Not sure which forum is best for this post so hopefully this is the right place. 😀
I would like to calibrate the output of my vinyl chain for recording purposes. I have the HFN test record and I am thinking the 300Hz test tracks could be used for that. I am having trouble making sense of the various units and required headroom though.
Assuming I use the +12dB 300Hz track for a reference level and I want to calibrate gain for -10dBV consumer line, what voltage should I see at the line output? I’m thinking 1.26 Vrms. Assuming I am on the right track here, what should I consider the peak level for music signals? Adding 20 dB to -10dBV I get 3.16 Vrms, is that the right number and is it safe? I have the capacity to limit the clicks in the recording chain so primarily interested in music signals.
Thanks in advance,
Marty
I would like to calibrate the output of my vinyl chain for recording purposes. I have the HFN test record and I am thinking the 300Hz test tracks could be used for that. I am having trouble making sense of the various units and required headroom though.
Assuming I use the +12dB 300Hz track for a reference level and I want to calibrate gain for -10dBV consumer line, what voltage should I see at the line output? I’m thinking 1.26 Vrms. Assuming I am on the right track here, what should I consider the peak level for music signals? Adding 20 dB to -10dBV I get 3.16 Vrms, is that the right number and is it safe? I have the capacity to limit the clicks in the recording chain so primarily interested in music signals.
Thanks in advance,
Marty
I believe you're over thinking this process. The goal is to get the audio into the recording device (the PC or some other recorder?) at the highest level WITHOUT clipping. I would start with the test disc and set the record level to something below clipping like minus 3dB (arbitrary start point) and then record your LP. If the levels are OK at the end, you're done. If there is clipping, back off the levels 3dB and do it again.
Keep in mind there is NO standard level for LPs. I've had some Columbia Masterworks discs that were a few dB lower because of a larger amount of time. Also keep in mind the equivalent bit depth for an LP isn't that large compared to the 16 bit digital so you have wiggle room.
Your digital systems have a very well defined max level point where you run out of bits and hard clipping begins. How THAT level relates to analog levels will depend on design choices and calibration. I've seen 2 Volts p-p and 2 Volts RMS for analog outputs at full scale. You can check this all easily by generating a signal in Audacity or Audition (or your favorite software) and just measure the analog levels.
Pro systems often use +4dBm at operating level ( 20dB below Full Scale )
G²
Keep in mind there is NO standard level for LPs. I've had some Columbia Masterworks discs that were a few dB lower because of a larger amount of time. Also keep in mind the equivalent bit depth for an LP isn't that large compared to the 16 bit digital so you have wiggle room.
Your digital systems have a very well defined max level point where you run out of bits and hard clipping begins. How THAT level relates to analog levels will depend on design choices and calibration. I've seen 2 Volts p-p and 2 Volts RMS for analog outputs at full scale. You can check this all easily by generating a signal in Audacity or Audition (or your favorite software) and just measure the analog levels.
Pro systems often use +4dBm at operating level ( 20dB below Full Scale )
G²
HaHa Likely so 🙂
The motivation was to come up with what would be a consistent level for archiving my LP’s to 24/96 with older standalone pro/semi-pro equipment. I would like the analog signal to be as hot as possible for universal use and repeatable in case I switch equipment during the project. I can calibrate the AD converter to a wide range of inputs so I’m thinking once I can predict the maximum input level it should be easy enough to avoid digital clipping.
The usual way is to set with a 0dB 1kHz reference track to somewhere between -20 and -24dBFS on your digitial recorder. This gives headroom for peaks as well as for ticks and pops.
LPs were cut all different levels. There's no "hard limit" like digital, or "soft limit" like tape. The primary limit is getting enough time per side (high levels are wider wiggles). Another limit is: can the consumer play it? There's fair limits on what the play needle can track. (Yes the cutting heads and amp have limits but they are often far past any real need.)
Preamp output level: we see some pres with 35dB gain and 12V supply, but "better" pres often run 35dB-40dB gain with +/-15V supplies. This means your output level "may be" 7V or 9V RMS.
Since this is a "final cut" (I assume you will burn your LPs and players once digitized), I would drive the preamp to clipping, call that digital Full Scale, and copy. If it didn't clip in analog it won't be clipped in the ADC. The resulting files may have max peaks 15dB-20dB below full scale; but it is easy to normalize (ugh) in software.
Preamp output level: we see some pres with 35dB gain and 12V supply, but "better" pres often run 35dB-40dB gain with +/-15V supplies. This means your output level "may be" 7V or 9V RMS.
Since this is a "final cut" (I assume you will burn your LPs and players once digitized), I would drive the preamp to clipping, call that digital Full Scale, and copy. If it didn't clip in analog it won't be clipped in the ADC. The resulting files may have max peaks 15dB-20dB below full scale; but it is easy to normalize (ugh) in software.
Before you go 24/96 I strongly suggest you watch this 24 minute video on Xiph.org. It will put a lot of things in perspective. I found it fascinating.
Xiph.Org Video Presentations: Digital Show & Tell
G²
24/96 for ripping has a number of advantages even if you then downsample to 16bits for replay. Not least headroom just in case.
