriate, just "save as" and choose "16 bit mono" in the menu.
Then, to you to verify what quality you can listen to to your taste for the final choice of the bitrate. If the material is similar on all the tapes, you can try on one file first, and you needn't to verify much for all the other ones, that you can convert directly with the same settings. By prudence take one or the other mp3 at random and listen to a random section before to erase the WAV files.
For my part, I always keep the original untouched sound in a WAV file in a separate folder as long as my restoration & editing is not finished, and I erase it only when I'm sure to get the good final result. For the music restoration, I keep all these original files on one, two even sometimes trhree Data CD-R for the case where something happens with the original support. But with my Alsacian theater records who are the unique known copies and a unique historic patrimony of the region, or 1900's-1920's records, the security level isn't forcefully the same that you need for your purpose. To keep the original untouched until the work is definitely finished let you redo it if the result is too bad and can be improved.
You can use the noise reducer by picking a sample of only tape noise without any voice or room's/street's noise to the clipboard and using the clipboard option in the noise reducer. If there is much noise, you may reduce the scale to avoid artifacts (if you hear starwar noises when you verify the result, these are the artifacts)
You can filter the useless frequencies. If you are a complete beginner, a solution is to seek it in the highpass/lowpass filter. You set it on "highpass" and to the maximal steepness. Then you increase the Hz value until you don't hear anything from the voice, you change the setting to "lowpass" and reduce the steepness to about 5 and click "OK". Then you open the same effect, but you start to monitor with the "lowpass" enabled and reduce the frequency until you hear nothing from the voice. Once you have found, you enable the "highpass" and apply it with the steepness at five.
The recording isn't stereo, but what I thought was that the player is. If it is the case, you can verify with the "channel mix" by monitoring "LEFT & LEFT" and "RIGHT & RIGHT" the one who gives the best sound. By selecting only one channel when the tape is played on a stereo device, you avoid also some phasing effect : variable gaps in the frequency spectrum, giving a kind of "psychedelic" sound and due to micro-desynchronizations of less than a milliseceond, the tape has a trend to warp with age.
Try to correct that with the equalizer with a similar correction:I really don’t have much quality? I can hardly understand the words sometimes on some tapes.
After you have restored what you can do, you will convert the WAV to mp3, and for that you must decide by comparing the quality by listening. For what you described, I think that a 22050 or 16000 Hz sample rate will be enough. When still in *.WAV format, use the resample button.
A monaural format will be the most appropriate, just "save as" and choose "16 bit mono" in the menu.
Then, to you to verify what quality you can listen to to your taste for the final choice of the bitrate. If the material is similar on all the tapes, you can try on one file first, and you needn't to verify much for all the other ones, that you can convert directly with the same settings. By prudence take one or the other mp3 at random and listen to a random section before to erase the WAV files.
For my part, I always keep the original untouched sound in a WAV file in a separate folder as long as my restoration & editing is not finished, and I erase it only when I'm sure to get the good final result. For the music restoration, I keep all these original files on one, two even sometimes trhree Data CD-R for the case where something happens with the original support. But with my Alsacian theater records who are the unique known copies and a unique historic patrimony of the region, or 1900's-1920's records, the security level isn't forcefully the same that you need for your purpose. To keep the original untouched until the work is definitely finished let you redo it if the result is too bad and can be improved.
Last edited by Blandine Catastrophe on Thu Aug 04, 2005 9:19 am, edited 3 times in total.
Gloup? :-°