EQing for flat cuts on T560

[Synthlord]

Hello all,

This is my first post after reading all of the great information posted over the years in the forum. As the new owner of a T560 many of the posts have saved me a lot of time and blanks, so thank you to all the contributors!

I've set myself a task of getting as close to the eq curve of a digital source as possible when cutting. In short my method is:

- Find a great modern pressed vinyl record
- Find a great digital copy of the same record
- Cut the digital copy using the T560
- Compare the cut to the pressed original

What has amazed me is just how close the digital and pressed record sound. However, my cut is very obviously missing high frequency content and the EQ is not flat against the digital source. I am using Souri's recommended settings at the moment.

My plan to remedy this is to:
- Replicate the RIAA curve from the main unit digitally. (I would like to go direct to the cutting amp for maximum flexibility)
- Cut some test tones / sweeps to assess the response
- Adjust the EQ to match

Before I do this however I would like to save myself some trial and error by understanding the following. Please let me know if you can answer any of these points:
- What is the best EQ response I can hope for from the T560?
- Is there a highest and lowest possible frequency that can be cut?
- What conditions need to be meet to do this assuming a full range digital source? (Heat, RIAA EQ settings etc)
- What should I automatically filter going to the cutting amp (40hz and below, 20khz and above?)

Any help would be greatly appreciated! I will share any discoveries I make on the way.

Thanks,

SL

[markrob]

Hi,

I don't own a VR, but here are a few things to think about:

I would first capture the open loop response of the head with no IRIAA so that you can see what you are dealing with and determine the way to proceed. Use caution when running any test tones, noise or sweeps into the head. This is particularly true if you employ IRIAA in your testing due to the large high frequency boost that the curve employs. From the your initial testing, you should be able to determine the main system resonance of the head along with any secondary resonances that could aid or impede you ability to flatten the response. Using the full IRIAA curve on a moving coil head is potentially a duplicated effort in the range of 500 Hz to system resonance. Below resonance, the open loop response of a typical head falls at -6db/oct. This matches the IRIAA curve, so if you employ the textbook curve you will end up having to EQ it back out later. It is better to use a modified curve that only has the 50 Hz and 2122 Hz turnovers. This gives you needed pre-emphasis but does not alter the midrange. Even with the modified IRIAA curve in place, you will still need quite a bit of high frequency boost to flatten the head. In a perfect world, a textbook moving coil head open loop response falls at -6db/oct above system resonance. So if you assume a 1 Khz resonance you would need about +20db of gain at 20 Khz in addition to the gain needed for the IRIAA. It gets pretty scary. Depending, you may find some secondary resonances that actually help you out at the high end. However, you can see that due to the extreme boosts required, you will need large amounts of power a the frequency extremes to flatten the head. You might find that you need to limit the response of of the head EQ up at the extremes and/or limit you source material to prevent meltdown. Making good measurements of the response can be tricky. If you use a white or pink noise source, make sure you are not measuring the noise floor of the system. Cut a reference silent section and look at the spectrum to see where the floor is. It can be a good idea to cut with program material so you can A/B and hear if the highs are really being cut. Also note that if you playback with a standard phono preamp, you will need the IRIAA in the path somewhere or your measurements will be incorrect. A good way to deal with this is to build up a phono preamp that has flat response (no RIAA). This gives you the true velocity response of the head since your typical magnetic pickup is velocity responding. If you can't do that, then use either use the modified IRIAA curve for your cuts or roll the IRIAA into your final EQ. In that case you'll need to be able to boost as much as +30db at 20khz (if you try to go that far). The other option is to use the EQ provided with the VR and just add some additional to flatten the head.

One other thought. Depending of the mechanical impedance of the head, the cutting media and conditions will contribute to the head response. If the head is really stiff, the effect of the media might be small. However a stiff head typically means it is not very sensitive.