Thanks for all of the input
I think I’m seeing things more clearly now. With all the headroom and ability to normalize it is just not as critical as I thought it was. It shouldn’t take much fooling around to come up with a level that always works
Marty
I think I’m seeing things more clearly now. With all the headroom and ability to normalize it is just not as critical as I thought it was. It shouldn’t take much fooling around to come up with a level that always works
Marty
Yes, you can find a level which always works. For some LPs it will be suboptimal, so you may not find a level which is always right.
Look at the record in a shallow angle, so that the light reflect on the tracks. Look for the rough portion on the surface, put down your needle there. The modulation most likely will be the highest there and adjust your meters. Every record is different.
EPs for DJ / club use can be pretty loud, maybe there are some of those in the collection. Aim for -6 dBFS on the peaks, and you should be pretty much set. If the test record contains some tracks of unreasonably high level, go for -3 dBFS maximum.
BTW, switch recording input level display on the computer to dB and avoid any settings below 0 dB - chances are it's just digital attenuation at this point. Actually, for quite a number of devices the software-accessible control is all digital attenuation. (This applies to Asus Xonar cards, for example.) This could result in a false sense of security when the input is actually clipping.
Onboard audio tends to have several stages of additional analog gain, this should not be necessary here and can be left at 0 dB (it helps improve SNR with weak sources).
These days, just about any line input in a decent quality sound device should provide a dynamic range in excess of 100 dB, and at times up to 120 dB, more than good enough for recording vinyl with some headroom. Getting a 24-bit recording at the desired sample rate can be a bit involved though. The Windows user tends to be best off using ASIO (with ASIO4All if need be). Otherwise there may be up to three different settings that have to be matched (recording software, interface sample rate and hardware sample rate). If your preferred recording software is Audacity, be aware that the only sound API to allow 24-bit (and bigger) samples in a stock build is WASAPI, and that will only complain about a sample rate mismatch but cannot adjust anything. ASIO is not included for licensing reasons by default, but you can compile your own build with it included relatively easily (as far as these things go).
BTW, switch recording input level display on the computer to dB and avoid any settings below 0 dB - chances are it's just digital attenuation at this point. Actually, for quite a number of devices the software-accessible control is all digital attenuation. (This applies to Asus Xonar cards, for example.) This could result in a false sense of security when the input is actually clipping.
Onboard audio tends to have several stages of additional analog gain, this should not be necessary here and can be left at 0 dB (it helps improve SNR with weak sources).
These days, just about any line input in a decent quality sound device should provide a dynamic range in excess of 100 dB, and at times up to 120 dB, more than good enough for recording vinyl with some headroom. Getting a 24-bit recording at the desired sample rate can be a bit involved though. The Windows user tends to be best off using ASIO (with ASIO4All if need be). Otherwise there may be up to three different settings that have to be matched (recording software, interface sample rate and hardware sample rate). If your preferred recording software is Audacity, be aware that the only sound API to allow 24-bit (and bigger) samples in a stock build is WASAPI, and that will only complain about a sample rate mismatch but cannot adjust anything. ASIO is not included for licensing reasons by default, but you can compile your own build with it included relatively easily (as far as these things go).
Thanks guys, Just to be clear, there is no PC in the recording chain. I’m using all standalone hardware so I probably have a bit more flexibility than I would with a usb interface. After digesting the earlier replies it finally dawned on me that I might have been approaching this from the wrong side of the dynamic range. As I see it now, there is headroom to spare at 24/96 and the ability to normalize so why worry about setting levels by the high end of the dynamic range when I can set up so that I am just seeing the noise of the analog side in the recording. As a practical matter, it is a lot easier to find the super quiet high dollar pressings in my collection than trying to pick out the hottest of the hot. Even if I set up so the noise is 10dB above digital black I end up with way more headroom than I need for vinyl.
Thanks again to you all for the advice
Marty
Thanks again to you all for the advice
Marty
Another question. Are you using an RIAA preamp or are you going to do the
RIAA curve in digital?
G²
RIAA curve in digital?
G²
Marty: A quicky and dirty with a lot of recorders is that they have a green and red level lights. Green is at -20dBFS. So play a 1kHz 0dB test track and adjust till that light just comes on and you should be good to go. (or just off for a bit more headroom).
A note on clicks and pops. They are usually very short duration (1 sample in many cases) so allowing clipping on those is not generally a problem and any clickrepair software should remove them.
A note on clicks and pops. They are usually very short duration (1 sample in many cases) so allowing clipping on those is not generally a problem and any clickrepair software should remove them.
The chain is LOMC > SUT (19dB) > RIAA pre (41 dB)> Line driver (10 dB)> ADC> HDD Recorder.
You are correct, the ADC has a light at -20dB FS and I can zoom in to 1dB per led so accurately setting that level should not be a problem. I don’t think I have a test record with a 1kHz test tone though. Is there one currently available that would work?
most test records have that on. Analogue Productions-The Ultimate Analogue Test LP-Turntable Set Up Tools|Acoustic Sounds as example, but there are cheaper ones if you check ebay.
- Status
- Not open for further replies.