Hope you find this useful

Mark

[okpanda]

comparing a pressed record to lathe-cut vinyl isn't exactly apples to apples. a pressed record was initially cut into lacquer on most likely a very high end machine. you are cutting into pvc with souri's dynamic cutterhead (i would assume) and idk what turntable. i'm not saying you can't get good cuts with the vinylrecorder, but it takes alot of time, research, and money.

[2bitcomputer]

@markrob,

I have a related question I've been wondering about for a while:
Since the high-frequency response of the outside(towards the edge) of the record is noticeably better than the high-frequency response of the inside(towards the label) where do you cut the testones/testnoise for your EQing?
In the middle?
Is there a standard for this??

Thanks...

[markrob]

Hi,

I would cut at the outside and at a faster speed if you can. For the most part, the losses at inner diameter is due to playback issues, not cutting. So you don't want this to affect your measurements. Let the losses at the inside happen but don't try to EQ them out in your head calibration. That's my take, but I would be interested in other opinions.

Mark

[Synthlord]

Thank you Mark and okpanda for your input.

@okpanda Do you have any experience on what the best possible cut I can get from the VR is? If decent treble is not possible due to the limitations of the system rather than the processing before the cutting amp then I will not push it too hard. The challenge is I do not have a reference to aim for, only the best pressed records I have.

@markrob Thank you for the detailed explanation. I am trying to keep the EQ adjustments in the computer for simplicity. Would the following steps make sense given I do not have a phono amp without RIAA?

- Measure RIAA response of phono amp using low level pink noise from line source
- Create IRIAA curve in computer to flatten pink noise
- Cut pink noise with (no EQ)
- Access difference and EQ as appropriate
- The final EQ becomes a combination of the IRIAA curve and secondary EQ adjustments

Thanks,

SL

[markrob]

Hi,

Based on your setup, I would roll the IRIAA into your final EQ and not bother with it. If you cut a noise source (make sure you pick the correct one based on your analyzer) and playback with a standard phono preamp that has the correct RIAA built in, you will find that the head is flat from about 50 Hz to near system resonance. You will see the peak in the response due to that resonance (could be 10-20 db) and then see a fast drop off above. Depending on the head, you may also see some peaks and valleys due to secondary resonances. Add EQ to flatten the response as needed. In addition to the high frequency EQ, you will need to add a bit of bass boost below 50 Hz to flatten the head response. Be careful here as your playback spectrum can get swamped by low frequency rumble and 60/50 Hz mains pickup. If you end up too far down in the mud at the high frequency extremes, then adding in high frequency pre-emphasis (just a simple shelving EQ starting around 2Khz and dropping out at 20 Khz will do) on the cutting side will help get above the nose floor. The shelving EQ then becomes part of the total corrective EQ needed on the cutting side.

You can also improve S/N if you use a sweep rather than noise. This is because the energy is spread through the entire bandwidth of the noise signal. If you go that route, make sure the chain is flat into the head or you could very easily burn it up if there are any excessive boosts. Its not as easy to make adjustments using sweeps.

Finally, if you are really trying to push the bandwidth of the head, you might consider trying half speed mastering. In that case the highest you will have to correct is 10 Khz. But this approach opens another can of worms.

Mark

[Thelurker]

I made a perfect DSP chain for the Vinyl Recorder when I owned one. It took over 100 blanks and numerous diamonds but the end result was perfect “press” like cuts every time. Forget Souri’s settings, he’s a great mechanical engineer/designer but far from a half decent sound engineer. The fact he peddles that crappy 2 channel mixer is a bit of a give away.

As explained by others on this thread, you'll need to correct the response of the head using pink noise but don’t do this on full output, the head or fuses will pop.

You’ll also need to add some carefully placed gentle multi band dynamics mainly in the mid to high bands to tame the sibilance's and harsh peaks.

[tragwag]

A bold task! I did something far less intensive when I got my T560.
my head tends to muck up the high mid range (around 6khz) far before 0db ref +4dbu
so I've focused on that area and taming it in regards to overall volume tradeoff